TOWARDS AN OBJECTIVE EVALUATION OF TEACHER PERFORMANCE Journal website: http://epaa.asu.edu/ojs/ Manuscript received: 24/6/2016 Facebook: /EPAAA Revisions received: 1/5/2017 Twitter: @epaa_aape Accepted: 17/5/2017 education policy analysis archives A peer-reviewed, independent, open access, multilingual journal Arizona State University Volume 25 Number 76 July 17, 2017 ISSN 1068-2341 Compulsory Education Laws or Incentives from Conditional Cash Transfer Programs? Explaining the Rise in Secondary School Attendance Rate in Argentina María Edo Universidad de San Andrés and CONICET Mariana Marchionni & Santiago Garganta CEDLAS-Universidad Nacional de La Plata and CONICET Argentina Citation: Edo, M., Marchionni, M., & Garganta, S. (2017). Compulsory education laws or incentives from conditional cash transfer programs? Explaining the rise in secondary school attendance rate in Argentina. Education Policy Analysis Archives, 25(76). http://dx.doi.org/10.14507/epaa.25.2596 Abstract: Argentina has traditionally stood out in terms of educational outcomes among its Latin American counterparts. Schooling of older children, however, still shows room for improvement especially among the more vulnerable. Fortunately, during the last years a sizeable improvement in attendance rates for children aged 15 through 17 took place. This could be related to the 2006 National Education Law that made upper -secondary education compulsory. In this paper, instead, we claim that the Asignación Universal por Hijo (Universal Child Allowance, AUH) —a massive conditional cash transfer program implemented in 2009 in Argentina — may be mostly responsible for this improvement. Using a difference -in-difference strategy we estimate that the program accounts for a 3.9 percentage point increase in the probability of attending secondary school among eligible children aged 15 through 17. The impact seems to be http://epaa.asu.edu/ojs/ http://dx.doi.org/10.14507/epaa.25.2596 Education Policy Analysis Archives Vol. 25 No. 76 2 led by boys and is more relevant for children living in larger families where the head of household has a lower educational level. Keywords: conditional cash transfers; education; attendance; Argentina ¿Leyes de educación obligatoria o incentivos de programas de transferencias condicionadas? Explicando el aumento de la tasa de asistencia al secundario en Argentina Resumen: Argentina tradicionalmente se ha destacado en términos de resultados educativos entre sus pares latinoamericanos. La escolarización de los adolescentes, sin embargo, todavía debe mejorarse, especialmente entre los más vulnerables. Afortunadamente, durante los últimos años se produjo un considerable aumento en las tasas de asistencia escolar al secundario entre los jóvenes de 15 a 17 años. Esto podría estar relacionado con la Ley Nacional de Educación de 2006 que transformó la enseñanza secundaria en obligatoria. En este artículo, en cambio, sostenemos que la Asignación Universal por Hijo (AUH), un programa masivo de transferencias monetarias condicionales implementado en 2009 en Argentina, puede ser el principal responsable de esta mejora. Utilizando una estrategia de diferencias en diferencias, estimamos que el programa representa un aumento de 3,9 puntos porcentuales en la probabilidad de asistir a la escuela secundaria entre los jóvenes elegibles de 15 a 17 años. El impacto parece estar liderado por los varones y es más relevante para los jóvenes que viven en familias más grandes donde el jefe del hogar tiene niveles educativos más bajos. Palabras-clave: programas de transferencias condicionadas; educación; asistencia; Argentina ¿Leis de educação obrigatória ou incentivos de programas de transferências condicionadas? Explicando o aumento na taxa de frequência escolar no ensino secundário na Argentina Resumo: A Argentina tradicionalmente se destacou em termos de resultados educacionais entre seus pares latino-americanos. A escolarização de crianças mais velhas, no entanto, ainda mostra espaço para melhoria, sobretudo entre as mais vulneráveis. Felizmente, durante os últimos anos teve lugar uma melhoria considerável das taxas de frequência escolar para crianças de 15 a 17 anos. Isso poderia estar relacionado à Lei de Educação Nacional de 2006 que tornou obrigatório o ensino secundário. Neste documento, porém, afirmamos que a Asignación Universal por Hijo (Subsídio Universal para Crianças, ou AUH) - um amplo programa de transferencia condicionada de renda implementado em 2009 na Argentina - pode ser o principal responsável por essa melhoria. Usando uma estratégia de diferença em diferença, estimamos que o programa representa um aumento de 3,9 pontos percentuais na probabilidade de frequentar a escola secundária para crianças elegíveis de 15 a 17 anos. O impacto parece ser liderado por meninos e é mais relevante para as crianças que vivem em famílias maiores, onde o chefe de família tem um nível educacional mais baixo. Palavras chave: transferências condicionadas de renda; educação; frequência escolar; Argentina Introduction Argentina has traditionally stood out within Latin America in terms of education. Since the very creation of the National Education System in 1884, primary education has been mandatory in Argentina. This and the free public provision of educational services have allowed to reach almost perfect rates of primary school attendance, which have remained relatively stable above 97% since the 1980s and are comparable and even higher than those of developed countries (Marchionni & Alejo, 2015). Explaining the rise in secondary school attendance rate in Argentina 3 In contrast, secondary education has not always been mandatory in Argentina. By the early 1990s, only the seven years of primary education were compulsory. In 1993, the Federal Education Law 24,195 (Ley Federal de Educación) increased compulsory education from seven to 10 years, thus including the first stage of secondary education. The National Education Law 26,206 (Ley Nacional de Educación) passed in December 2006 extended compulsory education by three more years, making mandatory also the upper-secondary education level. Secondary education indicators improved markedly since the mid-1990s, and some argue that these improvements are a consequence of the successive expansions of compulsory education (DiNIECE, 2011). For the case of the 1993 Federal Education Law, Alzúa et al. (2015) find a positive effect on school enrollment and attainment, but the mechanisms remain unclear since the 1993 reform combined an expansion in compulsory education with deep institutional and curricular modifications, among other changes. Over the last decade, the net school attendance rate for the group aged 15 to 17—the upper- secondary age range—rose by almost 4 percentage points, from 82.9% in 2004 to 86.6% in 2014.1 Our first hypothesis is that this improvement was not caused by the 2006 law. First, neither the law nor accompanying policies had enforcement mechanisms embedded in their design. Therefore, it is unclear through which channels the law may have affected school attendance. Second, three years after the law was passed, attendance rates for the group aged 15 to17 remained virtually unchanged. Only since 2010 school attendance for individuals in this age group started to show clear signs of growth. But if the 2006 National Education Law showed no impact on net attendance of those aged 15-17, what is driving the increase in those rates as of 2010? What is bringing children aged 15 to 17—especially those most poor—to stay in school? In this paper we claim that the Asignación Universal por Hijo (Universal Child Allowance, AUH), a program implemented in Argentina in late 2009, may be driving this increase in attendance rates. The AUH is a massive conditional cash transfer program (CCT) targeted at children under 18 years old living in poor families with no registered workers in the formal employment sector. The benefit consists of monthly cash transfers per child, and it is paid up to a maximum of five dependent children. As any typical CCT, the goal of the AUH is twofold: to provide social protection to the more vulnerable families while promoting the formation of human capital to break the intergenerational transmission of poverty. Therefore, cash transfers are conditional on complying with children’s health checkups and school attendance at all compulsory levels. Concerning the potential impact of AUH on schooling, the economic incentives introduced by the program and its conditions may both reduce the likelihood of beneficiary children dropping out of school and encourage dropouts to get back to the education system. Since the program covers a large proportion of Argentinian children and the cash transfer represents a large increase of household income for beneficiary families (Garganta et al., 2016), the impact on school attendance could be potentially strong. Nevertheless, only the empirical evidence can determine the actual relevance of this effect. Estimating the causal impact of the AUH on school attendance, however, represents a difficult task. The AUH was not assigned randomly nor was it accompanied by a publicly available comprehensive dataset that allows for assessing the program. We thus resort to the Permanent National Household Survey (Encuesta Permanente de Hogares, EPH) carried out in Argentina. Based on this data, we classify children in upper-secondary age-range (15 to 17) as potential beneficiaries 1 Net school attendance rate is the percentage of children in a given age group that attend the educational level that officially corresponds to that age (UNESCO). Education Policy Analysis Archives Vol. 25 No. 76 4 according to whether their parents comply with the program’s eligibility requirements. We thus compare the probability of secondary school attendance of both groups (eligible and not eligible) over time following a difference-in-difference approach. Our estimates suggest that the AUH increased the probability of attending secondary school among eligible children aged 15-17 by 3.9 percentage points. The impact is more relevant for children living in larger families where the head of household has lower education levels, and seems to be driven by boys’ behavior. The effect on younger children is statistically significant yet very small: 0.4 percentage points for those in primary school age range (6 through 11) and 0.8 percentage points for those in lower secondary (12 through 14 years old). The results hold across different specifications and robustness analysis. These findings are in line with the evidence available to date from other similar programs. Overall, CCTs have had a positive and significant impact on school enrollment and attendance, especially among children from the more vulnerable households, whose initial attendance rates are the lowest (Saavedra & García, 2012; Fiszbein et al., 2009). Moreover, the impact is often larger for secondary school and increases with the size of transfers. This paper intends to make contributions in several realms. First of all, it adds to the literature on the impact of CCT programs on educational outcomes. Secondly, it provides evidence of the effects of the Asignación Universal por Hijo, thus generating input for future improvements of the program. Finally, this work also seeks to highlight the fact that compulsory education laws by themselves are not enough to affect schooling. The rest of the