Files
Abstract
The Indian Integrated Child Development Services (ICDS) aims to improve the physical
and psychological well-being of children younger than five. However, previous evaluations find
that ICDS fails to significantly impact child stunting and that program placement is faulty. My
results contradict the lack of a significant treatment effect, but are consistent with problematic
program placement. Previous analyses of ICDS used probit to study placement, but the distribution of state-wise ICDS coverage is negatively skewed violating the normality assumption
of probit. To address this, I use beta regression to study placement and compare results with
probit analysis. In addition, using Propensity Score Matching (PSM) I find evidence of a significant, positive average and quantile treatment effects on stunting. Data are from the most
recent Indian Family and Health Survey (NFHS-3).