Journal of Public Economics 92 (2008) 1416–1440 www.elsevier.com/locate/econbase Giving children a better start: Preschool attendance and school-age profiles ☆ ⁎ Samuel Berlinski a,b, , Sebastian Galiani c, Marco Manacorda d,e,f a University College London, UK b Institute for Fiscal Studies, UK c Department of Economics, Washington University in St Louis, Campus Box 1208, St Louis, MO 63130-4899, USA d Queen Mary University of London, UK e CEP, London School of Economics, Houghton Street WC2A 2AE, London, UK f CEPR, USA Received 25 February 2007; received in revised form 23 October 2007; accepted 29 October 2007 Available online 23 December 2007 Abstract We study the effect of pre-primary education on children's subsequent school outcomes by exploiting a unique feature of the Uruguayan household survey (ECH) that collects retrospective information on preschool attendance in the context of a rapid expansion in the supply of pre-primary places. Using a within household estimator, we find small gains from preschool attendance at early ages that get magnified as children grow up. By age 15, treated children have accumulated 0.8 extra years of education and are 27 percentage points more likely to be in school compared to their untreated siblings. Instrumental variables estimates that attempt to control for non random selection of siblings into preschool lead to similar results. Pre-primary education appears as a successful and cost-effective policy to prevent early grade failure and its long lasting consequences in low income countries. © 2007 Elsevier B.V. All rights reserved. JEL classifications: I2; J1 Keywords: Preschool; Pre-primary education; Primary school performance 1. Introduction This paper estimates the effect of early exposure to pre-primary education on school stay-on rates and levels of completed education among individuals aged 7–15. We exploit a rather unique feature of the Uruguayan Encuesta Continua de Hogares (ECH) for the years 2001–2005 that collects retrospective information on the number of years of ☆ We are grateful to two anonymous referees, the editor Dennis Epple, Andres Peri, Emiliana Vegas, and especially Andrea Vigorito for many helpful comments and suggestions. Marco Manacorda gratefully acknowledges financial help from the Nuffield Foundation New Career Development Fellowship in the Social Sciences and ESRC grant no. 000-22-0131. Samuel Berlinski gratefully acknowledges financial help from the ESRC-DFID grant RES-167-25-0124. ⁎ Corresponding author. Department of Economics, University College London, Gower Street, London WC1E 6BT, UK. E-mail addresses: [email protected] (S. Berlinski), [email protected] (S. Galiani), [email protected] (M. Manacorda). 0047-2727/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.jpubeco.2007.10.007 S. Berlinski et al. / Journal of Public Economics 92 (2008) 1416–1440 1417 preschool attended. In order to control for unobserved household characteristics that are common to all children in the household and that might affect simultaneously exposure to pre-primary education and school progression we rely on a within household estimator that only exploits variability in the outcome and treatment variables across siblings. In order to account for the possibility of siblings' systematic differences in treatment and outcomes, we complement this strategy by instrumenting preschool attendance with average attendance rates by locality of residence and birth cohort. A major expansion in the provision of public pre-primary education in Uruguay over the last decade that led to an acceleration in preschool attendance among subsequent birth cohorts and that mainly affected more disadvantaged children generates sufficient variation in exposure to preschool education to warrant identification. We find a significant positive effect of preschool attendance on completed years of primary and secondary education. This works through both a fall in grade retention rates since the early school years and a reduction in drop- out rates among teenagers. The gains from attending preschool increase as children grow older, so that exposure to pre- primary education leads to gradually diverging paths in school attainment between treated and untreated children. We show that pre-primary education, by reducing grade retention early in school life, reduces the probability of subsequent grade failure and the incentives for early drop-out. Preschool exposure appears as a successful policy to prevent early school failure and its long lasting consequences. In poor countries, a large share of the population is excluded from the education system already at an early age and well before completion of the compulsory schooling cycle. Exclusion from the school system encompasses, in varying combinations, failure to enroll, late entry, intermittent and irregular attendance, high repetition rates and eventually early drop-out (UNESCO, 2005). In Uruguay, as in many other Latin American countries, the system is unable de facto to retain children in junior high school, despite this being in principle compulsory. Although graduation rates from primary school and enrollment in the first year of junior high are almost universal, as of 2001 about 25% of 25–29 years old declared not having completed junior high school. In this context, early exposure to the school system appears as a possibly successful policy option. A large body of literature in neuroscience, psychology and cognition makes the case for early childhood interventions. Research has established that learning is easier in early childhood than later in life, and that nutrition and cognitive stimulation early in life are critical for long-term skill development (see, among others, Bransford, 1979; Shonkoff and Phillips, 2000; Shore, 1997 and Sternberg, 1985). Thus, learning starts well before the day children enter primary school. The process of cognitive development starts at home and it is expected that pre-primary education facilitates this process by planning and providing systematic activities for children. Indeed, there is a widespread belief among educators that the benefits of pre-primary education are carried over to primary school. In particular, teachers identify lack of academic skills as one of the most common obstacles children face when they enter school (see Rimm-Kaufman et al., 2000). Also, they perceive preschool education as facilitating the process of socialization and self-control necessary to make the most of classroom learning (see Currie, 2001). In the economic literature, Carneiro and Heckman (2003) and Cunha et al. (2006) make a strong case for early investment in education. They suggest that the return to the investment in human capital declines exponentially during the life cycle, being the highest earlier in life. Not only the earlier the investment, the longer the time available for recovering it, but also some inputs are likely to have low returns when adopted later in life (e.g., it is hard to achieve any gains in IQ after a certain age) and potential complementarities arise among different types of investment, implying that higher levels of past inputs (and therefore of current human capital) yield higher returns to current investment in human capital. While there is substantial empirical evidence that intensive early education interventions targeted specifically to disadvantaged children lead to significant benefits, both in the short and in the long run (see, among others, Lee et al., 1990; Barnett, 1993; Barnett, 1995; Currie and Thomas, 1995; Reynolds 1998; Karoly et al., 1998; Danziger and Waldfogel, 2000; Currie, 2001; Garces et al., 2002; Blau and Currie, 2004; Schweinhart et al., 2005; Belfield et al., 2006),1 much less is known about the benefits of expanding pre-primary education for the population as a whole. Using variation across US states in the funding of kindergarten initiatives in the late 60s and early 70s Cascio (2004) finds that kindergarten reduces subsequent grade retention. Using data from the Early Childhood Longitudinal Study, Magnuson 1 Although some of this work reports (e.g., Barnett, 1995; Currie and Thomas, 1995; Currie, 2001) that the gains in IQ and standardized achievement tests that are caused by these interventions may fade out over time, it also reports long lasting benefits on measures of scholastic achievement (e.g., grade retention, special education placement, school drop-out, high school graduation and college enrollment) as well as a reduction in crime and welfare usage. 1418 S. Berlinski et al. / Journal of Public Economics 92 (2008) 1416–1440 et al. (2007) find that pre-primary education in the US is associated with higher reading and mathematics skills at primary school entry, but that these effects dissipate for most children by the end of first grade. Exploiting a natural experiment for Argentina, Berlinski et al. (2006) find a positive effect of pre-primary school attendance on third grade standardized Spanish and Mathematics test scores and on primary school pupils' behavioral outcomes such as attention, effort, class participation, and discipline. A major challenge in identifying the causal effect of pre-primary school attendance on later school outcomes is non- random selection into early education. Positive selection, whereby parents whose children attend pre-primary school possess characteristics that promote better school performance, would result in a spurious positive correlation between preschool and later academic outcomes.