The distribution of child nutritional status across countries and over time

Priya Bhagowalia, Purdue University

Abstract

Malnutrition is manifested in various degrees of under- and over-nutrition, with differences and rapid changes in prevalence and severity. In 2000 persistence of underweight caused an annual loss of about 138 million disability-adjusted life years (DALYs). The worldwide effect of high body mass index was linked to a loss of 33 million DALYs (WHO 2002). Standard measures of a population’s nutritional status, based on the “headcount” number of people exceeding specific thresholds for under-and overweight, do not account for differences at other levels of bodyweight, whether at the extremes or within the “normal” range. Nutritional status captured with a crude measure like bodyweight, is a continuous rather than a discrete variable, and changes at every level are correlated with changes in health risks. So, simple headcounts of the fraction of people exceeding a threshold leave much to be desired. This dissertation introduces a new approach to characterizing the distribution of a population’s nutritional status. Recently, continuous measures from economics that go beyond “headcounts” by considering severity beyond the threshold have been applied to monitor nutritional status (Sahn and Stifel, 2002; Jolliffe, 2004; Madden, 2006; Sahn, 2007). I draw upon the poverty literature to construct Foster-Greer-Thorbecke measures for the incidence and severity of under- and overweight, based on deviations in either direction from the median of a healthy population. I apply these measures to the nutritional status of over 400,000 preschool children, using 130 DHS surveys covering 53 countries from 1986 to 2006. Unlike conventional threshold-based methods, the new approach counts changes in every child’s bodyweight. Regressions of the weight-for-height FGT measures on real income, agricultural productivity, gender equality and income equality indicate that the new measures are more sensitive. When testing these variables’ influence on nutritional status, estimated coefficients are two to three times when using the new measure. In my sample, real income and gender equality are crucial to the nutritional status of pre-school children perhaps more than income inequality. Since health risks vary with body weight throughout the normal range and in the extremes, sub-threshold levels of under- and over-nutrition should be included in measures of malnutrition.

Degree

Ph.D.

Advisors

Masters, Purdue University.

Subject Area

Agricultural economics|Public policy

COinS