Hitting beyond the target: How early childhood interventions benefit siblings

There is widespread recognition that 250 million children in low-income countries are at risk of failing to reach their developmental potential [Black et al. 2017]. Fostering child development is one of the UN’s Sustainable Development Goals, and global commitments to invest in Early Childhood Development (ECD) are substantial: since 2000, the Inter-American Development Bank has committed more than 1.7 billion USD to ECD projects, and the World Bank invested 3.3 billion USD in ECD programmes between 2000 and 2013 [Berlinski and Schady 2015, Sayre et al. 2015]. 

A large body of evidence shows the potential for early childhood interventions to have significant positive impacts on targeted children [Attanasio et al. 2022, Garcia and Heckman 2023]. However, we know much less about how the gains are distributed across children within families, i.e. including those who are not explicitly targeted. Other things equal, early life interventions could increase inequality between children in the same household by only improving the developmental outcomes of a targeted child. Whether other siblings also benefit depends on how parents respond to the intervention as they make decisions on how to allocate resources across all their children. Parents may decide to withdraw resources from other children to focus on the better-endowed target child with potentially higher returns to investment, or they may attempt to equalise inputs and outcomes across children, despite differences in their endowments. The trade-offs parents face across siblings are likely most acute in more resource deprived environments [Rosenzweig and Schultz 1982]. 

Experimental evidence of the distributional impacts of an early childhood intervention in Northern Nigeria

In our research [Carneiro, Rasul and Salvati 2023], we study how inequality across siblings within families develops in the context of a targeted early childhood intervention in Northern Nigeria, where 70% of households live below the poverty line and baseline levels of human capital deprivation among children are high. We study an intervention that provides information and cash transfers to randomly selected households triggered by the verified pregnancy of a target child. The intervention takes place during the target child’s first 1000 days of life (including in utero) – a period in which the foundations of physical and cognitive development over the life cycle are laid [Grantham-McGregor et al. 2007].

The information component of the intervention consists in the public provision of messages to households on best practices and nutrition during the 1000-day window. The resource component provides sizeable monthly unconditional cash transfers to mothers upon verification of pregnancy when the intervention begins, lasting for two years. 

While the provision of information and resources has potential benefits for all children, the intervention is triggered by the verified pregnancy of a specific target child. Thus, the intervention creates inequality between the targeted child and its older siblings due to improvements in the initial endowment of the targeted child. Our earlier work documents substantial benefits of the intervention for the targeted child, leading to significant reductions in the incidence of stunting [Carneiro et al. 2021]. Our current study provides novel experimental evidence on the impacts of such interventions across all children within the household. We examine not only the distribution of outcomes but also the distribution of parental investments across children. We thus establish the cost effectiveness of the intervention based on a fuller set of impacts.

We shed light on the effects of the intervention across siblings by considering both the contemporaneous effects of the intervention on older siblings of the target child, as well as dynamic effects on younger siblings born to the same mother after the target child over a four-year post-intervention period. We exploit child-specific data on physical development, health, nutrition and inputs for the target child, their older sibling, and a younger sibling.

Treated parents increase inputs across all siblings in the household, despite differences in their children’s early development

Our results show that there are differential impacts of the intervention across children in the household. For the older sibling of the target child, height and stunting outcomes do not improve in response to the intervention, which contrasts with sizeable impacts on stunting for the target child (Figure 1, Panel A). This may occur because the intervention affects older siblings outside of the critical first 1000 days of life, and stunting outcomes are harder to shift later in life as they reflect past deprivation. However, on a range of other health and nutrition outcomes, such as the likelihood of being ill or injured, the likelihood of having diarrhoea and an index measuring dietary diversity we find significant benefits for older siblings (Figure 1, Panel B). In addition, we find significant improvements in parental behaviours towards older siblings that are not directly related to the information received during the intervention, for example the provision of de-worming medication (Figure 1, Panel C). Overall, the evidence suggests that parents increase investments in both the target child and their older sibling, despite differences in their early-life physical development. 

For younger siblings, we also do not observe gains on stunting (Figure 1, Panel A). However, on a range of other outcomes including other aspects of their health, and a range of parental behaviours, we document significant gains from the intervention for the younger sibling (Figure 1, Panels B and C). 

