• TABLE OF CONTENTS
HIDE
 Front Cover
 Front Matter
 Title Page
 Table of Contents
 Foreword
 Acknowledgement
 Main
 Reference
 Back Cover






Title: Overcoming malnutrition : is there an ecoregional dimension ?
CITATION THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00085376/00001
 Material Information
Title: Overcoming malnutrition : is there an ecoregional dimension ?
Physical Description: Book
Language: English
Creator: Sharma, Manohar
Garcia, Marito
Qureshi, Aamir
Brown, Lynn
Publisher: International Food Policy Research Institute
Place of Publication: Washington, D. C.
Publication Date: February, 1996
Copyright Date: 1996
 Record Information
Bibliographic ID: UF00085376
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: 34729778 - OCLC

Table of Contents
    Front Cover
        Front Cover
    Front Matter
        Page i
    Title Page
        Page ii
    Table of Contents
        Page iii
        Page iv
    Foreword
        Page v
    Acknowledgement
        Page vi
    Main
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
    Reference
        Page 19
        Page 20
    Back Cover
        Page 21
Full Text











Overcoming Malnutrition:
Is There an Ecoregional
Dimension?


Manohar Sharma, Marito Garcia,
Aamir Qureshi, and Lynn Brown


2 20
VISION








































"A 2020 Vision for Food, Agriculture, and the Environment" is an initiative of
the International Food Policy Research Institute (IFPRI) to develop a shared
vision and a consensus for action on how to meet future world food needs
while reducing poverty and protecting the environment. It grewv out of a
concern that the international community is setting priorities for addressing
these problems based on incomplete information. Through the 2020 Vision
initiative, IFPRI is bringing together divergent schools of thouthi on thsee
issues, generating research, and identify\ in recommendations.

This discussion paper series presents technical research results that encom-
pass a wide range of subjects drawn from research on policy-relevant
aspects of agriculture, poverty, nutrition, and the environment. The discus-
sion papers contain material that IFPRI believes is of key interest to those
involved in addressing emerging Third World food and development prob-
lems. These discussion papers undergo review but typically do not present
final research results and should be considered as works in progress.






Food, Agriculture, and the Environment Discussion Paper 10


Overcoming Malnutrition:

Is There an Ecoregional

Dimension?

Manohar Sharma, Marito Garcia,
Aamir Qureshi, and Lynn Brown













International Food Policy Research Institute
1200 Seventeenth Street, N.W.
Washington, D.C. 20036-3006 U.S.A.
February 1996















Contents


Foreword v
Conceptual and Empirical Framework: Poverty in Ecoregions 2
Description of the Nutrition Database 6
Results 6
Conclusions 16
References 19














Tables



1. Regression analysis of the relationship of per capital gross national
product (GNP) and prevalence of underweight children 5
2. Malnutrition and food production by ecoregions, 1990 7
3. Estimates of malnourished children in Asia, 1990 8
4. Regional distribution of malnourished children, 1990 10
5. Distribution of malnourished children by agroecological zone, 1990 10



Illustrations

1. Factors linking environment and human nutrition 3
2. Malnutrition (underweight) by geographical regions, 1990 9
3. Malnutrition (underweight) by agroecological zones, 1990 9
4. Malnutrition by ecoregions: Prevalence of underweight preschool
children by agroecological zones, 1990 12
5. Malnutrition by ecoregions: Prevalence of underweight preschool
children by regions and agroecological zones, 1990 13
6. Malnutrition (stunting) by agroecological zones, Sub-Saharan Africa,
1987-92 15
7. Malnutrition (stunting) by land type, Sub-Saharan Africa, 1987-92 15
8. Malnutrition (underweight) by agroecological zones, Sub-Saharan
Africa, 1987-92 17
9. Malnutrition (underweight) by highland and lowland areas, Sub-Saharan
Africa, 1987-92 17















Foreword


As part of IFPRI's 2020 Vision initiative, which seeks to develop an international consensus
on how to meet future world food needs while reducing poverty and protecting the environ-
ment, IFPRI has been looking at projections of global supply and demand for food through
the year 2020. In agrarian developing countries such projections can be greatly influenced by
the natural environment because climate, terrain, and soil characteristics drive the agricultural
system. The health and nutritional status of rural families can also be affected if natural
conditions encourage disease and malnutrition, thus affecting productivity.
The Consultative Group on International Agricultural Research (CGIAR) is supporting a
system-wide emphasis on developing the concept of data collection and research based on
ecoregional zones, rather than on individual countries. As part of this initiative, IFPRI has
undertaken research aimed at mapping the worldwide distribution of poverty as indicated by
the prevalence of malnourished children across ecoregions. The results presented in this
discussion paper represent our beginning efforts toward this goal.
We at IFPRI feel that understanding the ecoregional dimensions of poverty and malnutri-
tion potentially could be a major step forward in alleviating poverty sustainably by the year
2020 because it may help to target scarce resources more effectively.

Per Pinstrup-Andersen
Director General, IFPRI














Acknowledgments


This paper has benefited from comments provided by Michael Collinson, Lawrence Haddad,
Per Pinstrup-Andersen, Steve Vosti, Jay Willis, and participants of the workshop entitled
"Ecoregions of the Developing World-A Lens for Assessing Food, Agriculture, and the
Environment to the Year 2020," held November 7-9, 1994, in Airlie, Virginia, U.S.A., as part
of IFPRI's 2020 Vision initiative.




















The mandate of the Consultative Group on
International Agricultural Research (CGIAR)
is to generate sustainable improvements in the
productivity of agriculture, forestry, and fisheries in
developing countries, in order to alleviate poverty
and eliminate or reduce hunger (CGIAR 1994). The
CGIAR is proposing to develop its future strategies
and priorities using a concept of ecoregional zones,
based on the proposition that agroecological condi-
tions largely determine the production potential and
the population-supporting capacity of developing
countries. But food security-access to enough food
for a healthy and productive life by all people at all
times-is not guaranteed from sustainable food
production alone. To gain access to the food pro-
duced, people usually must have sufficient incomes
to purchase the food. Given the CGIAR's mandate,
this raises three fundamental questions:
What conceptual and empirical issues are
raised in addressing poverty and malnutrition
at the agroecological level?
What is already known about poverty and mal-
nutrition in different agroecological zones of
the world?
What are the implications for priority setting
by the CGIAR of the distribution of poverty
and malnutrition across agroecological zones?
Switching to an agroecological framework for
resource management, rather than a regional- or country-
specific framework, requires that information critical
for planning and targeting be organized by agro-
ecological zones (AEZs). This presents a potential
problem. To date, almost all of the studies on the
location, magnitude, and distribution of poverty are by
individual country or by subnational aggregates
within political or administrative boundaries that gen-
erally do not coincide with agroecological boundaries.
Surveys that document poverty even at regional levels
are available for only a limited number of countries.
Thus, analysis of socioeconomic data from an AEZ
perspective suffers from use of indirect rather than
direct measurements of poverty. The most complete