paper is organized as follows. The next section expands on compulsory education legislation in Argentina while presenting evidence on the evolution of net attendance rates over the last decade. The third section describes the AUH and discusses the channels that may affect schooling decisions. Then, we present the data and methodology, and in the fifth and sixth sections discuss the results. The last section concludes and points to further research. Compulsory Education Laws and School Attendance in Argentina Compulsory education laws are motivated by the potential social benefits and positive externalities coming from an expansion of the overall education attainment which promotes economic development (Oreopoulos, 2006a). These laws may affect attendance rates through different channels. In the first place, the human capital model of school choice perceives education as an investment (Becker, 1975) and hence depends on intertemporal benefits and costs of schooling. Consequently, compulsory education may prevent a probably optimal decision of leaving school. However, compulsory attendance laws may raise lifetime welfare if they generate positive externalities or under the presence of suboptimal school attainment (Oreopoulos, 2006a; Eckstein & Zilcha, 1994), which is likely among the more vulnerable children in developing countries like Argentina. Secondly, these legislations may trigger implicit enforcement mechanisms, by imposing social stigma to those who fail to comply with the rule. Fulfillment of mandatory schooling may also affect future opportunities in the labor market if, for instance, legal educational requirements are set as a condition to enter the formal employment sector (Alzúa et al., 2015). Finally, other public policies accompanying the launch of these legislations may have an impact on attendance rates by affecting the direct costs of education (abolition of tuition fees), the quality of education (increase in educational budget, drastic changes in the curricula) or the availability of nearby educational facilities (large-scale infrastructure programs), among others. Unfortunately, evidence of the impact on attendance rates of changes in compulsory education laws is relatively scarce. Most studies concentrate on the effects regarding labor market outcomes (Acemoglu & Angrist, 2000; Angrist & Kruger, 1991; Oreopoulos, 2006a, 2006b). Even Explaining the rise in secondary school attendance rate in Argentina 5 though some studies document the improvement of attendance rates following mandatory education laws (Goldin and Katz, 2008; Lleras Muney, 2002; Oreopoulos, 2006a), the mechanism through which the effect operates is not entirely clear. Compulsory education laws are usually launched together with other policies aiming at increasing school attendance. Therefore, some or all of the abovementioned channels operate at the same time, hindering the possibility of isolating the impact of the expansion of compulsory education by itself. Regarding Argentina, while primary education has always been mandatory, it was only in the early 1990s that compulsory schooling expanded to secondary education. The Federal Education Law, passed in 1993, increased mandatory education from seven to 10 years of schooling, thus including the first stage of secondary education (children up to 14 years old). Later, in 2006, the National Education Law added three more years of compulsory education, covering also the upper- secondary level (youths between 15 and 17 years old).2 Table 1 summarizes the timing and scope of these reforms. Table 1 Extension of compulsory education in Argentina Age Common Education Law Federal Education Law National Education Law Modification to National Education Law Year: 1884 Year: 1993 Year: 2006 Year: 2015 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Compulsory Years 7 10 13 14 Sources: Common Education Law (1884), Federal Education Law (1993), National Education Law (2006). Some argue that these successive expansions in mandatory schooling are responsible for the observed improvements in secondary education indicators since the early 1990s in Argentina (DiNIECE, 2011). However, the evidence is not so clear. Alzúa et al. (2015) evaluate the impact of the 1993 law by taking advantage of the different timing in the implementation of the reform. They find that the 1993 law was followed by a notable increase in gross enrollment rates and had a positive impact on years of schooling for children aged 13-14. However, as stated by the authors, the main mechanism driving the effect is hard to identify since the new legislation was accompanied by 2 Only four other Latin American countries have passed equivalent legislation (i.e. mandatory schooling for both primary and secondary education): Uruguay in 2008, Chile and Brazil in 2009 and Mexico in 2013 (Ruiz & Schoo, 2014). Education Policy Analysis Archives Vol. 25 No. 76 6 changes in the curricula and a strong expansion of the education budget to finance investment in school infrastructure as well as teacher’s training. Figure 1. Net school attendance rates by age group Source: own estimations based on EPH. Note: net school attendance rate is the percentage of children in a given age group that attend the educational level that officially corresponds to that age (UNESCO). Ages 6-11 correspond to primary school; ages 12-14 and 15-17 correspond to lower and upper secondary school, respectively. Figure 1 shows that by 2004, net attendance rates for children aged 6 to 11 (primary school age) and 12 to 14 (lower-secondary school age) were above 97% and remained rather stable over the following decade. Compared to these younger children, those aged 15 to 17 exhibit markedly lower attendance rates (82% in 2004). Even though for this latter group education became compulsory in 2006, net attendance rates remained mostly unchanged over the following three years.3 Only after 2009 net attendance rates started to significantly grow for 15-17 year-olds, from 82.9% in 2009 to 86.6% in 2014, i.e. an almost 4-percentage-point increase. Administrative data shows a similar pattern for secondary school enrollment. In fact, the number of students enrolled in that educational level remained rather stable during the period in between the National Education Law and 2010, increasing by only 1.5%. On the contrary, the number of students enrolled in secondary school increased by 7.6% since 2010 (DiNIECE: http://portales.educacion.gov.ar/diniece/).4 The preliminary evidence in Figure 1 suggests that the 2006 National Education Law had no impact on net attendance rates on the first three years after its implementation, which is not surprising given that there were no companion measures that could have encouraged school attendance. In fact, even though there was a large expansion of the educational budget, new funds were almost entirely absorbed by salaries, with no investment in training or systematic infrastructure 3 In fact, attendance rates for the group of 15-17 year-olds follow a similar pattern to the 12-14 year-old group over the 2004-2009 period, even though the latter group was not affected by the law. This is confirmed by a difference-in-difference estimation. These results are available upon request. 4 It is worth noting, however, that these figures are not strictly comparable to the attendance rates presented in Figure 1 for two reasons: they represent the number of students enrolled rather than net attendance rates and they are not specific to the 15-17 age group. Unfortunately, administrative data is not available by age. 8 0 8 5 9 0 9 5 1 0 0 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 6-11 12-14 15-17 http://portales.educacion.gov.ar/diniece/ Explaining the rise in secondary school attendance rate in Argentina 7 development, and only quite limited changes in the curricula.5 Moreover, despite some specific programs were developed to complement the new education law they were more focused on establishing an adequate normative framework and on improving institutional arrangements than in providing direct or indirect incentives to school attendance (UNICEF, 2012). 6 Figure 2. Net attendance rates for 15-17 year olds by income quintile Source: own estimations based on EPH. Note: quintiles of the distribution of per capita family income. But if the 2006 National Education Law had no impact on attendance rates for those aged 15-17 three years after its implementation, what is driving the increase as of 2010 shown in Figure 1? In this paper we claim that the Asignación Universal por Hijo (Universal Child Allowance, AUH) program, implemented in Argentina in late 2009, is responsible for encouraging children aged 15 through 17, especially poor children, to stay at (or return to) school. In fact, Figure 2 shows that the improvement of upper-secondary net attendance rates after 2009 was driven by the most vulnerable children, i.e. the target group of the AUH. Net attendance rates for youths aged 15 to 17 in the first quintile of the income distribution increased eight percentage points in the last decade: almost three percentage points between 2004 and 2009 (from 72.8% to 74.6%) but more than five percentage points between 2009 and 2014 (from 75% to 80.5%). Net attendance rates for those in the top quintiles have remained mostly unchanged over the last decade. 5 The 2005 Education Funding Law 24,075 (Ley de Financiamiento Educativo) introduced a gradual expansion of the educational budget, with the aim of reaching 6% of GDP by 2010. This implied an increase in per-student expenditure in Argentina, but the country lacked improvements in terms of the efficiency of this investment, in particular the pedagogical and organizational transformations to facilitate the improvement of education results (Auguste, 2012). 6 For instance, the Plan Nacional de Educación Secundaria Obligatoria (2009-2011 and 2012-2016). 5 0 5 5 6 0 6 5 7 0 7 5 8 0 8 5 9 0 9 5 1 0 0 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 1st Quintile 2nd Quintile 3rd Quintile 4th Quintile 5th Quintile Education Policy Analysis Archives Vol. 25 No. 76 8 The AUH Program and the Incentives to School Attendance The AUH was launched in November 2009 and represents a massive conditional cash transfer (CCT) program that focuses on children under 18 years old living in poor and informal households. It was designed to extend the social protection network in Argentina, which used to be tied to the formal employment sector, to the more vulnerable groups of the population. The magnitude of the benefit as well as the expansion in the number of beneficiaries have no precedents in the Argentinian social policy, formerly characterized by small scale and targeted programs. The AUH awards a monetary transfer to households with children where neither parent is registered in the formal sector. This includes inactive, unemployed or informal workers earning less than the minimum wage.7 Each beneficiary household can perceive a transfer per child under 18 years old up to a maximum of five dependent children.8 Currently, more than 3.5 million children and youths benefit from this program, representing almost 29% of all individuals under 18 years old and approximately 15% of total households in the country (ANSES, 2014). Regarding its annual budget, the AUH is one of the largest CCT programs in Latin America, with resources representing almost 0.8% of the country’s GDP (Stampini & Tornarolli, 2013). CCT programs may impact on school enrollment and attendance by relaxing family’s budget constraints but also through the conditions they impose. Typically, the requirements associated to CCT programs involve enrollment and actual attendance at school and health controls. The rationale of these conditions is to redirect household consumption to the formation of human capital in order to break the intergenerational transmission of poverty. For instance, as education may be regarded as a normal good its consumption could increase with household income. The conditions—or co-responsibilities—set an additional incentive to bias this increase in consumption towards investment in education. In particular, the AUH imposes sanitary and educational conditions in terms of periodical health controls and vaccination for children under 5 and pregnant women, and school attendance at all compulsory levels from ages 5 through 18. For this purpose, the program sets a particular payment mechanism: 80% of the subsidy is automatically received by beneficiary families on a monthly basis, and the remaining 20% is paid annually, once compliance with the conditions is proven. 