Thumbnail Image
Echenique_umd_0117E_22591.pdf(11.47 MB)
No. of downloads:
Publication or External Link
Echenique, Juan Agustin
Bhargava, Alok
In April 2007, the Government of Chile established a comprehensive support system for early childhood named ”Chile Crece Contigo” (ChCC). This social protection system was the first one of its class in Latin America. However, ten years later, the empirical evidence about the benefits of this policy is still scarce. This dissertation looks to fill that knowledge gap and answer broader questions about promoting skills formation in the first years of life. Chapter 1 introduces the reader to the complex structure of the comprehensive support system Chile Crece Contigo, revises the literature that argues the need for this program, and finally revises the existing evidence about this policy. In Chapter 2, I study a novel service offered as part of the comprehensive early childhood support system Chile Crece Contigo: the stimulation workshops. This service consists of workshops for children and their parents designed to foster and accelerate the learning process of those lagging in their developmental benchmarks based on psycho-motor evaluation. The paper exploits variations in the provision of the service (only children with lags in development are offered the referral to sensory rooms) and non-compliance in treatment attendance to test if the treatment can meet their target of closing the gap in human capital formation between children who received help for enhancing their development trajectory and those who followed the usual path of development. Based on information from Electronic Health Records from an urban district in Santiago, Chile, I document in first place take-up rate behavior based on observable characteristics and baseline evaluations indicating positive sorting into the treatment. Second, I show how the stimulation services offered by Chile Crece Contigo have a higher rate of effectiveness in reducing developmental lags in children with lower test scores at baseline. The average difference for children diagnosed with developmental lags at baseline ranges between 0.7-0.4 standard deviations between 8mo-18mo and 1.0-0.5 standard deviations for 18-36mo. These results suggest that the effectiveness indicator used by policy-makers underestimates the returns of the intervention due to its heterogeneity. Second, efforts to increase the take-up rates for children close to the cutoff values are needed to improve the overall returns of the intervention. Chapter 3 studies the effects of a comprehensive early childhood support system on human capital accumulation. Specifically, we explore differences in educational achievement of the first generations of children exposed to the comprehensive child development support system ”Chile Crece Contigo.” To study this, we exploit the gradual implementation of the policy and the age eligibility requirements to estimate the returns of availability of the policy on data from seven cohorts (2012-2018) of fourth-grade students in Chile. We find sizable positive effects in mathematics (0.21 of a standard deviation) and language (0.23 of astandard deviation) test scores for municipalities that started the program during or before their prenatal stage compared to children that the program began when they were older than sixty months. Estimates from an event-study design show that the exposure returns dissipated for children thirty-six months old or older when the policy started. This result is consistent with the schedule of interventions and early detection instruments established. When we look at the difference in the returns to exposure across gender and socioeconomic status, we find evidence that (i) a comprehensive child support system has higher returns on boys, which could be explained partially by differences in access to need-based services, (ii) these differences across gender differences occur in children with higher levels of exposure, and (iii) we do not find relevant differences between students classified as low-socioeconomic background and not classified in this category.