analysis attempted so far is a report to the Technical
Advisory Committee (TAC) of the CGIAR (Broca
and Oram 1991). The principal indicator of poverty
used in that report is calorie consumption, specifically
the population falling below a calorie consumption
cutoff. The report explicitly recognizes the limitations
of the analysis: ". .. there are no direct survey data on
food consumption, nor food expenditure, available for
some major thermal regions and agroecological
zones." Thus, a comprehensive mapping of poverty
and malnutrition across AEZs is as yet unavailable.
The present paper is an attempt to expand under-
standing of poverty and malnutrition within and
across AEZs. It is a significant departure from pre-
vious work on this issue because it uses direct mea-
surements of nutritional status to characterize pov-
erty within and across AEZs. It is also more
comprehensive than previous works because the sur-
vey information covers about 94 countries in the
developing world, encompassing more than 90 per-
cent of the total population of developing countries.
It should be stated at the outset, however, that
the results given in this paper are more precise for
Sub-Saharan Africa than for the other regions. This
is because agroecological boundaries have been
identified within countries in Sub-Saharan Africa
where the country boundary encompasses more than
one AEZ. This has not yet been accomplished for
many countries outside Sub-Saharan Africa. Once
these boundaries have been finalized, it will be pos-
sible to present precise numbers of poor and mal-
nourished people by AEZ on a worldwide basis.
The paper is organized into four main sections.
First, a conceptual framework draws out the linkages
between poverty and malnutrition and the interaction
of these linkages within an agroecological setting. The
second section defines the main indicator of poverty
used and describes how the nutrition database was
compiled and organized and the method used in aggre-
gating data to the agroecological level. The third section
presents results of a global analysis together with a more
detailed discussion of the picture in Sub-Saharan
Africa. Finally, the conclusions are summarized.









Conceptual and Empirical
Framework:
Poverty in Ecoregions

Using Nutritional Status in Ecoregional
Analysis: Conceptual Issues
The poverty and nutritional status of any population
group is the result of the interaction between the
natural environment and human action. The inter-
relationships involved are highly complex and con-
stantly evolving over time.
The main elements of the analytical framework
are described in Figure 1. The natural environment
(air, land, and water) defines, principally, the natural
resource base or production potential of land.' The
seasonality of crop production, cropping patterns,
choice of crops, and yield rates are fundamentally
influenced by the natural environment. The agricul-
tural system and consequent productivity are thus,
by and large, determined by climatic and soil charac-
teristics. The natural environment also encompasses
factors that adversely affect the productivity of hu-
man action. These include conditions that encourage
or mitigate the growth and spread of certain diseases
(for example, malaria-carrying mosquitos survive
only in certain temperature zones). Additionally, the
natural environment includes those topological fac-
tors that affect activity levels of the resident popula-
tion, either directly (through terrain, for example) or
indirectly (by influencing the type of farming system
adopted, for example).
Human action is applied to the natural environ-
ment in order to produce goods and services for
consumption. However, both the nature of the action
applied and the resulting output depend on a number
of factors, including (1) the level of technology,
(2) the social organization of resource management
(especially ownership and control), and (3) the
demographic attributes of the resident population.2
A few examples will illustrate the point.
Technology. Irrigation provides an obvious ex-
ample. Irrigated areas of arid zones generally have a
higher carrying capacity than unirrigated areas of
arid zones. This means that irrigated areas can sup-


port the same level of welfare for higher population
densities or higher levels of welfare for the same
population density.
Organization. The effects of organization-
related factors are important at both micro and
macro levels. At the micro level, for example, the
institutional framework governing property rights
has a direct effect on the intensity of cultivation and
on the choice of technology for agricultural produc-
tion. Farmers are unlikely to invest in future land
productivity potential when they are unsure of their
future right to use the land. Differences in agricul-
tural productivity between market-based and non-
market-based economies are clear illustrations of
this point. At the macro level, government policies
and actions not only determine resource allocation
between competing sectors, but also directly influ-
ence household and individual welfare through the
public distribution of food, health, and education. In
Sri Lanka, for example, government policies on sub-
sidized food distribution, health, and educational in-
puts have resulted in a far lower prevalence of mal-
nutrition than in other areas of the region.
Demography. Even a casual look at the demo-
graphic map of the world shows that people gravitate
to the most promising areas. Because population
movements occur in response to differences in returns
to labor, there is a natural tendency for people to move
to more favorable AEZs where it may be possible to
earn a higher income. Over the long run, however,
wages decline when population becomes too large;
therefore, returns to labor can be expected to converge
across AEZs, particularly within individual countries
where there may be few impediments to labor move-
ment. Not surprisingly, then, the most fertile tracts of
land, such as the Indo-Gangetic plains in India, are
precisely where population densities are the highest,
with the result that the rich natural resource base is
thinly spread across the population.
The two basic factors-natural environment and
human action-interact to create what may be termed
a "livelihood environment." This interaction deter-
mines the parameters that influence private household
production and consumption decisions as well as the
production and distribution of public goods such as
health and education. The livelihood environment


IThe natural environment affects nonagricultural production possibilities as well as agricultural. However, this paper deals with
developing countries where agriculture is overwhelmingly the dominant sector, so most references relate to agriculture.
2For example, recent studies indicate that human physiology is an important variable, in that it adapts and responds significantly to
the surrounding environment (Payne and Lipton 1994).










Figure 1-Factors linking environment and human nutrition


I1 I


Natural environment


Political and economic structure


LIVELIHOOD ENVIRONMENT
Agricultural Government policy Off-farm Infrastructure Price structure
production Sector priorities employment
system Macro policies opportunities
Welfare planning




HOUSEHOLD
Income Access to food Education Access to health and
sanitation inputs


Dietary intake


I>


I I --


o-



Nutritional status
of individual


comprises the following components: agricultural and
farming systems; nonagricultural income-earning
opportunities; government policies such as price poli-
cies, trade, and other macroeconomic policies; and
infrastructure relating to health services, education,
transportation, and communication.
The system shown in Figure 1 is interactive both
horizontally and vertically, producing both forward
and backward linkages. Changes in government
policies, for example, may trigger changes in tech-
nology adoption through agricultural extension,
which affects soil conditions and land productivity


Prevalence of disease


through human action. On the other hand, environ-
mental degradation may affect cultivation practices
that, in turn, may induce government action regard-
ing technology dissemination through agricultural
extension services.
The household, where decisions regarding the
welfare of family members are made, responds to the
livelihood environment, seeking to optimize its con-
trol of and access to resources such as consumption
goods, including food, health, education, and clean
water. The results of these decisions directly affect
the well-being of individuals through their effects on


Human action





"'_Z





N"_


_-'Z


SI


I









dietary intake and incidence of disease. Anthro-
pometric measures capture the physiological aspects
of general welfare levels, which is one aspect of the
outcome of household decisions.


