Explaining Cross-Region Disparities in Childhood Stunting in India

In the present study, we have tried to explore the drivers of the large disparities in childhood stunting (height-for-age Z scores distributions) among different regions of India. We compared the regions with relatively poor height-for-age Z scores (HAZ) by keeping the best performer southern region as benchmark and employed unconditional quantile regression (recentred influence function regression per se) and counterfactual decomposition (QR-CD in short) methods using unit level fourth round data of National Family Health Survey (NFHS). Results indicated that covariate effects (endowments) and coefficient effects (returns to endowments) vary across regions and quantiles when compared to the benchmark region. It has been observed that HAZ disparities are primarily attributable to the differences in returns to endowments, while endowments try to reduce such disparities when southern region is compared with the northern region. However, magnitude of the differences varies across quantiles. On the contrary, both covariate and coefficient effects have significantly increased the disparities in HAZ outcomes when southern region was compared with the central region. It was also noted that covariate effects dominate over coefficient effect with different magnitudes in various quantiles.