9 According to the regulations, noncompliance with the conditions implies not only the 20% is not perceived but also the termination of the benefits which implies the loss of the future transfers until the child turns 18. Since the AUH was launched as a permanent program with a wide support of all political parties, the transfers should be perceived as permanent income and the expected present value of the transfers should be large, thus reinforcing the commitment of beneficiaries with conditionalities. 7 It is important to note, however, that monitoring this condition is not feasible in practice. This implies that informal workers earning more than the minimum wage could become beneficiaries. Nevertheless, as shown later on, both quantitative and qualitative evidence suggest that these situations are scarce, probably due to social responsibility or stigma. 8 Transfers for disabled children have no age limit. 9 Concerning the condition on school attendance, the program originally required that the child must be enrolled in a public school. This clause, however, was never made effective given the large public opposition that claimed for a considerable fraction of vulnerable children who attends publicly subsidized private schools. In fact, 16% of all primary school students belonging to the first two quintiles of the equivalized income distribution were attending a private school in 2010. The corresponding figure for secondary school students is 14% (SEDLAC, 2015). Explaining the rise in secondary school attendance rate in Argentina 9 In practice, however, the AUH does not take such a hard line. The permanent exclusion of beneficiaries for noncompliance may be morally questionable. Since the program aims to target the more vulnerable groups of the population, noncompliance may be a manifestation of some kind of obstacle that the family cannot overcome. Consequently, in practice, failure to comply with the conditions of the AUH implies the loss of the accumulated 20% and the blockage of monthly payments, but usually not the permanent suspension of the benefit. The child may rejoin the program by demonstrating that he or she is enrolled in the current school year. Of course, these softer rules may also imply lower incentives for compliance. The amount of the AUH transfer has been modified several times to cope with inflation.10 As of June 2014, the monthly transfer for each child – i.e. 80% of the total transfer– was ARS 515, which represented almost 15% of the minimum legal wage in Argentina. For a typical poor family with three children, this implied an almost 30% increase of total monthly family income. The remaining 20% amounted to ARS 1,400 per child per year, i.e. 62% of total family income for the same typical family and almost 100% of the minimum legal wage. According to the literature on the impact of CCT programs, the effects on ‘access to school’ indicators such as enrollment and attendance are usually positive (Cecchini, 2014; Fiszbein et al., 2009), even though the size of the effect varies with other factors: it is larger for groups with low attendance rates, among the most vulnerable families and in programs with more generous transfers (Saavedra & García, 2012). Besides these general findings, some particular results are worth noticing. Typically, the size of the effect is larger in the secondary school level than in the primary level. For instance, both the Oportunidades program in Mexico (formerly known as PROGRESA) and Familias en Acción in Colombia significantly contributed to increase attendance rates, especially among secondary school children (Attanasio et al., 2008; De Brauw and Hoddinott, 2008; Schulz, 2004). Also, even when focusing on secondary education, the size of the effect exhibits considerable variation: from a two-percentage-point increase in the case of Ingreso Ciudadano in Uruguay to a 12- percentage-point increase in the case of Oportunidades in Mexico and Bolsa Escola in Brazil (Saavedra & García, 2012).11 In summary, even though the impact differs across programs and population groups, in general CCT programs improve the so-called ‘intermediate objectives’: better access to school, higher enrollment rates and higher attendance (Bastagli, 2008; Cecchini, 2014 ). Given this evidence and the importance of the AUH—both in terms of coverage and generosity of the benefits—it is likely that it contributed to the improvement of attendance rates documented the previous section, which took place precisely after the program’s inception in late 2009. However, only the empirical evidence can reveal whether this potential effect of the AUH on school attendance is significant and quantitatively relevant. Evidence of the impact of the AUH on education results is still scarce. Among a large set of wellbeing indicators, Paz and Golovanevsky (2014) find large and positive effects in attendance rates—around seven percentage points—of the AUH for eligible children aged 13-17 when comparing the years 2009 and 2010 through a difference-in-difference methodology. Jimenez and Jimenez (2016) apply a Propensity Score Matching approach to the 2012 National Expenditure Household Survey (ENGHo) and find that the AUH reduced the dropout rate among teenagers. In a recent working paper based on aggregate data from administrative sources, Cigliutti et al. (2015) 10 The nominal monthly benefit per child, initially set at ARS $180, has increased on average more than 20% per year and hence its real value has remained relatively constant since 2009 (Garganta et al., 2016). 11 Additional evidence from the Mexican PROGRESA/Oportunidades, the oldest and most studied program in Latin America, shows also a significant reduction in dropouts (SEDESOL, 2008), a fall of the gender gap in secondary enrollment (Parker, 2003) and an increase in indigenous children attendance (Escobar & De la Rocha, 2008). Education Policy Analysis Archives Vol. 25 No. 76 10 find that secondary gross enrollment rates in Argentina rose by 2.25 percentage points due to the AUH compared to a synthetic control that consists of a linear combination of other Latin American countries.12 The present work provides new evidence regarding the impact of the AUH on eligible’s secondary school attendance. By using micro-data and following a difference-in-difference approach we extend the period of analysis to cover six years before and five years after the AUH implementation. Furthermore, we zoom into the group aged 15-17 which allows for relating our findings to the extension of mandatory schooling while deepening the analysis of the nature of the effect by exploring heterogeneities across different sub-groups. In particular, and based on the international evidence discussed above, we expect a larger impact on school attendance for older children from the most vulnerable households, whose initial attendance rates are the lowest. We also expect the effect to increase as the number of siblings grows, since the household’s total transfer raise with the number of children. Data and Empirical Strategy The AUH was neither randomly assigned nor accompanied by a publicly available comprehensive dataset that may allow for follow-ups of the beneficiary population. The absence of these features greatly determines both the data and the empirical strategy for assessing the program’s impact on any outcome. We use microdata from the Permanent National Household Survey (EPH) carried out by the Argentinian national statistical office (INDEC). The EPH gathers data on demographic, education, income and employment issues and covers 31 large urban conglomerates, representing 62% of the total population of the country. We focus on the 2004-2014 decade. The pre-intervention period (before) includes years 2004 through 2009—the AUH was launched in November 2009—while the post-intervention period (after) covers years 2010 through 2014. Unfortunately, the EPH does not include questions that allow us to identify AUH beneficiaries. Consequently, we define the ‘treatment’ and ‘control’ groups based on AUH eligibility, and thus perform an intention-to-treat analysis. Given that the proportion of eligible households not participating in the program is small, the take up seems to be random and there are virtually no beneficiaries in formal households since the registration system of formal workers implies the automatic exclusion from the program (Garganta et al., 2016), this approach likely provides a lower- bound estimate of the actual effect of the AUH on the beneficiaries.13 Our sample includes children aged 15-17, i.e. in the upper secondary age range. We aim at determining if the child is a potential beneficiary of the program by checking whether he/she meets the AUH eligibility criteria. Particularly, we define the treatment and control groups based on children’s eligibility according to their parents’ labor status. A child is classified as belonging to the treatment group whenever his/her parents are either inactive, unemployed, informal or self- employed workers. Because of a special regulation, children whose parents are registered employees working in the domestic service are also eligible for the AUH and hence are included in the treatment group.14 As for the control group, it includes all children aged 15-17 for whom at least one of their parents is employed in the formal sector. As an additional requirement for eligibility, the AUH imposes that earnings are below the minimum legal wage. Even though this condition is not verifiable for informal workers, qualitative and quantitative evidence suggests that middle and high-income informal workers opt out of the 12 D’Elia et al. (2014) provide evidence of the AUH impact on education quality indicators. 13 See for instance Ravallion (2008) and Duflo et al. (2006). 