Nutritional Status as a
Poverty Indicator
Assessing poverty at the agroecological level re-
quires indicators that possess four important attrib-
utes. First, the indicator should be a good proxy for
describing the level of poverty. Second, it should be
possible to standardize the indicator across popula-
tions and ethnic groups to facilitate comparisons
across countries and regions of the world. Third, the
indicator should be such that it can be disaggregated
or measured at smaller subnational levels, preferably
corresponding to AEZs, to obtain direct measure-
ments. Fourth, the indicator should be sensitive to
changes in the magnitude of poverty.
In this paper, anthropometric indexes of pre-
school children are used (body weight and height
standardized by age) as a measure of nutritional status
and an indicator of poverty. Anthropometric indexes
can be termed outcome measures, a quantification of
growth failure. These individual health and develop-
ment outcomes are the result of the complex inter-
actions of socioeconomic factors and environmental
determinants (UN ACC/SCN 1993). In poor commu-
nities, inadequacies in diet, brought about by low
agricultural productivity and consequent low in-
comes, and infection and diseases are often the major
determinants of growth failure. Environmental and
climatic factors exacerbate these outcomes through
their impact on land productivity and the spread and
prevalence of infectious disease. The use of anthro-
pometry has been described by the United Nations
Administrative Committee on Coordination-Sub-
Committee on Nutrition (UN ACC/SCN 1993) as "the
best general proxy for constraints to human welfare of
the poorest, including dietary inadequacies, functional
impairment, and/or environmental health risks."
Also, anthropometric indexes for preschool children
are collected extensively on a worldwide scale.
Anthropometric indicators are objective in that
they can be compared with internationally accepted
normative standards. An accepted global measure,
anthropometric indexes can be used in comparing
the growth patterns of children across countries and
across ethnic groups. The underlying premise is that
healthy preschool children of any race or ethnic ori-
gin will grow at the same rate (WHO 1983).


The three most commonly used anthropometric
indexes of nutritional status are stunting (low height-
for-age), wasting (low weight-for-height), and under-
weight (low weight-for-age). Stunting is widely used
as a measure of long-term malnutrition, whereas wast-
ing is used to characterize short-term malnutrition and
underweight is a reflection of both stunting and wast-
ing. Of the three, the most broadly reported indicator
is the prevalence of underweight children.
The internationally recognized comparative
standard is that of the U.S. National Center for Health
Statistics (NCHS). Anthropometric data are standard-
ized, using the NCHS reference population, and
Z-scores (or standard deviations) are calculated, based
on the median NCHS reference population age or
height. Thus, those below -2 standard deviations of
the NCHS median reference height-for-age are de-
fined as stunted, and those below -2 standard devia-
tions of the median weight-for-age are underweight.
Anthropometric indexes are broadly available
for preschool children, as they represent a measure
of developmental impairment for one of the most
vulnerable population groups. While the indexes
pertain only to preschool children, they are also a
good (but not perfect) indicator of deprivation at the
household level.
How sensitive are anthropometric indexes of nu-
tritional status to the level of poverty in populations?
An analysis relating nutritional status to incomes at the
national level presented in the Second Report on the
WorldNutrition Situation (UN ACC/SCN 1992) indi-
cates that the relationship is strongest at the lower end
of the range. Increasing GNP per capital from US$300
to US$600 is associated with a decline in the preva-
lence of underweight children from about 34 percent
to about 17 percent, or a reduction of about 50 percent.
Beyond $900 of annual income the effect of increas-
ing incomes on nutrition diminishes. Regression
analysis, using cross-national data from 93 countries,
illustrates the strong, statistically significant relation-
ship between GNP per capital and prevalence of under-
weight children, even when controlling for factors
such as education, public social expenditures, and
regional peculiarities (Table 1). This evidence further
justifies using nutritional status as a poverty indicator.


Data Availability and Implicationsfor
Data Analysis and Interpretation
A major difficulty in analysis relates to data avail-
ability by AEZs. Where available information on
the natural resource base is limited, socioeconomic










Table 1-Regression analysis of the relationship
of per capital gross national product
and prevalence of underweight
children
Independent Variable Coefficient t-Statistic

Constant 27.40 12.57**
GNP per capital -0.002 2.11**
Female education -0.11 2.81**
Percent public expenditure
for social support -0.11 -1.67*
South Asia (dummy) 54.67 12.68**
R-2 0.89
F-statistic 94.3
Number of countries 93

Source: UN ACC/SCN 1993.
Note: Dependent variable: percentage of underweight children, by
country.
*Significant at the 5 percent level.
**Significant at the 1 percent level.



information delineated by natural resource-based
or agroecological boundaries (as opposed to a po-
litical or administratively set boundary) is in even
shorter supply.
This study uses ecoregions as defined by the
Technical Advisory Committee (TAC) of the
CGIAR as the natural resource delineation system.
The TAC/CGIAR ecoregions are essentially aggre-
gations of the original AEZs constructed by the Food
and Agriculture Organization of the United Nations
(FAO) between 1978 and 1981. FAO AEZs display
a "broadly uniform physical response to the applica-
tion of production technologies" and thus describe
"a geographical area that is homogeneous with re-
spect to its environment and natural resources, for
example, climate, land form, soils, and water bodies"
(Wood and Pardey 1993, 3). This exercise resulted in
45,000 unique agroecological cells for the develop-
ing world (Hunt 1993).
The TAC/CGIAR classification represents an at-
tempt to achieve a level of aggregation that strikes an
appropriate balance between practicality, on the one
hand, by generating a manageable set of zones, and
accuracy on the other, by aggregating countries or
regions into zones that are in some sense "homogene-
ous" (Pardey and Roseboom 1991, 12). The TAC


aggregation was based on two agroecological charac-
terizations: "major climate" and "length of growing
period,"3 reducing the original FAO classification
into nine broad AEZs. These zones are categorized as


AEZI =
AEZ2 =
AEZ3 =
AEZ4 =
AEZ5 =
AEZ6 =
AEZ7 =
AEZ8 =
AEZ9 =


warm, semi-arid tropics,
warm, subhumid tropics,
warm, humid tropics,
cool tropics,
warm, semi-arid subtropics,
warm, subhumid subtropics,
warm/cool, humid subtropics,
cool subtropics (summer rainfall), and
cool subtropics (winter rainfall).


In essence, these zones are really agroclimatic
zones, since information on variations in soil and
terrain was lost during aggregation. As Pardey and
Roseboom (1991) note, the potential bias arising from
the aggregation procedure clearly depends on the use
to which this zoning method is put. For example, if
nutritional outcomes are substantially affected by in-
come and public health facilities, and if these vary
significantly within agroclimatic zones, one would
expect significant aggregation bias in cross-zonal
comparisons of nutritional outcomes. Indeed, the
TAC was aware of such difficulties and recognized
that "socioeconomic circumstances important to ...
research, as well as the support needs of national
programs, are better differentiated by national and
regional boundaries" (TAC 1992, 4). The TAC there-
fore sorted the nine AEZs by major geographic re-
gions and used the term "regional agroecological zones
(RAEZs)" or "ecoregions" to describe them. These
ecoregions constitute more homogeneous units for
analysis and planning than the AEZs. The developing-
country regions are Asia (including the Pacific), Sub-
Saharan Africa, Latin America and the Caribbean, and
West Asia-North Africa. There are 23 ecoregions.4
A major omission in the TAC AEZ classifica-
tion is the failure to distinguish between lowland and
highland areas. This omission is important because
cropping systems and yield potentials tend to vary
markedly between the two areas (Broca and Oram
1991). Fortunately, in Sub-Saharan Africa, the Inter-


3"Major climate" is determined by mean daily temperature during the growing season; "length of growing period" is defined as the
number of days when both moisture and temperature permit crop growth.
4Not all geographical regions contain all the AEZs. For example, Sub-Saharan Africa contains only four out of the nine zones.









national Institute of Tropical Agriculture (IITA) has
disaggregated the TAC AEZs into upland areas
(more than 800 meters above sea level) and lowland
areas (below 800 meters above sea level) so that this
distinction can be included when using the Sub-
Saharan Africa data.



