Parents and Child Care in Colombia: Choices, Interactions, and Child Development

• In Event: 2-166 - Effective Scalable Models for Interventions to Improve Center-based Care and Education for Preschool-Age Children in LMICs: Progress and Future Directions

Integrative Statement

A large body of evidence connects child care programs and child developmental outcomes as does evidence showing effects of parental investments on child outcomes. However, less is known about how parents make child care decisions, the effects of child care choices on parental investments, and the interactions between parental and child care investments in the production of early human capital. Empirical evidence is scarce in both developing and developed countries, first because data with child care, parental investments, and child outcomes are rare and, second, because the estimation of such effects and interactions is complicated by potential endogeneity issues that arise because parents might choose child care programs and their investments in children based on a variety of unobserved parent and child characteristics, which can directly impact child developmental outcomes.
This paper uses a rich dataset from a randomized control trial of two different child care modalities in Colombia to explore how parents make child care choices, how their decisions affect the ways in which they invest in their children, and how parental and child care inputs interact in the production of developmental outcomes of children younger than 6. We estimated the production of early human capital considering the fact that parents might react in different ways to changes in the availability and cost of different child care modalities and that parents make their investments in children based on unobserved characteristics which also matter for child development. We used several sources of exogenous variation to estimate the model, including random assignment to the child care programs under study.

The sample consisted of 1,400 socioeconomically vulnerable children younger than 5 at baseline who participated in a cluster randomized control trial in 14 large cities in Colombia between 2010 and 2012 and were enrolled in 300 public home-based child care units at baseline. We used the Early Childhood Environmental Rating Scale - Revised (ECERS-R; Harms, Clifford & Cryer, 1998), the Infants and Toddlers Environmental Rating Scale - Revised (ITERS-R; Harms, Cryer & Clifford, 2003), and the Family Daycare Rating Scale (FDCRS; Harms & Clifford, 1989) to construct two factors of child care quality measuring aspects of process and structural quality of early education in different child care modalities. Similarly, we used rich household data to construct one factor for parental investments. In addition, we used a variety of developmental measures to construct developmental factors in three domains: cognitive, socioemotional, and health development. We used these data to estimate a three-staged model: (1) determinants of parental child care choices, (2) determinants of parental investments in children given child care choices, and (3) early human capital production function in which both parental and child care inputs interact. The second and third stages correct for potential endogeneity of parental choices by using a control function approach.
Parents choose different child care modalities based on structural quality but not on process quality features, which might reflect the fact that the former can be more easily observed by parents than the latter. The literature shows that process quality is crucial for developmental outcomes. In addition, the most relevant determinant of child care choices was distance to child care. Once selection is controlled, parents do not invest differentially depending on the child care modality chosen. More educated, richer parents invest more in their children. Parents invest more in the firstborn child versus siblings. Parental investments matter crucially for child development, and the magnitude of their effect significantly increases once a correction for endogeneity is implemented. In terms of child care inputs, a large effect of process quality on developmental outcomes emerged, but no significant effect of structural features of early education quality. Finally, high degrees of persistence of developmental outcomes over time, and complementarities across different developmental domains emerged, with cognition being relevant for further socioemotional skills and health.

Authors

• Raquel Bernal, Los Andes University
  Presenting Author

• Orazio Attanasio, University College London & Institute for Fiscal Studies
  Non-Presenting Author

• Federico Rodríguez, Los Andes University
  Non-Presenting Author

• Francis Vella, Georgetown University
  Non-Presenting Author