14 Special Social Security Scheme for Domestic Service Employees (Law 25,239, Title XVIII). Explaining the rise in secondary school attendance rate in Argentina 11 program due to social responsibility and stigma, and hence the inclusion error is small.15 Therefore, we further restrict the sample to only include children from poor households, defined as those in the first four deciles of the per capita income distribution.16 In order to estimate the intention-to-treat impact of the AUH on secondary school attendance of eligible children we follow a difference-in-difference methodology by comparing the differences in the probability of secondary school attendance of the treatment and control groups, before and after the inception of the program. The identification assumptions are that secondary attendance rates of treatment and control groups would have evolved similarly in the absence of the program and that there was no other contemporaneous event to the implementation of the AUH that could have caused differences in the evolution of school attendance between the treatment and control groups. The latter does not appear to be a strong assumption considering no major initiatives affecting educational outcomes took place in 2009 (infrastructure expansion, teacher’s training, school meals, etc.). Regarding the first assumption, it cannot be proven but we provide evidence in its favor in the next section. As for the difference-in-difference model, we use the standard linear specification in equation (1). 𝐴𝑡𝑡𝑒𝑛𝑑𝑠𝑖 =∝ +𝛽1𝑇𝑟𝑒𝑎𝑡𝑖 + 𝛽2𝐴𝑓𝑡𝑒𝑟𝑖 + 𝛾(𝑇𝑟𝑒𝑎𝑡𝑖 . 𝐴𝑓𝑡𝑒𝑟𝑖 )+𝜃 𝑋𝑖 + 𝑢𝑖 (1) The output variable Attends is a binary indicator that takes the value 1 for children attending secondary school and 0 otherwise; 17 Treat is an indicator variable for the treatment group; After tags years after the AUH implementation (2010-2014), and 𝑋 includes a set of child and household level controls (child’s gender, age and squared age; head of household’s gender, age, squared age, educational level and employment status) as well as other household characteristics (household size, per capita income, single parent household, female headed household, number of children under 18). We also control for time (year and quarter) and regional fixed effects, as well as for regional trends.18 If the unobserved characteristics that remain after adding all these controls do not have a differential impact on attendance between both groups before and after the implementation of the AUH, we may claim that the 𝛾 parameter represents the causal effect of the program (Angrist & Pischke, 2009). 15 From the experience of public officials in charge of the registration to the AUH, non-poor individuals –yet not belonging to the formal sector- tend to opt out of the program either by not even starting the procedure or by not complaining when they are suspended from the benefit following audits (Pautassi et al., 2013). Evidence from the last National Consumption Survey (ENGHo 2012) points in the same direction: very few children belonging to the upper income deciles—less than 2% in the two top deciles—receive benefits from the AUH (Gasparini & Cruces, 2015). 16 Results are robust to other income measures as well as other cut-offs. Moreover, since our data comes from the national household survey (EPH) and not from an evaluation survey of the program, it is very unlikely that people misreport incomes or any other variable in order to affect the probability of becoming eligible to the AUH program. 17 Unfortunately, even though the EPH includes information on the education level being attended, it does not inform the specific school year. 18 We use data for the first semester of each year by combining EPH’s samples from the first two quarters and control for quarter fixed effects. Education Policy Analysis Archives Vol. 25 No. 76 12 Results Table 2 shows average net attendance rates for treatment and control groups before and after the inception of the AUH. Even though attendance rose for both groups, the increase was considerably larger among eligible children: 5.1 percentage points as compared to 1.9 for the control group. This preliminary unconditional evidence suggests that the AUH may have had the effect of rising secondary school attendance of eligible children aged 15-17 by 3.2 percentage points. Table 2 Net secondary school attendance rates. Children between 15 and 17 years old Treatment (i) Control (ii) (i)-(ii) Before AUH 75.1 87.0 -11.9 After AUH 80.2 88.9 -8.7 Difference (After-Before) 5.1 1.9 3.2 Source: own estimations based on Encuesta Permanente de Hogares. Note: Treatment includes children whose parents are inactive, unemployed, informal or self-employed workers (or registered employees working in domestic service). Control includes all children for whom at least one of their parents is employed in the formal sector. Before AUH includes years 2004-2009; After AUH includes years 2010-2014. It is worth noting, however, that given the very nature of the program—non-random assignment –, treatment and control groups differ by construction. Table 3 shows that even though the two groups share on average some features (gender, age, household’s size), potential AUH beneficiaries belong to poorer households and exhibit a larger proportion of single-parent and female headed households where the head of household has lower educational attainment and is more likely to be unemployed, both in pre and post-intervention periods. Table 3 Descriptive statistics. Children between 15 and 17 years old Variables Before After Treatment Control Diff. P-v Treatment Control Diff. P-v Child Male 51.5 51.0 0.5 0.6 50.1 51.4 -1.3 0.9 Age 15.9 15.9 0.0 0.5 16.0 15.9 0.1 0.0 Head of HH Single 34.7 14.4 20.3 0.0 36.9 16.6 20.3 0.0 Female 36.6 18.2 18.4 0.0 42.0 22.7 19.3 0.0 Age 46.4 45.4 1.0 0.0 46.1 45.4 0.7 0.0 Education years 7.9 9.0 -1.1 0.0 8.4 9.5 -1.1 0.0 Employed 73.5 89.9 -16.4 0.0 71.3 89.4 -18.1 0.0 HH HH Size 5.8 5.8 0.0 0.4 5.7 5.7 0.0 0.1 # of Children 3.2 3.1 0.1 0.0 3.1 3.0 0.1 0.0 Per capita income 184.3 285.4 -101.1 0.0 741.7 1012.5 -270.8 0.0 Observations 12,466 6,363 10,002 6,171 Source: own estimations based on Encuesta Permanente de Hogares. Note: Treatment includes children whose parents are inactive, unemployed, informal or self-employed workers (or registered employees working in domestic service). Control includes children for whom at least one of their parents is employed in the formal sector. Before AUH includes years 2004-2009; After AUH includes years 2010-2014. # of Children is the total number of children under 18 living in the household. HH stands for household. Explaining the rise in secondary school attendance rate in Argentina 13 In fact, as Figure 3 shows, treatment and control groups differ in their school attendance rates prior to the program, which is in part due to those differences in characteristics. Nevertheless, albeit attendance rates levels differ before the inception of the AUH, the time patterns are similar. This is confirmed by a pre-program common trends test: we do not find enough evidence to reject the null hypothesis that the pre-treatment trends were equal, thus reinforcing the confidence in our identification assumption.19 However, since 2010, just after the AUH implementation, the gap in school attendance between groups started to shrink because the attendance rate of eligible children grew faster than that of the control group. Figure 3. Net attendance rates for 15-17 year olds. Treatment and control groups Source: own estimations based on Encuesta Permanente de Hogares. Note: Treatment Group includes children whose parents are either inactive, unemployed, informal or self-employed workers (or are registered employees working in the domestic service). Control Group includes children for whom at least one of their parents is employed in the formal sector. Children in both groups are aged 15 through 17 and belong to the first four deciles of the per capita family income distribution. We now assess whether this result holds in a multivariate difference-in-difference framework and is robust to several types of controls. Table 4 shows the results of estimating the linear model of school attendance in equation 1. Models 1, 2 and 3 in the table progressively control for child’s and head of household’s characteristics (child’s gender, age and squared age; head of household’s gender, age, squared age, educational level, employment status), other household features (household size, per capita income, single-parent household, female headed household, number of children under 18), region and time fixed effects (year and quarter), as well as regional trends. The coefficient of the 19 We run a model of our outcome of interest (attendance) on a constant, the treatment dummy, year dummies and the interactions between these latter variables including only pre-intervention years. We then apply an F test in which the null hypothesis (Ho) states that all the coefficients for the interaction terms are jointly equal to zero. We find no evidence to reject the null: Ho: F(5, 18,817)=0.47, Prob>F=0.80. We then run a new model that includes both pre and post-program years. The null hypothesis is now easily rejected: Ho: F(10, 34,980)=2.19, Prob>F=0.015. AUH 6 0 7 0 8 0 9 0 1 0 0 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Treatment group Control Group Education Policy Analysis Archives Vol. 25 No. 76 14 interaction term is positive and statistically significant across specifications, suggesting a positive impact of the AUH on school attendance of eligible children aged 15-17 of almost four percentage points.20 Table 4 Probability of attending secondary school. Children between 15 and 17 years old (1) (2) (3) Treatment*After 0.0320*** 0.0392*** 0.0388*** (0.00817) (0.00890) (0.00885) Treatment -0.119*** -0.0771*** -0.0757*** (0.00728) (0.00623) (0.00622) After 0.0195*** 0.000711 0.0309 (0.00655) (0.00700) (0.0433) Child and head of HH characteristics Yes Yes Yes Other HH Characteristics No Yes Yes Regional and Time Dummies, Regional Trends No No Yes Observations 35,002 35,002 35,002 Source: own estimations based on Encuesta Permanente de Hogares. Note: OLS estimations. Dependent binary variable: Attends, equals 1 if the child is 15-17 years old and attends secondary level; Treatment equals 1 for eligible children and 0 for non-eligible children; After equals 1 in the period 2010-2014 and 0 for the period 2004-2009; child´s and/or head of household´s characteristics (child’s gender, age and squared age, head of household’s gender, age, squared age, educational level and employment status), other household characteristics (household size, per capita income, single parent household, female headed household, number of children under 18), region fixed effects (6 regions), time fixed effects (year and quarter) and regional time trends. Clustered robust standard errors in parenthesis; * p<0.10, ** p<0.05, *** p<0.01. These results thus support our hypothesis that the AUH did contribute to the improvement of attendance rates for children in the upper-secondary age range documented in the second section of this paper. Moreover, they inform that the size of the effect is certainly non-trivial. According to our estimates, the 3.9 percentage-point impact in secondary school net attendance implies that the AUH helped around 20,000 eligible children aged 15-17 to stay at secondary school over the period 2010-2014. In terms of education gaps, it represents a 20% closure of the net attendance rate gap between the treatment group and those belonging to the richest quintile. Moreover, compared to other Latin American CCT programs, the impact we find for the AUH is between the two- percentage-point effects of the Brazilian Bolsa Escola and the Uruguayan Ingreso Ciudadano, and the 12-percentage-point effects of Familias en Acción in Colombia and Oportunidades in Mexico (Saavedra & García, 2012). 