Description of the Nutrition
Database
A distinct advantage in using nutritional status as an
indicator of poverty is that a large number of institu-
tions have been carrying out nutrition surveys world-
wide. These have been compiled and analyzed by
groups such as the World Health Organization
(WHO), the UN ACC/SCN in Geneva, the United
Nations Children's Fund (UNICEF), and the World
Bank. In addition, various organizations, such as the
United States' Demographic and Health Surveys
Group and the Centers for Disease Control in
Atlanta, Georgia, have been compiling large data-
bases that have subnational nutritional status data,
which can be adapted to the ecoregional classifica-
tion. The Second Report on the World Nutrition
Situation (Volumes I and II), published by the UN
ACC/SCN (1992, 1993), and the subsequent Update
on the Nutrition Situation, 1994 provide good sources
of data. The WHO Global Database on Child Growth
(de Onis and Bloessner 1994) has been compiled and
periodically updated at WHO. The database used for
the present analysis is compiled from data from
these various sources. Anthropometric indicators are
defined for the 0-to-60-months age group in nearly all
of the surveys.
The database used here consists of data on the
prevalence of underweight and stunted children from
94 countries and more than 90 percent of the popula-
tion in the developing world. In the TAC/CGIAR
classification of AEZs, some 99 countries fall within
a single AEZ, whereas 33 countries have two or more
AEZs. For those countries that had multiple AEZs, the
national population was allocated among the zones,
based on the proportion of total arable land allocated
to each zone. Then the number of malnourished in
each zone was computed. The prorating factors were
based on Kassam (1991) and reported in Pardey and
Roseboom (1991). In the case of Sub-Saharan Africa,
a more detailed map of the AEZs prepared by the IITA
was available so that site-specific survey results could
be directly plotted for the relevant AEZ. How data
availability affects data analysis is summarized below:


Only national-
level data
available


Subnational
data available


Country
within a
Single
Ecoregion

Use data
as is



Use
weighted
average


Country
with
Multiple
Ecoregions

Use prorating
factors based
on Pardey and
Roseboom
1991
Map data
points to
specific
ecoregions


Results
The results presented here represent an initial attempt
to map the worldwide prevalence of underweight
children by ecoregions. As indicated previously, the
fact that nutritional status is a relatively good poverty
indicator means that the results also throw light on the
distribution of poverty among ecoregions. At this
stage, however, there are limitations to the analysis,
owing more to the incomplete nature of ecoregional
mapping than to a lack of nutritional status data. Since
an ecoregional map, which details the boundaries for
countries occupying more than one ecoregion, has
been finalized only for Sub-Saharan Africa, the results
provided here are preliminary, pending a complete
mapping of the coordinates of the ecoregions beyond
Sub-Saharan Africa.

Global Results
Tables 2 and 3 summarize the results aggregated for
the nine AEZs and four geographic regions. Because
Asia has an extremely large population and includes
countries that are diverse in population and income,
it has been broken down into two distinct regions-
Asia and the Pacific, and South Asia-for purposes
of analysis (Table 3). The tables show both the
prevalence of malnutrition and the numbers of mal-
nourished children. TAC (1992) data on population
and food production are also presented.
Figures 2 and 3 also present the data using a
technique called box plotting. In Figure 2, for exam-
ple, 50 percent of the undernourished children in the
sample for a given region fall within the box (ranging
from the 25th percentile to the 75th percentile). The
heavy line through the box represents the median. It
lies at 26 percent for Sub-Saharan Africa. After ex-
treme and outlying observations are eliminated, a line













Table 2-Malnutrition and food production by ecoregions, 1990

Agro- Irrigated Total Malnourished Children
ecological Food Arable Arable Popu-
Regions/Agroecological Zones Zone(AEZ) Productiona Land" Land' nation' Percent Numberc

(million (million (million (million) (million)
TGEd) hectares) hectares)


Sub-Saharan Africa
Warm, semi-arid tropics
Warm, subhumid tropics
Warm, humid tropics
Cool tropics
West Asia-North Africa
Warm, semi-arid tropics
Cool tropics
Cool subtropics (winter rainfall)
Asia
Warm, semi-arid tropics
Warm, subhumid tropics
Warm, humid tropics
Warm, semi-arid subtropics (summer rainfall)
Warm, subhumid subtropics (summer rainfall)
Warm/cool, humid subtropics (summer rainfall)
Cool subtropics (summer rainfall)
Latin America and the Caribbean
Warm, semi-arid tropics
Warm, subhumid tropics
Warm, humid tropics
Cool tropics
Warm, semi-arid subtropics (summer rainfall)
Warm, subhumid subtropics (summer rainfall)
Warm/cool, humid subtropics (summer rainfall)
Cool subtropics (summer rainfall)
Cool subtropics (winter rainfall)
Overall
Warm, semi-arid tropics
Warm, subhumid tropics
Warm, humid tropics
Cool tropics
Warm, semi-arid subtropics (summer rainfall)
Warm, subhumid subtropics (summer rainfall)
Warm/cool humid subtropics (summer rainfall)
Cool subtropics (summer rainfall)
Cool subtropics (winter rainfall)


104.2
33.3
22.7
33.4
14.8
65.2
0.3
0.9
64.0
732.6
113.0
69.4
124.6
117.9
54.2
138.1
115.6
141.8
11.8
21.1
23.4
33.1
4.4
3.0
20.5
20.6
4.0
1,043.8
158.4
113.2
157.9
48.0
122.3
57.2
158.5
136.1
68.0


5.22 161.8 501.1
3.69 54.0 166.6
0.43 43.8 106.3
0.44 37.3 152.3
0.66 16.7 75.9
18.66 83.0 316.0
0.10 0.2 5.5
0.25 1.4 8.0
18.31 81.4 302.5
135.75 462.5 2,739.7
22.15 85.9 466.2
7.70 40.5 228.9
14.50 45.0 474.5
43.02 106.0 456.6
10.14 32.5 212.9
22.77 78.4 485.9
15.47 72.4 414.7
14.07 161.4 447.7
1.76 10.9 37.7
2.16 26.1 70.3
1.80 21.8 87.3
2.02 15.4 130.2
2.59 8.1 13.5
0.47 7.1 3.8
1.14 33.7 62.5
0.10 32.2 27.8
2.03 6.1 14.6
173.70 868.7 3,996.5
27.70 161.0 676.0
10.29 110.4 405.5
14.94 82.3 714.1
2.68 32.1 206.1
45.61 114.1 470.1
10.61 39.6 216.7
23.91 112.1 548.4
15.57 106.4 442.5
20.34 87.5 317.1


28.3
10.4
6.1
6.7
5.1
9.3
0.1
1.0
8.1
139.8
37.6
13.5
30.3
31.2
7.4
9.4
10.5
6.0
0.7
1.0
1.1
2.0
0.5
0.0
0.6
0.0
0.0
183.4
47.9
20.6
38.0
8.1
31.7
7.4
10.4
10.6
8.2


aData are from TAC 1992.
bData are from UN ACC/SCN (1993).
CBased on anthropometric measurement of underweight in preschool children (-2 standard deviations of the median weight-for-age using U.S. National
Center for Health statistics standard) in 91 developing countries.
dTotal grain equivalent.











