Inequalities in Childhood Malnutrition in India: Some Evidence on Group Disparities

Abstract

This paper examines inequalities in child malnutrition in India through three distinct — although inter‐related — types of empirical analysis. First, it reports the socio‐economic inequalities in childhood malnutrition across different Indian states. Second, it decomposes the gap in malnutrition between children belonging to poor and non‐poor households to understand the disadvantageous distribution of health determinants and their effects. This analysis indicates that the distribution of endowments and positive maternal characteristics are significant in widening the gap between the child malnutrition among poor and non‐poor households. Third, it examines the inter‐group disparities in child malnutrition and notes that child groups privileged in terms of income, mother’s nutritional status and education have lower malnutrition, whereas the group adverse in all three characteristics endures the most. The paper concludes that policies to reduce malnutrition inequalities should recognize that endowment revisions can be more effective if appended with behavioural interventions.

Key words:

• Child health
• Malnutrition
• Health inequality
• Group inequality
• Decomposition
• India

Acknowledgements

We are grateful to the two referees and to the Editors for their comments and suggestions on this paper, to the participants at the HDCA Annual Conference, New Delhi 2008 for helpful discussions, and to T. K. Sundari Ravindran, D. Narayana, V. Santhakumar and T. R. Dilip for helpful discussion on the subject matter of this paper. The usual caveat applies.

Notes

1 The CI has certain attractive properties as compared with certain other measures of health disparities (see Wagstaff et al., 1991; Kakwani et al., 1997). Recently, Erreygers (2006) presented a generalization of the traditional CI that preserves its main properties and overcomes several of its methodological shortcomings.

2 Alternatively, Equation (5) can be rewritten as Equations (i) and (ii) where the interaction effect is placed in the unexplained and the explained components, respectively, as follows:

3 This sub‐group classification could depend upon the relevance of any particular grouping either from the policy perspective or along the concerns of data availability and accuracy. Perhaps in certain countries there can be more than three important dimensions that could be equally amenable for policy. Given such possibilities, there is no reason to restrict ourselves only to three identities and the applicability of the Subramanian and Majumdar (2002) approach is immediate.

4 For a detailed discussion of the measure and its properties see Subramanian and Majumdar (2002); for an application, see Majumdar and Subramanian (2001).