21 20 The result is also robust to a nonlinear specification of the probability model. For instance, from rerunning the model in column 3 in Table 4 using a Probit model we find similar effects to those obtained from the linear model. The marginal effect of the interaction term evaluated on the sample average of the treated is 2.6 percentage points, with an associated p-value of 0.001. The complete estimation results from the Probit model are available upon request. 21 Some additional clarifications must be made in order to assess a fair comparison. Firstly, the average baseline of secondary school attendance in most Latin American countries was considerably lower than that Explaining the rise in secondary school attendance rate in Argentina 15 False Experiments We perform a series of counterfactual experiments or placebo exercises to gain more confidence in the validity of the identification assumption. In this regard, we run again the model in column 3 from Table 4 but using only pre-treatment observations, and pretending that the program took place in any year previous to 2009—the actual implementation date of the AUH. Table 5 shows the results for five alternative fake dates: 2004, 2005, 2006, 2007 and 2008. In all cases the coefficient accompanying the interaction term is small and not statistically significant, meaning that only after 2009 some event caused a differential shift on the attendance rates for the treatment group, but clearly not before. Table 5 Probability of attending secondary school. Placebo regressions Intervention in 2004 2005 2006 2007 2008 Treatment*After 0.0192 0.0252 0.0179 0.0157 0.0155 (0.0184) (0.0167) (0.0140) (0.0128) (0.0151) Treatment -0.0856*** -0.0872*** -0.0796*** -0.0751*** -0.0719*** (0.0166) (0.0147) (0.0105) (0.00831) (0.00753) After -0.0525** 0.0437 0.0489 0.0507 0.0759* (0.0249) (0.0392) (0.0393) (0.0386) (0.0421) Child and head of HH characteristics Yes Yes Yes Yes Yes Other HH Characteristics Yes Yes Yes Yes Yes Regional and Time Dummies, Regional Trends Yes Yes Yes Yes Yes Observations 18,829 18,829 18,829 18,829 18,829 Source: own estimations based on Encuesta Permanente de Hogares. Note: OLS estimations. Dependent binary variable: Attends, equals 1 if the child attends upper secondary level; Treatment equals 1 for eligible children and 0 for non-eligible children; After is defined ad-hoc for each year (for example in 2006 it equals 0 in the period 2004 to 2006 and 1 in the period 2007-2009). For a description of control variables included, refer to Table 4. Clustered robust standard errors in parenthesis; * p<0.10, ** p<0.05, *** p<0.01. Alternative Definitions of the Pre- and Post-intervention Periods The National Education Law of 2006 extended compulsory schooling for children aged 15- 17. If this legislation altered schooling incentives differently for the treatment and control groups, the effects we find cannot be adjudicated solely to the AUH. As discussed above, we do not believe this is the case because neither the law nor accompanying policies had enforcement mechanisms embedded in their design, and thus it is unclear through which channels the law may have affected school attendance. Moreover, the results of the placebo experiment with 2006 as the false intervention date—column 3 in Table 5—provide evidence against this possibility. of Argentina. Furthermore, we focus on upper-secondary attendance rates while the evidence presented above corresponds to the complete secondary school level. Education Policy Analysis Archives Vol. 25 No. 76 16 To reinforce this point we additionally assess the AUH impact on secondary attendance by establishing an alternative shorter pre-intervention period: from 2007 to 2009, rather than 2004 to 2009, i.e. we only consider post-law years. Column 1 in Table 6 shows the original results—the same results reported in Table 4, column 3—while column 2 presents the estimated results when restricting the sample to years 2007-2014 and defining 2007-2009 as the pre-intervention period. Coefficients are quite similar in terms of size and statistical significance, reinforcing the hypothesis that it was the AUH in 2009 that caused the increase in attendance rates of poor children living in informal households. Table 6 Probability of attending secondary school. Alternative pre and post-intervention periods Pre-Intervention Period 2004-2009 2007-2009 2007-2009 Post-Intervention Period 2010-2014 2010-2014 2010-2011 Treatment*After 0.0388*** 0.0328*** 0.0224* (0.00885) (0.00975) (0.0112) Treatment -0.0757*** -0.0707*** -0.0683*** (0.00622) (0.00831) (0.00911) After 0.0309 -0.100*** -0.0268** (0.0433) (0.0358) (0.0127) Child and head of HH characteristics Yes Yes Yes Other HH Characteristics Yes Yes Yes Regional and Time Dummies; Regional Trends Yes Yes Yes Observations 35,002 27,035 17,383 Source: own estimations based on Encuesta Permanente de Hogares. Note: OLS estimations. Dependent binary variable: Attends, equals 1 if the child attends upper secondary level; Treatment equals 1 for eligible children and 0 for non-eligible children; After is defined ad-hoc in each model (column 1: it equals 1 in the period 2010-2014 and 0 for the period 2004-2009; column 2: it equals 1 in the period 2010-2014 and 0 for the period 2007-2009; column 3: it equals 1 in the period 2010-2011 and 0 for the period 2007-2009). For a description of control variables included, refer to Table 4. Clustered robust standard errors in parenthesis; * p<0.10, ** p<0.05, *** p<0.01. We also consider another alternative sample using 2007-2009 as the pre-intervention period and restricting the post-intervention period to years 2010-2011. With this exercise we intend to address concerns of a too large post-intervention period in which contamination from other sources could arise. Column 3 in Table 6 shows that in this case the interaction coefficient is 2.2 percentage points, still significant but smaller than the coefficient we obtain when taking the entire post- intervention period. That is, even when focusing in just a couple years before and after its inception, the AUH has a significant positive effect on attendance rates of eligible children aged 15 to 17. However, the impact increases over time possibly as a result of the expansion of the program and also due to the cumulative effect of increased attendance at earlier ages, i.e. the increased attendance rates at earlier ages would push up attendance of older children a few years later. Alternative Samples Since the EPH does not include information to identify AUH beneficiaries we relied on children’s eligibility based on their parents’ labor status. However, some limitations of the survey Explaining the rise in secondary school attendance rate in Argentina 17 may lead to classification errors. To start with, we do not have information on one or both parents when they do not live within the household.22 Furthermore, even if parents live with their child it is not always straightforward to identify this relationship given the fact that the EPH collects information on the family linkage of each household member only in terms of the head of household.23 To assess the extent to which these limitations may affect results, we define three alternative nested samples that account for different possible situations: (i) a first sample that only contains those children for whom both parents live in the household; (ii) an alternative larger sample that includes children for whom at least one parent is present; and finally (iii) one that also incorporates those children living in households where neither parent is present. Considering our universe is composed by all children aged 15-17 belonging to the first four income deciles, then sample (i) represents 64.4% of that target population, sample (ii) adds up a considerable fraction of children leading to a total coverage of 94.1%, while sample (iii), by construction, holds the total universe. In all three samples, whenever more than one adult could be identified as the mother or father of the child, the child was only considered eligible if all of the ‘potential’ parents met the eligibility conditions.24 Table 7 Probability of attending secondary school. Alternative samples Sample (i) (ii) (iii) Treatment*After 0.0358*** 0.0388*** 0.0376*** (0.00984) (0.00885) (0.00865) Treatment -0.0761*** -0.0757*** -0.0792*** (0.00750) (0.00622) (0.00594) After 0.0287 0.0309 -0.00799 (0.0459) (0.0433) (0.0134) Child and head of HH characteristics Yes Yes Yes Other HH Characteristics Yes Yes Yes Regional and Time Dummies, Regional Trends Yes Yes Yes Observations 23,953 35,002 37,207 Source: own estimations based on Encuesta Permanente de Hogares. Note: sample (i) includes children aged 15-17 for whom both parents live in the household; sample (ii) includes children aged 15-17 for whom at least one parent is present; sample (iii) includes all children aged 15-17, irrespective of whether both, one or neither parent in present in the household. See Table 4 for a description of the variables included. Clustered robust standard errors in parenthesis; * p<0.10, ** p<0.05, *** p<0.01. 22 The latter generally includes households where grandparents are in charge of their grandchildren. 23 For instance, suppose a family is composed by the head of household, two of his daughters, two sons in law and a grandson between 15 and 17 years old. In such a case we would not be able to identify who the father and mother of the child are. 24 In the example set in the previous note, this would imply that both daughters and both sons-in-law should meet the requirements. These cases, however, only represented 0.8% of sample (i) and 1.8% of sample (ii). Education Policy Analysis Archives Vol. 25 No. 76 18 Table 7 shows that the estimated effects of the program are not altered when using these alternative samples, neither in magnitude nor in terms of statistical significance. Given the robustness of the main result to different samples, we choose to conduct the analysis on the basis of sample (ii). Indeed, all the results shown previously relied on this last group of children. The choice is grounded on conceptual reasons. On the one hand, it extends sample (i) by including many single- parent households, mostly female headed households, where poverty rates are usually higher and are thus possibly more prone to belong to the treatment group. On the other hand, sample (ii) excludes those children for whom we have no information on neither of their parents working conditions— sample (iii). The chosen sample, of course, suffers from the risk of including in the treatment group children that should belong to the control group: when the parent living with the children meets the program’s eligibility conditions but the parent not living within the household does not. Nevertheless, even making very pessimistic assumptions, we estimate that only 9% of sample (ii) could be wrongly classified in the treatment group.25 Heterogeneous Effects Our estimates show that the AUH increased net secondary attendance rates for those eligible children aged 15 to 17 years old by almost four percentage points, but heterogeneities may be hidden behind this average effect. In this section we explore whether the impact of AUH on attendance rates varies across groups. Firstly, we look for heterogeneous effects by age and gender of children. Secondly, we assess whether the impact is related to household characteristics: number of children and education level of the head of household. Heterogeneities by Age Table 8 shows that the effect varies considerably across age groups. Compared to the almost 4-percentage-point increase for the group aged 15-17, the effect is only 0.8 for the group aged 12-14 —lower-secondary age range. For children aged 6-11 —primary school age— the effect is even smaller but still significant (0.4 percentage points) while for the youths between 18 and 20 years old the estimated effect of the AUH is not statistically significant. The latter result is consistent with the fact that individuals older than 18 years old are not eligible for the program, so no effect of the AUH is expected in terms of their schooling. Regarding the age groups covered by the program (6 to 11, 12 to 14 and 15 to 17), the results are consistent with the existing international evidence on the impact of CCT programs on schooling: the effect of the AUH is larger for higher levels of education, where baseline attendance rates are lower (Fiszbein et al., 2009; Saavedra & García, 2012). Indeed, even though the explicit cost of attending school may be similar at all educational levels, the opportunity costs certainly increase with age: older children may work in the labor market or allow for other adults in the household to do so by taking care of younger siblings or performing other household chores.26 Therefore, it is plausible that the economic incentives introduced by the AUH may have lower or even insignificant effects for younger school- aged children, whose educational decisions are less sensitive to economic changes, thus explaining the larger impact for the oldest eligible children. 