Figure 1: Impacts of the intervention across siblings in the household

Notes: The figure reports intention-to-treat estimates in terms of percentage impacts relative to the control mean. For older sibling and target child effects are measured two years after the intervention begins. For younger sibling effects are measured four years after the intervention begins. Confidence intervals at 95%.

The lack of physical development gains for younger siblings is a puzzle, since they benefit from improved parental knowledge throughout their pregnancy. We provide three explanations for this puzzle. First, we show that the intervention reduces birth spacing between the target child and their younger sibling, with potential detrimental effects for the latter in terms of physical development. Second, the intervention increases maternal labour supply over time, thereby potentially reducing maternal time investments in the younger sibling. Finally, there is fade-out in maternal knowledge related to the practice of immediately breastfeeding a newborn, which is reflected in a significant reduction in the likelihood of the younger sibling being put to breast immediately relative to the target child. The fade-out of this specific practice may have persistent negative effects on physical development in the early years.

The differential effects of the intervention across siblings may not have been intended by treated parents. If anything, they appear to privilege equal treatment of children, as parents increase their investments by similar amounts for different siblings, as opposed to investing more in their most healthy, or in their least healthy children (which would, respectively, reinforce or compensate health differences produced by better pre-natal behaviours resulting from the intervention).

When accounting for gains across siblings, the rate of return to the early childhood intervention rises ten-fold

We establish the cost effectiveness of the intervention based on this fuller set of impacts on family welfare rather than those related to the target child alone. If gains from the intervention are measured exclusively through changes in physical development, then the annualised rate of return (IRR) to the intervention is the same as when only gains to the target child are considered. However, if we consider gains across a broader set of dimensions, the IRR to the intervention rises substantially. To illustrate, we combine our estimates of the intervention impacts on de-worming rates across siblings with existing estimates of long-run earnings impacts of de-worming in early life [Hamory et al. 2021]. We show that the annualised IRR rises ten-fold when gains for older and younger siblings are accounted for, relative to when only gains to the target child are considered. Thus, such interventions can generate larger returns than currently appreciated if families respond by improving the human development of all their children, despite differences in their endowments.

References

Attanasio O, S Cattan, and C Meghir (2022), “Early Childhood Development, Human Capital and Poverty,” Annual Review of Economics 14: 853-92.

Berlinski S, and N Schady (2015), “The Early Years: Child Well-Being and the Role of Public Policy,” New York: MacMillan.

Black M M, S P Walker, L C Fernald, C T Andersen, A M DiGirolamo, C Lu, D C McCoy, G Fink, Y R Shawar, J Shiffman, A E Devercelli, Q T Wodon, E Vargas-Baron, and S Grantham-McGregor (2017), “Early childhood development coming of age: Science through the life course” The Lancet 389(10064): 77-90.

Carneiro P, L Kraftman, G Mason, L Moore, I Rasul, and M Scott (2021), “The Impacts of a Multifaceted Pre-natal Intervention on Human Capital Accumulation in Early Life,” American Economic Review 111: 2506-49.

Carneiro P, I Rasul, and F Salvati (2023), “Families as Drivers of Inequality: Experimental Evidence from an Early Childhood Intervention,” Working Paper.

Garcia J L and J J Heckman (2023), “Parenting Promotes Social Mobility Within and Across Generations,” Annual Review of Economics 15: 1.

Grantham-McGregor S, Y B Cheung, S Cueto, P Glewwe, L Richter, and B Strupp (2007), “Developmental Potential in the First 5 Years for Children in Developing Countries,” The Lancet 369: 60-70.

Hamory J, E Miguel, M Walker, M Kremer, and S Baird (2021), “Twenty-year Economic Impacts of Deworming,” PNAS 118(14).

Rosenzweig M R and T P Schultz (1982), “Market Opportunities, Genetic Endowments, and Intrafamily Resource Distribution: Child Survival in Rural India,” American Economic Review 72: 803-15.

Sayre R, A E Devercelli, M J Neuman, and Q Wodon (2015), “Investing in Early Childhood Development: Review of the World Bank’s Recent Experience” Washington, DC: World Bank Group.