Table 3-Estimates of malnourished children in Asia, 1990


Regions/Agroecological Zones


Agroecological
Zone (AEZ)


Asia and the Pacific
Warm, semi-arid tropics 1
Warm, subhumid tropics 2
Warm, humid tropics 3
Warm, semi-arid subtropics
(summer rainfall) 5
Warm, subhumid subtropics
(summer rainfall) 6
Warm/cool, humid subtropics
(summer rainfall) 7
Cool subtropics (summer rainfall) 8
South Asia
Warm, semi-arid tropics 1
Warm, subhumid tropics 2
Warm, humid tropics 3
Warm, semi-arid subtropics
(summer rainfall) 5
Warm, subhumid subtropics
(summer rainfall) 6
Cool subtropics (summer rainfall) 8
All Asia


Population
(million)


1,466
5
68
333

110

89

486
375
1,274
461
160
142

347

124
40
2,740


Malnourished Children
Percent Number
(million)


21 9.4
21 8.3
60 96.6
63 37.5
62 11.2
65 12.9

54 27.6

63 5.1
54 2.3
41 139.8


Source: TAC 1992.
Note: Malnourishment is based on anthropometric measurement of underweight in preschool children
weight-for-age using U.S. National Center for Health statistics standard) in 91 developing countries.


is drawn below the box that represents the smallest
value that is not an outlier. The line above the box
indicates the highest value that is not an outlier.5
The box plot technique is well-suited to a study
of the spread of observations, but the median value
shown is misleading, since the box plot weighs each
observation equally (irrespective of the size of popu-
lation that it represents). Weighted means are there-
fore computed for each regional and agroecological
group separately (see Tables 2 and 3) and are pre-
sented in the next section using bar charts. For exam-
ple, the mean for each region is computed as


(-2 standard deviations of the median


2-", pw
Sw,



where pi is the prevalence rate in the ih unit and w, is
the population of the i" unit.
Overall, results show that in 1990, 34 percent of
preschoolers-approximately 183 million chil-
dren-in developing countries were malnourished.
Where the Poor Live: A Regional Perspective.
The largest share of underweight children is by far the
highest in South Asia; the average rate for this region


5The sketch below shows how a box plot is constructed.
values more than 3 box-lengths from the 75th percentile (extremes)
o values more than 1.5 box-lengths from the 75th percentile (outliers)
largest observed value that is not an outlier

50 percent of 75th PERCENTILE
cases have values MEDIAN
within the box 25th PERCENTILE

smallest observed value that is not an outlier
o values more than 1.5 box-lengths from the 25th percentile (outliers)
values more than 3 box-lengths from the 25th percentile (extremes)









Figure 2-Malnutrition (underweight) by geographical regions, 1990
Prevalence of underweight preschool children (percent)
70



40- ----------------------------------------- -----hanistan-

Iran
30 ---- -- ------------------------------------

0 - ------ --- - ------------- Yemen People's
20 - - -------- Democratic
Republic


0-

10
-0 ------------------------------------------------------- -----------I---------------------------------
Sub-Saharan Asia and the Latin America South Asia West Asia-
Africa Pacific and the Caribbean (N = 10) North Africa
(N= 55) (N= 15) (N= 52) (N= 18)
Geographical Region (TAC)
Note: Underweight is more than two standard deviations below the median weight-for-age of the U.S. National Center for Health Statistics.


Figure 3-Malnutrition (underweight) by agroecological zones, 1990
Prevalence of underweight preschool children (percent)
70
SOBangladesh
60 ----


50
40


-10 1
Warm, Warm, Warm, Cool Warm, Warm, Warm/Cool Cool Cool
Semi-arid Subhumid Humid Tropics Semi-arid Subhumid Humid Subtropics Subtropics
Tropics Tropics Tropics (N = 19) Subtropics Subtropics Subtropics (Summer (Winter
(N = 34) (N = 26) (N = 38) (Summer (Summer (Summer Rainfall) Rainfall)
Rainfall) Rainfall) Rainfall) (N= 5) (N= 16)
(N = 6) (N = 3) (N = 3)
Agroecological Zone (TAC)
Note: Underweight is more than two standard deviations below the median weight-for-age of the U.S. National Center for Health Statistics.


. . - . I


-----oAfghanistar


- - - - - -


I"f~~










as a whole is 60 percent (Table 4). Within South Asia,
the rate ranges from a little over 40 percent to about
65 percent (Figure 2). Only a few countries in West
Asia-North Africa, Latin America and the Caribbean,
and Sub-Saharan Africa have malnutrition prevalence
rates that even approach the minimum rate in South
Asia. Because South Asia is one of the most populated
regions in the world, the absolute number of malnour-
ished children-96.6 million-is also the highest,
accounting for more than half of the world's total.
Sub-Saharan Africa has the next highest re-
gional prevalence rate of malnutrition, as measured
by underweight children. The spread between the
highest and the lowest country rates in the region is
wider than in any other region, ranging from under
10 percent to close to 50 percent. The average for the
region as a whole is 31 percent with the number of
malnourished children totaling 27.3 million, which
represents about 16 percent of the total malnourished
children in the world.
While the mean prevalence rate of undernutrition
for Asia and the Pacific is relatively low at 25 percent,
the high population density results in large absolute
numbers of malnourished children-43.2 million.
The distribution of malnutrition for Asia and the
Pacific (excluding South Asia) is similar to that of
Sub-Saharan Africa in range, but in essence it is
shifted downward, with the magnitudes observed at
both the lower and the upper ends of the distribution
smaller than in Sub-Saharan Africa. The malnutrition
prevalence rates for both Latin America and the Car-
ibbean and West Asia-North Africa are considerably
lower than for the other developing regions, although
both regions contain pockets of high prevalence such
as Guatemala and Yemen. In some parts of West
Asia-North Africa, malnutrition prevalence rates are
high and similar to rates in parts of South Asia. By
contrast, the highest rates found in Latin America and
the Caribbean are only about 25 percent.


Table 4-Regional distribution of malnourished
children, 1990
Malnourished Children
Region Percent Number
(million)
Sub-Saharan Africa 30.5 28.3
West Asia-North Africa 18.8 9.3
Asia and the Pacific (excluding
South Asia) 25.0 43.2
South Asia 60.0 96.6
Latin America and the Caribbean 10.0 6.0
Source: IFPRI estimates (preliminary).


Table 5-Distribution of malnourished children
by agroecological zone, 1990


Agroecological Zone


Malnourished Children
Percent Number


(million)
Warm, semi-arid tropics 49.0 48.8
Warm, subhumid tropics 36.4 20.6
Warm, humid tropics 37.0 38.0
Cool tropics 26.0 8.1
Warm, semi-arid subtropics
(summer rainfall) 44.0 31.7
Warm, subhumid subtropics
(summer rainfall) 38.0 7.4
Warm/cool, humid subtropics
(summer rainfall) 19.0 10.0
Cool subtropics (summer rainfall) 23.0 10.6
Cool subtropics (winter rainfall) 17.4 8.2
Source: IFPRI estimates (preliminary).