25 This is based on the assumption that all non-present parents live and are recognized as such. Also, we assume that their formality rate is similar to that of parents living with their children—around 36%. 26 The legal minimum working age in Argentina is 16 years old (Ley de empleo infantil 26,390). Explaining the rise in secondary school attendance rate in Argentina 19 Table 8 Probability of attending secondary school. Heterogeneities by age range Age Range 6-11 12-14 15-17 18-20 Treatment*After 0.00422*** 0.00809** 0.0388*** 0.0170 (0.00153) (0.00315) (0.00885) (0.0151) Treatment -0.00383*** -0.0153*** -0.0757*** -0.0867*** (0.00113) (0.00256) (0.00622) (0.00941) After -0.0229* -0.00251 0.0309 -0.109 (0.0120) (0.0172) (0.0433) (0.0695) Child and head of HH characteristics Yes Yes Yes Yes Other HH Characteristics Yes Yes Yes Yes Regional and Time Dummies, Regional Trends Yes Yes Yes Yes Observations 69,332 34,904 35,002 28,792 Source: own estimations based on Encuesta Permanente de Hogares. Note: OLS estimations. Dependent binary variable: Attends, equals 1 if the child attends the corresponding level; Treatment equals 1 for eligible children and 0 for non-eligible children; After equals 1 in the period 2010-2014 and 0 for the period 2004-2009. For a description of the variables included, see Table 4. Clustered robust standard errors in parenthesis; * p<0.10, ** p<0.05, *** p<0.01. Heterogeneities by Gender Table 9 shows that the increase in attendance rates was mostly driven by improvements in boys’ attendance: the estimated impact for boys is above five percentage points while that of girls is below two percentage points and not statistically significant. Table 9 Probability of attending secondary school. Heterogeneities by gender Boys Girls Treatment*After 0.0583*** 0.0165 (0.0108) (0.0122) Treatment -0.100*** -0.0499*** (0.00813) (0.00731) After 0.0263 0.00823 (0.0393) (0.0696) Child and head of HH characteristics Yes Yes Other HH Characteristics Yes Yes Regional and Time Dummies, Regional Trends Yes Yes Observations 17,822 17,180 Source: own estimations based on Encuesta Permanente de Hogares. Note: for a description of variables, see Table 4. Clustered robust standard errors in parenthesis; * p<0.10, ** p<0.05, *** p<0.01. Education Policy Analysis Archives Vol. 25 No. 76 20 Once again, more than one mechanism may explain these results. As stated before, different baseline levels of attendance may be in part responsible. In fact, initial attendance rates were lower for boys: around 70% as compared to 80% for girls among the treatment group—the control group showed higher rates: 86% and 88%, respectively. Also, according to the literature, family decisions on girls’ schooling seem to be more tied to cultural factors which are less affected—at least in the short term—by changes in household income. For instance, previous evidence for Argentina (Sosa Escudero & Marchionni, 1999) suggests that girls’ attendance is rather inelastic as compared to boys’. Heterogeneities by Household Characteristics Table 10 shows the AUH effect on net school attendance according to the the number of children under 18 present. The impact is statistically significant for all groups, but it increases with the number of children. In particular, the effect for larger households almost doubles that of families with one or two children. Table 10 Probability of attending secondary school. Heterogeneities by number of children in the household Number of Children 1 or 2 3 or 4 5 or more Treatment*After 0.0267** 0.0326** 0.0506** (0.0130) (0.0143) (0.0218) Treatment -0.0612*** -0.0686*** -0.104*** (0.0106) (0.00872) (0.0127) After -0.0760* 0.0652 0.0535 (0.0390) (0.0451) (0.0700) Child and head of HH characteristics Yes Yes Yes Other HH Characteristics Yes Yes Yes Regional and Time Dummies, Regional Trends Yes Yes Yes Observations 13,799 14,301 6,902 Source: own estimations based on Encuesta Permanente de Hogares. Note: for a description of variables included, see Table 4. Clustered robust standard errors in parenthesis; * p<0.10, ** p<0.05, *** p<0.01. According to the existing evidence discussed in the third section of this paper, the size of the effect is usually larger in programs with more generous benefits (Saavedra & García, 2012). Since the AUH transfers are on a per-child basis up to a maximum of 5 dependent children, more eligible children in the household imply a rising cash benefit and a potentially stronger income effect. Therefore, this result suggests that larger families may show more commitment with the conditionalities of the program. Finally, table 11 explores whether the effect varies with household structure –two-parent or single-parent families– and with the education level of the head of household. The positive effect of AUH on school attendance is present in both two-parent and single-parent households, and the size of the effect is similar for the two family types. Moreover, we find that the average effect is mostly driven by children from families whose head of household has a low educational level. Children with Explaining the rise in secondary school attendance rate in Argentina 21 parents with low education, unemployed or in informal jobs are the most vulnerable, their attendance rates are the lowest and, therefore, the potential of the program to affect their schooling is greater. On the contrary, the effect for children whose head of household has a high educational level is smaller and not statistically significant. Table 11 Probability of attending secondary school. Heterogeneities by characteristics of the head of household Single- Parent Two- Parent Low Education High Education Treatment*After 0.0369** 0.0398*** 0.0360*** 0.0240 (0.0157) (0.0105) (0.0104) (0.0142) Treatment -0.0685*** -0.0765*** -0.0841*** -0.0435*** (0.00970) (0.00758) (0.00752) (0.0100) After -0.00686 -0.00582 0.121** 0.0269* (0.0300) (0.0128) (0.0549) (0.0153) Child and head of HH characteristics Yes Yes Yes Yes Other HH Characteristics Yes Yes Yes Yes Regional and Time Dummies, Regional Trends Yes Yes Yes Yes Observations 10,994 24,008 25,505 9,497 Source: own estimations based on Encuesta Permanente de Hogares. Note: OLS estimations. “Low Education” includes household which head has less than secondary school education, “High Education” refers to households where the head completed secondary education. For a description of the variables included, see Table 4. Clustered robust standard errors in parenthesis; * p<0.10, ** p<0.05, *** p<0.01. Concluding Remarks and Further Research Argentina has traditionally stood out in terms of educational outcomes among its Latin American counterparts. Schooling of older children, however, still shows room for improvement especially among the more vulnerable school-age children. Fortunately, during the last years a sizeable improvement in attendance rates for children aged 15 through 17 took place. Even though this could be potentially related to the 2006 National Education Law that made upper-secondary education compulsory, in this paper we claim that the rise in school attendance was not caused by this new legislation. One of the conceptual arguments is that the 2006 law was not accompanied by enforcement mechanisms or by other measures designed to encourage schooling. Instead, we show that the Asignación Universal por Hijo (Universal Child Allowance, AUH), a massive conditional cash transfer (CCT) program implemented in 2009 in Argentina, may be the main responsible for the rise in net attendance rates of upper-secondary school age children since 2010. The AUH aims at providing social protection for poor and informal households while encouraging investment in children’s human capital to break the intergenerational transmission of poverty. To this end, as in any typical CCT program, cash transfers are conditioned on compliance with health and schooling conditions. In particular, the AUH requires that children between 5 and 18 years old attend school at all compulsory levels in order to remain in the program. Education Policy Analysis Archives Vol. 25 No. 76 22 There are various reasons to believe that the incentives introduced by the AUH may have caused attendance rates to grow in Argentina. Arguments from the economic theory provide a first justification. There is an income effect associated to the money transfer, which is positive as long as children’s education may be regarded as a normal good, leading to more consumption of education. Moreover, the fact that transfers are conditional set an additional incentive to redirect family consumption to children’s education. A second justification comes from the vast empirical evidence on the positive effects of CCT programs on access to school indicators such as enrollment and attendance. Because of its design, these potential effects are also present in the AUH. However, only the empirical evidence can reveal whether its actual effect on school attendance is significant and quantitatively relevant. To this aim, we use a difference-in-difference strategy based on data from the Argentinian National Permanent Household Survey and estimate that the program accounts for a 3.9 percentage point increase in secondary school attendance among eligible children aged 15 through 17. This effect is robust to different specifications and a large set of robustness checks, supporting our hypothesis that the AUH did contribute to the improvement of attendance rates for children in the upper-secondary age range and informing that the size of the effect is certainly non-trivial. Also, we present evidence suggesting that this effect is not related to the expansion of compulsory education that took place in Argentina in 2006. Additionally, we find a