Adam Wagstaff, Naoko Watanabe
Poverty and Human Resources, Development Research Group
World Bank, September, 2000

Available as PDF file [37p.] at:
http://econ.worldbank.org/docs/1189.pdf

".... Despite the development community's shift in emphasis toward
the poor, malnutrition, like other dimensions of poor health, is con-
centrated among the worst off. Yet targets are still defined in terms
of population averages. Consider, then, this information about malnu-
trition rates among different economic groups in 20 developing coun-
tries.

Among the conclusions Wagstaff and Watanabe reach about malnutrition
rates among different economic groups:

* Inequalities in malnutrition almost always disfavour the poor.

* It's not just that the poor have higher rates of malnutrition. The
rate of malnutrition declines continuously with rising living stan-
dards. The tendency of poorer children to have higher rates of stunt-
ing and underweight is not due to chance or sampling variability.
Inequalities in stunting and underweight, as measured by the concen-
tration index, are statistically significant in almost all countries.

* Inequalities in underweight tend to be larger than inequalities in
stunting, which tend to be larger than inequalities in wasting.

* In most cases, whatever the malnutrition indicator, differences in
inequality between countries are not statistically significant.

* Even if attention is restricted to the cross-country differences
inequality that are statistically significant, interesting conclu-
sions merge. Egypt and Vietnam have the most equal distributions of
malnutrition, and Nicaragua, Peru, and, to a lesser extent, Morocco
have highly unequal distributions.

* Some countries (such as Egypt and Romania) do well in terms of both
the average (the prevalence of malnutrition) and the distribution
(equality). Others do badly on both counts. Peru, for example, has a
higher average level of stunting than Egypt and higher poor-nonpoor
inequality. But many countries do well on one count and badly on the
other. Brazil, for example, has a far lower (less than 20 percent)
stunting rate overall than Bangladesh (more than 50 percent) but has
four times as much inequality s measured by the concentration index).

* Use of an achievement index that captures both the average level
and the inequality of malnutrition leads to some interesting rank re-
versals in the country league table. With stunting, for example, fo-
cusing on the achievement index moves Egypt (a low-inequality coun-
try) from sixth position to fourth, higher than Brazil and Russia
(two countries with high inequality)...."