Where the Poor Live. An Agroecological Per-
spective. Table 5 summarizes the prevalence of mal-
nutrition across AEZs. The highest prevalence rates
for undernutrition occur in the semi-arid zones of
both the tropic and subtropic ecoregions, 49 percent
in the former and 44 percent in the latter. Coinciden-
tally, both these zones show wide variations in their
prevalence rates (Figure 3) ranging from less than
10 percent malnourished to more than 60 percent
malnourished. The likely driving factor behind this
high variation, particularly in the subtropics, is irri-
gation. In arid areas, irrigation leads to significantly
higher agricultural productivity, resulting in higher
incomes and improved nutritional status. Given the
relatively high population densities of these zones,
the number of malnourished children is also high,
particularly in the warm, semi-arid tropics. The dis-
ease burden in these areas may also be high, com-
pounding the problem of malnutrition.
High malnutrition prevalence rates also prevail
in the warm, humid and warm, subhumid areas of the
tropics (AEZ2 and 3) and the warm, subhumid areas
of the subtropics (AEZ6). Although agricultural pro-
ductivity is high in these zones, population densities
are also high, so that per capital availability of arable
land in the humid tropics (AEZ3) and in the sub-
humid tropics (AEZ6) ranks the lowest among all
AEZs. A potential compounding factor could be
insect-carried and waterborne diseases, which are
more prevalent in these regions and spread more
rapidly, given the high population density. The
prevalence of malnutrition is generally lower in the
cooler subtropics (AEZ7, AEZ8, and AEZ9). In
AEZ7 and AEZ9, the range of malnutrition preva-
lence rates across geographical zones is less than in










any of the other AEZs. In the cool subtropics with
summer rainfall, by contrast, the spread of observa-
tions is the greatest.
Combining Agroecological and Regional Per-
spectives for an Ecoregional View. Figures 4 and 5
illustrate the prevalence of malnutrition among chil-
dren by ecoregions. In Figure 4, regional data are
presented by AEZs and in Figure 5 the same data are
organized geographically. As already noted, the
spread in prevalence rates across geographical re-
gions is much larger than between ecoregions.
Among the AEZs, the prevalence rate of malnutri-
tion ranges from a low of 17.4 percent in the cool
subtropics with winter rain (AEZ9), to a high of 49
percent in the warm semi-arid tropics (AEZ1). Be-
tween geographical groupings, the range is from 60
percent malnourished in South Asia to 10 percent in
the Latin America and the Caribbean region.
Figures 4 and 5 also indicate that geographic
regional variation is substantial within each AEZ.
This is an important point, indicating that the impact
of the natural environment on welfare is signifi-
cantly modified by other socioeconomic factors. The
rates are consistently lower for Latin America and
the Caribbean in all AEZs. Not surprisingly, per
capital income is also the highest in this region. This
highlights the importance of national-level policies
that foster economic growth and address problems of
food insecurity and undernutrition.
In examining differences in rates among AEZs,
prevalence of malnutrition is generally found to be
higher in the warmer areas of the tropics and sub-
tropics (36 to 40 percent) than in the cooler areas (17
to 26 percent). Prevalence is highest of all in the
warm, semi-arid tropics (AEZ1). The reason behind
this pattern becomes clearer once the regional attri-
butes are also considered. The following explana-
tions are possible.

1. AEZI encompasses large areas in Asia (149 mil-
lion hectares) and Sub-Saharan Africa (1,246 mil-
lion hectares), where the majority of the poor are
concentrated. In South Asia, for example, 62 per-
cent of the people live in the semi-arid zone
(Broca and Oram 1991).
2. Agricultural productivity measured by agricul-
tural output per unit of land during 1981-85
(US$141 in 1980 prices) is the lowest in AEZ1,
with the exception of the cool tropical zone
(AEZ9), which nevertheless has a low prevalence
rate (Pardey and Roseboom 1991). Also, since
1961, the growth rate for agricultural productivity
for AEZ1 has been one of the lowest.


3. Due to the semi-arid nature of AEZ1, "the Green
Revolution has nearly bypassed this area" and
"increasing food production in the rainfed areas
in ways that conserve and enhance the resource
base is an extraordinarily difficult task, given the
uncertainty of rainfall" (TAC 1992, 49).

People in the warm areas of the arid and semi-
arid tropics may be sick more often than those in
other AEZs. Two populations in Mali illustrate this
point: the poorer nomadic population of the largely
desert northern regions and the resident population
of the Mopti Delta, located in the semi-arid tropics.
The infant and child mortality rates of the nomadic
children, who visit the Mopti Delta during the dry
season each year, are lower than those of the resident
Mopti Delta children. By the end of the dry season,
the waters and ponds of the Mopti Delta are heavily
polluted. With the onset of the rainy season, mosquito-
borne diseases become common, increasing the sick-
ness rate among the delta children. But these dis-
eases have little effect on the nomadic children
because they have already left the delta to return to
the desert north (Hill and Randall 1983). High dis-
ease burdens contribute significantly to increasing
rates of malnutrition.
Although the warm, humid tropical zone
(AEZ3) has the highest agricultural output per hec-
tare of cultivatable land, this is not reflected in a
lower prevalence of malnutrition among children; at
37 percent, the rate of malnutrition here is still high.
At 65 percent, South Asia has the highest malnutrition
rate for this AEZ. The TAC estimates that 70 percent
of this ecoregion's population lives below the pov-
erty line, which is indicative of the correlation be-
tween poverty and malnutrition. Even in Latin
America and the Caribbean, where incomes are
much higher, the prevalence rate of malnutrition in
the warm, humid tropical zone is relatively high at
10 percent, compared with just 2 percent in the cool
subtropics. That higher rates of malnutrition coexist
with a relatively rich resource base is yet another
indication that other factors also affect nutritional
outcomes. One important modifying factor in this
zone is population density. In Asia, for example, the
availability of arable land is just 0.09 hectare per
person; even in relatively land-abundant Latin
America and the Caribbean, it is only 0.25 hectare
per person (TAC 1992). Another important modify-
ing factor is the interaction between health, nutrition,
and disease. TAC (1992) reports that in Sub-Saharan
Africa, diseases such as malaria and human try-
panosomiasis are endemic to this zone. The combi-












Figure 4-Malnutrition by ecoregions: Prevalence of underweight preschool children by
agroecological zones, 1990


AEZI: Warm, semi-arid tropics
Sub-Saharan Africa
West Asia-North Africa
Asia and the Pacific (excluding South Asia)
South Asia
Latin America and the Caribbean
All developing countries
AEZ2: Warm, subhumid tropics
Sub-Saharan Africa
Asia and the Pacific (excluding South Asia)
South Asia
Latin America and the Caribbean
All developing countries
AEZ3: Warm, humid tropics
Sub-Saharan Africa
Asia and the Pacific (excluding South Asia)
South Asia
Latin America and the Caribbean
All developing countries
AEZ4: Cool tropics
Sub-Saharan Africa
West Asia-North Africa
Latin America and the Caribbean
All developing countries
AEZ5: Warm, semi-arid subtropics (SR)
Asia and the Pacific (excluding South Asia)
South Asia
Latin America and the Caribbean
All developing countries
AEZ6: Warm, subhumid subtropics (SR)
Asia and the Pacific (excluding South Asia)
South Asia
Latin America and the Caribbean
All developing countries
AEZ7: Warm/Cool, humid subtropics (SR)
Asia and the Pacific (excluding South Asia)
Latin America and the Caribbean
All developing countries
AEZ8: Cool subtropics (SR)
Asia and the Pacific (excluding South Asia)
South Asia
Latin America and the Caribbean
All developing countries
AEZ9: Cool subtropics (WR)
West Asia-North Africa
Latin America and the Caribbean
All developing countries


0 10 20 30 40 50 60 70
Percent of underweight preschool children


Notes: (SR) is summer rainfall, and (WR) is winter rainfall. Underweight is more than two standard deviations below the median weight-for-age of
the U.S. National Center for Health Statistics.