positive effect on school attendance for younger children of primary and lower-secondary school age range, but the size of the effect is small (0.4 and 0.8 percentage points, respectively). This result is consistent with the evidence from other CCT programs in Latin America, where the impact is lower for children with high enrollment or attendance rates at the beginning of the program. We also find no effect of the AUH on school attendance for the group aged 18 and over, which is precisely the age limit for participating in the program. Moreover, we find that the positive impact of the AUH on attendance rates is not homogenous within the group of children aged 15 through 17: the effect seems to be driven particularly by boys and is higher for children from larger families where the head of household has low educational attainment. Again, these results are consistent with empirical evidence from other studies. For instance, Sosa Escudero and Marchionni (1999) suggest that boys’ school attendance in Argentina is more elastic to economic incentives than girls’. Also, evidence from other CCT programs find that the size of the impact grows with the generosity of the transfers (Fiszbein et al., 2009; Saavedra & García, 2012), which in the case of the AUH increases with the number of children up to a maximum of five dependent children. Therefore, more eligible children in the household imply larger benefits and a potentially stronger income effect. Our result on the higher impact in families whose head of household has low level of schooling is also consistent with evidence from the abovementioned studies that find larger effects for the most vulnerable population groups. Children with parents with low education, unemployed or in informal jobs are the most vulnerable, their attendance rates are the lowest and, therefore, the potential of the program to affect their schooling is greater. Further research should point in several directions. A first relevant issue would be to unravel which mechanisms within the AUH are responsible for the increase in attendance rates. The effect may be driven by the monthly benefit itself or by the conditionality, or both mechanisms could be operating simultaneously. A deep understanding of these alternative channels is indeed relevant in terms of improving the design of CCT programs. Secondly, it would be interesting to explore if the AUH has not only increased secondary school attendance among eligible children but also affected other educational results, such as intra-annual dropouts or secondary school completion rates. Thirdly, it would also be relevant to disentangle if this increase in attendance rates is matched by a Explaining the rise in secondary school attendance rate in Argentina 23 similar result in the employment realm. It could be expected that an increase in attendance rates may contribute to a reduction in labor participation among the 15-17 age group. It could also be the case, however, that those upper-secondary school aged children were not working in the labor market before the AUH, but in charge of household chores such as taking care of their siblings. In that case the AUH may be altering instead other members’ labor participation. Although household decision processes are certainly difficult to assess, exploring these hypothesis would shed light on the mechanisms that are at work and thus further refine the AUH’s design. Acknowledgements We are grateful to Leonardo Gasparini, Walter Sosa Escudero, Paula Razquin, Joaquín Coleff, Emamnuel Vázquez, Leopoldo Tornarolli, seminar participants at Universidad de San Andrés and Universidad Nacional de La Plata, and two anonymous referees for valuable comments and suggestions. The usual disclaimers apply. References Acemoglu, D., & Angrist, J. (2000). How Large are the Social Returns to Education? Evidence from Compulsory Schooling Laws. NBER Macroeconomics Annual, 15, 9-59. Agencia Nacional de Seguridad Social. (ANSES) (2014). La Asignación Universal por Hijo para la Protección Social. Boletín Cuatrimestral. I Cuatrimestre 2014. Observatorio de la Seguridad Social. Alzúa, M.L., Gasparini, L., & Haimovich, F. (2015). Educational Reform & Labor Market Outcomes: the Case of Argentina’s Ley Federal de Educación. Journal of Applied Economics, 18(1), 21-46. Angrist, J., & Krueger, A. (1991). Does Compulsory School Attendance Affect Schooling & Earnings? Quarterly Journal of Economic, 106(4), 979-1014. Angrist, J. & Pischke, J. (2009). Mostly Harmless Econometrics. An Empiricist’s Companion. Princeton University Press. Auguste, S. (2012). La calidad educativa en la Argentina. Documento de Trabajo 116, Fundación de Investigaciones Económicas Latinoamericanas, Buenos Aires. Attanasio, O., Fitzsimons, E., Gómez, A., & López, D. (2008). Child education & work choices in the presence of a conditional cash transfer programme in rural Colombia. (WP06/13) London, Institute for Fiscal Studies. Bastagli, F. (2008). Conditionality in public policy targeted to the poor: promoting resilience? Social Policy & Society, 8(1). Becker, G. S. (1975). Human Capital: A Theoretical & Empirical Analysis, with Special Reference to Education. 2nd Edition, The National Bureau of Economic Research. Boffi, S. (2013). Políticas de transferencias de ingresos. Su impacto en el mercado de trabajo. Experiencias en Latinoamérica y los casos de la Asignación Universal por Hijo y el Programa de Inclusión Previsional en Argentina. Publication No 16, Centro de Investigación en Trabajo, Distribución y Sociedad. Cecchini, S. (2014). Educación, Programas de Transferencias Condicionadas y Protección Social en América Latina y el Caribe. In M. Feijóo & M. Poggi (Eds.) Educación y Políticas Sociales. Sinergias para la Inclusión. IIPE-UNESCO. Education Policy Analysis Archives Vol. 25 No. 76 24 Cigliutti, M., Echeverri Gómez, M. , Golinsky, F., Gutiérrez, A., & Sorá, M.G. (2015). Conditional Cash Transfer Programs: Measuring the Impact on Education. Anales de la Asociación Argentina de Economía Política, L Reunión Anual, Salta, Argentina. De Brauw, A., & Hoddinott, J. (2008). Must conditional cash transfer programs be conditioned to be effective? The impact of conditioning transfers on school enrollment in Mexico. IFPRI, Discussion Paper, 00757. D’Elia, V., & Navarro, A.I. (2013). Universal Child Allowance & School Delay of Children in Argentina. Revista de Análisis Económico, 28(2), 63-90. Dirección Nacional de Información y Evaluación de Calidad Educativa. (DiNIECE), http://portales.educacion.gov.ar/diniece/ DiNIECE (2011). La Transformación del nivel secundario. (2006–2009). Temas de Educación. Boletín DiNIECE year 6, no. 9, Dirección Nacional de Información y Evaluación de la Calidad Educativa, Ministerio de Educación, Buenos Aires, Argentina. Duflo, E., Glennerster, R., & Kremer, M. (2006). Using Randomization in Development Economics Research: A Toolkit. NBER Technical Working Paper 333. Eckstein, Z., & Zilcha, I. (1994). The effects of compulsory schooling on growth, income distribution & welfare. Journal of Public Economics, 54(3), 339-359. Escobar, A., & De la Rocha, M. (2008). Girls, mothers & poverty reduction in Mexico: Evaluating Progresa-Oportunidades. In S. Razavi (Ed.), The Gendered Impacts of Liberalization: Towards embedded liberalism? New York: Routledge. Fiszbein, A., Schady, N., Ferreira, F. H.G., Grosh, M., Keleher, N., Olinto, P., & Skoufias, E. (2009). Conditional Cash Transfers: Reducing Present & Future Poverty. World Bank Publications, The World Bank, number 2597. Garganta, S., Gasparini, L., Marchionni, M., & Tappata, M. (2016). The Effects of Cash Transfers on Fertility: Evidence from Argentina. Population Research & Policy Review. First online: October 28, 2016. doi:10.1007/s11113-016-9417-x Gasparini, L., & Cruces, G. (2015). Pobreza y Políticas de Transferencias Monetarias en Argentina. Diagnóstico y alternativas. Mimeo, CEDLAS. Goldin, C., & Katz, L. (2008). Mass Secondary Schooling & the State: The Role of State Compusion in the Hight School Movement. NBER Working Paper 10075. Jiménez, M., & Jiménez, M. (2016). Efectos del programa Asignación Universal por Hijo en la deserción escolar adolescente. Cuadernos de Economía, 35(69), 709-752. Lleras-Muney, A. (2005). The Relationship Between Education & Adult Mortality in the United States. Review of Economic Studies, 72, 189-221. Marchionni, M., & Alejo, J. (2015). Funding Basic Education Throughout the Demographic Transition in Argentina (Chapter 8). In M. Gragnolati, R. Rofman, I. Apella, I. & S. Troiano, (Eds.). As time goes by in Argentina. Opportunities & Challenges of the Demographic Transition. Washington D.C.: World Bank Group. Oreopoulos, P. (2006a). The compelling effects of compulsory schooling: evidence from Canada. Canadian Journal of Economics, Canadian Economics Association, 39(1), 22-52. Oreopoulos, P. (2006b). Estimating Average & Local Average Treatment Effects of Education When Compulsory Schooling Laws Really Matter. The American Economic Review, 96(1), 152- 175. Parker, S. (2003). Evaluación del impacto de Oportunidades sobre la inscripción escolar: primaria, secundaria y media superior. Resultados de la Evaluación Externa del Programa de Desarrollo Humano Oportunidades 2002, Mexico D.F., Secretaría de Desarrollo Social (SEDESOL). http://portales.educacion.gov.ar/diniece/ Explaining the rise in secondary school attendance rate in Argentina 25 Pautassi, L., Arcidiácono, P., & Straschnoy, M. (2013). Asignación Universal por Hijo para la Protección Social de la Argentina. Entre la satisfacción de necesidades y el reconocimiento de derechos, CEPAL, Serie Políticas Sociales N°184. Paz, J., & Golovanevsky, L. (2014). Programa Asignación Universal por Hijo para la Protección Social. Un ejercicio de evaluación de sus efectos sobre los hogares. III Jornadas Nacionales sobre Estudios Regionales y Mercados de Trabajo, Jujuy, Argentina. Ravallion, M. (2008). Evaluating Anti-Poverty Programs. Handbook of Development Economics, 4. DOI: 10.1016/S1573-4471(07)04059-4 Ruiz, M. C., & Schoo, S. (2014). La obligatoriedad de la educación secundaria en América Latina. Convergencias y Divergencias en cinco países. Foro de Educación, 12(16), 71-98. Saavedra, J. E., & García, S. (2012). Impacts of conditional cash transfer programs on educational outcomes in developing countries: a meta-analysis", Labor & Population Working Paper Series, WR-921-1. Secretaría de Desarrollo Social (SEDESOL). (2008). Oportunidades, un programa de resultados. México, D.F. Socio-Economic Database for Latin America & the Caribbean (SEDLAC) (2017). Centro de Estudios Distributivos Laborales y Sociales (CEDLAS) & the World Bank, http://sedlac.econo.unlp.edu.ar/eng/ Sosa Escudero, W., & Marchionni, M. (1999). Household Structure, Gender & the Economic Determinants of School Attendance in Argentina. Anales de la Asociación Argentina de Economía Política, XXXIV Reunión Anual, Rosario. UNICEF. (2012). Acerca de la Obligatoriedad de la Escuela Secundaria en Argentina. Análisis de la Política Nacional. About the Authors María Edo Universidad de San Andrés & CONICET mariaedo@gmail.com María Edo is a postdoctoral scholar at the National Council of Scientific and Technical Research (CONICET)-Universidad de San Andrés (UdeSA) in Argentina and Assistant Professor at UdeSA. María holds a PhD in Economics from UdeSA. Her research interests include Poverty Measurement, Social Policies and Gender Economics. Mariana Marchionni CEDLAS-Universidad de La Plata & CONICET marchionni.mariana@gmail.com Mariana Marchionni is Senior Researcher of the Centro de Estudios Distributivos, Laborales y Sociales (CEDLAS) at Universidad Nacional de La Plata (UNLP), and Associate Researcher of the National Council of Scientific and Technical Research (CONICET). Mariana holds a PhD in Economics from UNLP and was a Postdoctoral Research Fellow at University of British Columbia during 2013. She is a full-time professor and the Director of the Master Program in Economics at UNLP. Her main research interests are in the areas of Economics of Education, Gender Economics, Demographic Economics, and Social Policies, and she has published numerous articles in academic journals, books and book chapters on these topics. http://sedlac.econo.unlp.edu.ar/eng/ mailto:mariaedo@gmail.com mailto:marchionni.mariana@gmail.com Education Policy Analysis Archives Vol. 25 No. 76 26 Santiago Garganta CEDLAS-Universidad de La Plata & CONICET sangarganta@hotmail.com Santiago Garganta is Senior Researcher of the Centro de Estudios Distributivos, Laborales y Sociales (CEDLAS) at Universidad Nacional de La Plata (UNLP) in Argentina. He has a PhD in Economics from UNLP and holds a postdoctoral fellowship from the National Council of Scientific and Technical Research (CONICET). His research work is focused on the fields of Labor Economics, Poverty and Inequality, and Social Protection Systems. education policy analysis archives Volume 25 Number 76 July 17, 2017 ISSN 1068-2341 Readers are free to copy, display, & distribute this article, as long as the work is attributed to the author(s) & Education Policy Analysis Archives, it is distributed for non- commercial purposes only, & no alteration or transformation is made in the work. More details of this Creative Commons license are available at http://creativecommons.org/licenses/by -nc- sa/3.0/. All other uses must be approved by the author(s) or EPAA. EPAA is published by the Mary Lou Fulton Institute & Graduate School of Education at Arizona State University Articles are indexed in CIRC (Clasificación Integrada de Revistas Científicas, Spain), DIALNET (Spain), Directory of Open Access Journals, EBSCO Education Research Complete, ERIC, Education Full Text (H.W. Wilson), QUALIS A1 (Brazil), SCImago Journal Rank; SCOPUS, SOCOLAR (China). Please send errata notes to Audrey Amrein-Beardsley at audrey.beardsley@asu.edu Join EPAA’s Facebook community at https://www.facebook.com/EPAAAAPE & Twitter feed @epaa_aape. mailto:sangarganta@hotmail.com http://www.doaj.org/ mailto:audrey.beardsley@asu.edu https://www.facebook.com/EPAAAAPE Explaining the rise in secondary school attendance rate in Argentina 27 education policy analysis archives editorial board Lead Editor: Audrey Amrein-Beardsley (Arizona State University) Editor Consultor: Gustavo E. Fischman (Arizona State University) Associate Editors: David Carlson, Margarita Jimenez-Silva, Eugene Judson, Mirka Koro-Ljungberg, Scott Marley, Jeanne M. Powers, Iveta Silova, Maria Teresa Tatto (Arizona State University) Cristina Alfaro San Diego State University Ronald Glass University of California, Santa Cruz R. Anthony Rolle University of Houston Gary Anderson New York University Jacob P. K. Gross University of Louisville A. G. Rud Washington State University Michael W. Apple University of Wisconsin, Madison Eric M. Haas WestEd Patricia Sánchez University of University of Texas, San Antonio Jeff Bale OISE, University of Toronto, Canada Julian Vasquez Heilig California State University, Sacramento Janelle Scott University of California, Berkeley Aaron Bevanot SUNY Albany Kimberly Kappler Hewitt University of North Carolina Greensboro Jack Schneider College of the Holy Cross David C. Berliner Arizona State University Aimee Howley Ohio University Noah Sobe Loyola University Henry Braun Boston College Steve Klees University of Maryland Nelly P. Stromquist University of Maryland Casey Cobb University of Connecticut Jaekyung Lee SUNY Buffalo Benjamin Superfine University of Illinois, Chicago Arnold Danzig San Jose State University Jessica Nina Lester Indiana University Sherman Dorn Arizona State University Linda Darling-Hammond Stanford University Amanda E. Lewis University of Illinois, Chicago Adai Tefera Virginia Commonwealth University Elizabeth H. DeBray University of Georgia Chad R. Lochmiller Indiana University Tina Trujillo University of California, Berkeley Chad d'Entremont Rennie Center for Education Research & Policy Christopher Lubienski University of Illinois, Urbana-Champaign Federico R. Waitoller University of Illinois, Chicago John Diamond University of Wisconsin, Madison Sarah Lubienski University of Illinois, Urbana-Champaign Larisa Warhol University of Connecticut Matthew Di Carlo Albert Shanker Institute William J. Mathis University of Colorado, Boulder John Weathers University of Colorado, Colorado Springs Michael J. Dumas University of California, Berkeley Michele S. Moses University of Colorado, Boulder Kevin Welner University of Colorado, Boulder Kathy Escamilla University of Colorado, Boulder Julianne Moss Deakin University, Australia Terrence G. Wiley Center for Applied Linguistics Melissa Lynn Freeman Adams State College Sharon Nichols University of Texas, San Antonio John Willinsky Stanford University Rachael Gabriel University of Connecticut Eric Parsons University of Missouri-Columbia Jennifer R. Wolgemuth University of South Florida Amy Garrett Dikkers University of North Carolina, Wilmington Susan L. Robertson Bristol University, UK Kyo Yamashiro Claremont Graduate University Gene V Glass Arizona State University Gloria M. Rodriguez University of California, Davis Education Policy Analysis Archives Vol. 25 No. 76 28 archivos analíticos de políticas educativas consejo editorial Editor Consultor: Gustavo E. Fischman (Arizona State University) Editores Asociados: Armando Alcántara Santuario (Universidad Nacional Autónoma de México), Jason Beech (Universidad de San Andrés), Angelica Buendia (Metropolitan Autonomous University, Mexico), Ezequiel Gomez Caride (Pontificia Universidad Católica Argentina), Antonio Luzon (Universidad de Granada), José Luis Ramírez (Universidad de Sonora, Mexico) Claudio Almonacid Universidad Metropolitana de Ciencias de la Educación, Chile Juan Carlos González Faraco Universidad de Huelva, España Miriam Rodríguez Vargas Universidad Autónoma de Tamaulipas, México Miguel Ángel Arias Ortega Universidad Autónoma de la Ciudad de México María Clemente Linuesa Universidad de Salamanca, España José Gregorio Rodríguez Universidad Nacional de Colombia, Colombia Xavier Besalú Costa Universitat de Girona, España Jaume Martínez Bonafé Universitat de València, España Mario Rueda Beltrán Instituto de Investigaciones sobre la Universidad y la Educación, UNAM, México Xavier Bonal Sarro Universidad Autónoma de Barcelona, España Alejandro Márquez Jiménez Instituto de Investigaciones sobre la Universidad y la Educación, UNAM, México José Luis San Fabián Maroto Universidad de Oviedo, España Antonio Bolívar Boitia Universidad de Granada, España María Guadalupe Olivier Tellez, Universidad Pedagógica Nacional, México Jurjo Torres Santomé, Universidad de la Coruña, España José Joaquín Brunner Universidad Diego Portales, Chile Miguel Pereyra Universidad de Granada, España Yengny Marisol Silva Laya Universidad Iberoamericana, México Damián Canales Sánchez Instituto Nacional para la Evaluación de la Educación, México Mónica Pini Universidad Nacional de San Martín, Argentina Juan Carlos Tedesco Universidad Nacional de San Martín, Argentina Gabriela de la Cruz Flores Universidad Nacional Autónoma de México Omar Orlando Pulido Chaves Instituto para la Investigación Educativa y el Desarrollo Pedagógico (IDEP) Ernesto Treviño Ronzón Universidad Veracruzana, México Marco Antonio Delgado Fuentes Universidad Iberoamericana, México José Luis Ramírez Romero Universidad Autónoma de Sonora, México Ernesto Treviño Villarreal Universidad Diego Portales Santiago, Chile Inés Dussel, DIE-CINVESTAV, México Paula Razquin Universidad de San Andrés, Argentina Antoni Verger Planells Universidad Autónoma de Barcelona, España Pedro Flores Crespo Universidad Iberoamericana, México José Ignacio Rivas Flores Universidad de Málaga, España Catalina Wainerman Universidad de San Andrés, Argentina Ana María García de Fanelli Centro de Estudios de Estado y Sociedad (CEDES) CONICET, Argentina Juan Carlos Yáñez Velazco Universidad de Colima, México javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/819') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/820') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/4276') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/1609') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/825') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/797') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/555') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/823') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/798') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/814') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/801') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/1617') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/2703') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/802') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/816') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/826') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/804') javascript:openRTWindow('http://epaa.asu.edu/ojs/about/editorialTeamBio/3264') Explaining the rise in secondary school attendance rate in Argentina 29 arquivos analíticos de políticas educativas conselho editorial Editor Consultor: Gustavo E. Fischman (Arizona State University) Editoras Associadas: Geovana Mendonça Lunardi Mendes (Universidade do Estado de Santa Catarina), Marcia Pletsch, Sandra Regina Sales (Universidade Federal Rural do Rio de Janeiro) Almerindo Afonso Universidade do Minho Portugal Alexandre Fernandez Vaz Universidade Federal de Santa Catarina, Brasil José Augusto Pacheco Universidade do Minho, Portugal Rosanna Maria Barros Sá Universidade do Algarve Portugal Regina Célia Linhares Hostins Universidade do Vale do Itajaí, Brasil Jane Paiva Universidade do Estado do Rio de Janeiro, Brasil Maria Helena Bonilla Universidade Federal da Bahia Brasil Alfredo Macedo Gomes Universidade Federal de Pernambuco Brasil Paulo Alberto Santos Vieira Universidade do Estado de Mato Grosso, Brasil Rosa Maria Bueno Fischer Universidade Federal do Rio Grande do Sul, Brasil Jefferson Mainardes Universidade Estadual de Ponta Grossa, Brasil Fabiany de Cássia Tavares Silva Universidade Federal do Mato Grosso do Sul, Brasil Alice Casimiro Lopes Universidade do Estado do Rio de Janeiro, Brasil Jader Janer Moreira Lopes Universidade Federal Fluminense e Universidade Federal de Juiz de Fora, Brasil António Teodoro Universidade Lusófona Portugal Suzana Feldens Schwertner Centro Universitário Univates Brasil Debora Nunes Universidade Federal do Rio Grande do Norte, Brasil Lílian do Valle Universidade do Estado do Rio de Janeiro, Brasil Flávia Miller Naethe Motta Universidade Federal Rural do Rio de Janeiro, Brasil Alda Junqueira Marin Pontifícia Universidade Católica de São Paulo, Brasil Alfredo Veiga-Neto Universidade Federal do Rio Grande do Sul, Brasil Dalila Andrade Oliveira Universidade Federal de Minas Gerais, Brasil