1


I


M











Figure 5-Malnutrition by ecoregions: Prevalence of underweight preschool children by regions
and agroecological zones, 1990


Sub-Saharan Africa
Warm, semi-arid tropics
Warm, subhumid tropics
Warm, humid tropics
Cool tropics
West Asia-North Africa
Warm, semi-arid tropics
Cool tropics
Cool subtropics (WR)
Asia and the Pacific (excluding South Asia)
Warm, semi-arid tropics
Warm, subhumid tropics
Warm, humid tropics
Warm, semi-arid subtropics (SR)
Warm, subhumid subtropics (SR)
Warm/cool, humid subtropics (SR)
Cool subtropics (SR)
South Asia
Warm, semi-arid tropics
Warm, subhumid tropics
Warm, humid tropics
Warm, semi-arid subtropics (SR)
Warm, subhumid subtropics (SR)
Cool subtropics (SR)
Latin America and the Caribbean
Warm, semi-arid tropics
Warm, subhumid tropics
Warm, humid tropics
Cool tropics
Warm, semi-arid subtropics (SR)
Warm, subhumid subtropics (SR)
Warm/cool, humid subtropics (SR)
Cool subtropics (SR)
Cool subtropics (WR)
All developing countries
AEZ1: Warm, semi-arid tropics
AEZ2: Warm, subhumid tropics
AEZ3: Warm, humid tropics
AEZ4: Cool tropics
AEZ5: Warm, semi-arid subtropics (SR)
AEZ6: Warm, subhumid subtropics (SR)
AEZ7: Warm/cool, humid subtropics (SR)
AEZ8: Cool subtropics (SR)
AEZ9: Cool subtropics (WR)


0 10 20 30 40 50 60 70
Percent of underweight preschool children


Notes: (SR) is summer rainfall, and (WR) is winter rainfall. Underweight is more than two standard deviations below the median weight-for-age of
the U.S. National Center for Health Statistics.









nation of thinly spread resources and a relatively
high prevalence of infection and disease leads to
high rates of malnutrition, in spite of the well-
endowed natural resource base.
Data across AEZs show that the cool tropics
(AEZ4) have relatively lower prevalence rates of
malnutrition among children than the warm tropics,
even though they are close to the bottom of the
agricultural productivity ladder. However, when
data are further disaggregated by geographic regions,
it can be seen that the prevalence of malnutrition is
highest in the cool tropics of Latin America and the
Caribbean, West Asia-North Africa, and Sub-Saharan
Africa, the three regions that contain this zone. In
West Asia-North Africa, for example, where the cool
tropics zone includes countries such as Ethiopia,
prevalence rates are as high as 53 percent. In Latin
America and the Caribbean, the prevalence rate in
the cool tropics, which includes Bolivia, Colombia,
and Ecuador, is 14 percent. According to the 1992
TAC report, high population densities in this zone
have resulted in overexploitation of the land, leading
to long-term declines in soil fertility and widespread
soil erosion. The zone also is characterized by poor
soil conditions and steep slopes that put substantial
constraints on productivity. Among all the AEZs,
per capital availability of arable land is lowest in the
cool tropics. The apparent contradiction between
low prevalence rates of malnutrition in the cool tropics
when only AEZs are considered and high prevalence
rates in those geographic regions that have cool
tropic zones is driven by the high prevalence rates of
malnutrition in South Asia, a region that has no cool
tropics zone. Thus, at the AEZ level, the worst inci-
dence of malnutrition is in the semi-arid tropic and
subtropic zones, which are predominately Asian.
The cool subtropics (AEZs 7, 8, and 9) registered
the largest gains in agricultural productivity (Pardey
and Roseboom 1991). Although no time-series data
on nutrition are available for this area, the cross-
sectional data suggest much lower levels of malnutri-
tion in these areas than anywhere else. Further, this
observation holds true for every geographic region,
although it is notable that overall malnutrition preva-
lence rates are generally higher in South Asia. This is
thought to be due to a number of important interven-
ing factors. In Asia, the warm/cool, humid subtropics
(AEZ7) comprise the most densely populated areas of
the world. Again, in Asia, the cool subtropics with
summer rainfall zone (AEZ8) includes countries such
as Bhutan, Mongolia, Nepal, and parts of China and
India, where a mountainous topography severely con-
strains intensive cultivation.


There is little variation in the prevalence rates of
malnutrition among children across AEZs in the
Sub-Saharan Africa and South Asia regions, the two
regions that are home to most of the world's poor.
The similarity across AEZs in the poorest regions of
the world again indicates the importance of socio-
economic factors, especially those related to popula-
tion, in determining malnutrition outcomes inde-
pendently of environment-related factors.


Results for Sub-Saharan Africa
Unlike the global data, the special data set for Sub-
Saharan Africa is not only site-specific (as opposed
to national or regional), but data can be grouped
according to the TAC AEZs and according to land
elevation. This is possible because IITA has sub-
divided each AEZ into upland and lowland areas
(IITA undated). Population data on the survey areas,
however, are not available. This makes it impossible
to draw inferences about the population represented
by the survey data; therefore, the site-specific survey
data cannot be weighted according to population
size. Socioeconomic data that can be superimposed
on the survey areas are also not available. For this
reason, no means are computed and analysis is con-
ducted entirely in terms of the box plots.
Stunting (Low Height-for-Age). When the mal-
nutrition prevalence rates across AEZs are further
divided into highland and lowland areas (Figure 6),
it is clear from the plots that the highland-lowland
distinction is important in each of the TAC AEZs.
However, the difference between highland and low-
land stunting rates is least prevalent in the warm,
subhumid zones. That average rates of stunting seem
to converge around similar levels across AEZs
within both highland and lowland areas is another
discernible pattern. In the lowlands, stunting rates
converge between 20 and 40 percent. In the high-
lands, with the exception of the warm, subhumid
zone, the stunting prevalence rates cluster just under
60 percent.
Figure 7 compares the rates of stunting across a
broad classification of highland and lowland areas in
Sub-Saharan Africa. There is a marked difference
not only in the average prevalence of stunting, but
also in the range of stunting prevalence rates ob-
served across the two land elevations. While the
average rate of prevalence is close to 60 percent in
the highlands, it is approximately 30 percent in the
lowlands. Though the range of observations in high-
land and lowland areas is about the same, the distri-












Figure 6-Malnutrition (stunting) by agroecological zones, Sub-Saharan Africa, 1987-92

Prevalence of stunted preschool children (percent)
inn .


] - - - - - - - - --. _ _
O Madagascar










0 Lesotho


o Ethiopia 0 Ethiopia
0 Tanzania
@ Zambia 0 Ethiopia
8 Tanzania


-- ------- 7T-


U n I I
Highland Highland Highland Highland Highland Lowland Lowland Lowland
Cool Warm Arid/ Warm Warm Warm Warm Arid/ Warm Warm
(N = 21) Semiarid Cool Humid Subhumid Semi-arid Subhumid Humid
(N= 11) (N= 17) (N = 3) (N= 12) (N = 36) (N = 25) (N = 23)

Agroecological Zone

Note: Stunting is more than two standard deviations below the median height-for-age of the U.S. National Center for Health Statistics.




Figure 7-Malnutrition (stunting) by land type, Sub-Saharan Africa, 1987-92

Prevalence of stunted preschool children (percent)

100



80


Highland
(N = 64)


Lowland
(N = 84)


Land Type

Note: Stunting is more than two standard deviations below the median height-for-age of the U.S. National Center for Health Statistics.


0 Ethiopia
O__ 0 Tanzania
o Ethiopia
o Tanzania


- n










bution is much more compact in the lowlands than
the highlands, perhaps because the highlands consti-
tute a more heterogeneous environment (especially
in infrastructure) than the lowlands.
Prevalence of Underweight (Low Weight-for-
Age). The highland-lowland differential is main-
tained for the weight-for-age indicators (Figures 8
and 9), although the difference in the average preva-
lence rates is not as marked as in the case of stunting
indicators. The absolute prevalence rates for under-
weight are also much lower than those for stunting,
at a little less than 25 percent for the lowlands and
approximately 30 percent for the highlands.
Both the warm, semi-arid and warm, humid zones
show significant differences between highlands and
lowlands, with the highland underweight prevalence
rates, at 30 percent, exceeding those of the lowlands.
In the warm, subhumid zone, however, the situation is
reversed so that, at 28 percent, the prevalence of
underweight preschoolers is a little higher in lowland
areas than in highland areas. The range of observed
prevalence rates of underweight children also differs
across lowland and highland regions. While the rates
seem to lie between 20 and 30 percent in lowland
areas, underweight prevalence rates vary from 20 to
40 percent in the highlands. Thus, as with stunting, the
highland areas appear to constitute a more heteroge-
neous environment than the lowlands.


Conclusions

This analysis reveals, as summarized below, that
there is a correlation between the prevalence of mal-
nutrition and ecoregional boundaries:
Globally, the prevalence rate of malnutrition
is highest in the warm, semi-arid tropic and
subtropic zones, where agricultural productiv-
ity conversely is relatively low.
Where regions contain a cool tropic zone
(Sub-Saharan Africa, Latin America and the
Caribbean, and West Asia-North Africa), the
prevalence of malnutrition is often the highest
of all of the zones.
Where data availability allowed a comparison
of the prevalence rates across highland and
lowland areas-in Sub-Saharan Africa-both
the rates of stunting and underweight were
higher in highland areas than in lowland areas.
The prevalence rates of malnutrition were
generally lower in the cooler subtropic eco-
regions than in the warmer tropical zones.


Because there appears to be a rather strong cor-
relation between malnutrition and poverty, the above
findings also describe the possible correlation be-
tween poverty and ecoregions.
These observations, taken together, suggest the
following:

1. A linkage between agricultural productivity and
nutrition can be detected at the ecoregional level,
particularly when data are further disaggregated by
geographical regions. According to Pardey and
Roseboom (1991), the arid and semi-arid zone
(AEZ1); the warm, subhumid tropic zone (AEZ2);
and the cool tropic zone (AEZ4) experienced the
smallest increase in land and labor productivities.
Coincidentally, AEZI and AEZ2 are also zones
with higher prevalence rates of malnutrition.
2. Ecological characteristics also contribute to in-
creasing rates of malnutrition as evidenced by the
differentials in highland and lowland rates of
malnutrition. Although it is difficult to pinpoint
the exact nature of causation from the data, some
ideas come to mind.
Highland areas generally have an inferior
resource base relative to lowland areas. The
cooler and harsher climate combined with
steep slopes in the highland areas put major
constraints on agricultural growth. The limi-
tations on irrigation in highland areas im-
pede technological advancement, especially
in the production of rice and wheat.
The topography of highland areas tends to
impede the provision of infrastructure, limit-
ing communication, transportation, and mar-
keting opportunities. This not only limits the
marketing of agricultural inputs and outputs,
but also impedes the delivery of public ser-
vices such as health and education. The lack
of these services, in turn, contributes to in-
creased levels of malnutrition.
Highland areas may have benefited less
from agricultural research, both at national
and international levels, due to the percep-
tion that they were areas of lower agricul-
tural potential. Thus there have been fewer
technological advancements in farming sys-
tems in these areas.
3. Given the higher prevalence of malnutrition in
the hotter and more humid areas of the tropics, a
stronger interaction between climate and the
prevalence of infections in these zones cannot be
ruled out. Diseases such as malaria and human







17




Figure 8-Malnutrition (underweight) by agroecological zones, Sub-Saharan Africa, 1987-92

Prevalence of underweight preschool children (percent)
60


O Ethiopia
50 ----------------------------------------------------------------------------- a-------------

O Tanzania


40- ---------------------- --------



30 -



20- ---- ----- -



1 0 - - - - - - - - - - - - - - - - - - - - - - - - - -

10




10
-10 1 1
Highland Highland Highland Highland Highland Lowland Lowland Lowland
Cool Warm Arid/ Warm Warm Warm Warm Arid/ Warm Warm
(N = 20) Semi-arid Cool Humid Subhumid Semi-arid Subhumid Humid
(N= 11) (N= 13) (N=3) (N= 12) (N = 38) (N = 26) (N = 21)

Agroecological Zone

Note: Underweight is more than two standard deviations below the median weight-for-age of the U.S. National Center for Health Statistics.




Figure 9-Malnutrition (underweight) by highland and lowland areas, Sub-Saharan Africa, 1987-92

Prevalence of underweight preschool children (percent)

60

0 Ethiopia
50


40 -------


30 -- ------------ --




20 -------- -------- ------------ ------ --- - -------- --- -- - -- _--- -- -----
20 --








-10 ,---------------------------------------------- ------------------------------
-10
Highland Lowland
(N= 59) (N= 85)
Land Type

Note: Underweight is more than two standard deviations below the median weight-for-age of the U.S. National Center for Health Statistics.









trypanosomiasis (in Sub-Saharan Africa) are, for
example, endemic in the warm, humid tropics.
Since these zones also have higher population
densities, this also contributes to a faster spread
of infection.

This analysis indicates that, while the natural en-
vironment does influence the amount of poverty and
malnutrition in an area and its prevalence among chil-
dren, these effects are conditioned in large part by
other socioeconomic factors, principally technologi-
cal, demographic, and potentially health-related fac-
tors, resulting in considerable variations in income
and nutritional status even within AEZs. A clear illus-
tration of this is the widespread prevalence of malnu-
trition in precisely those areas that have the highest
agricultural productivity (warm, humid zones). The
regional (as opposed to agroecological) aspect of the
distribution of the malnourished population also can-
not be ignored. By and large, the malnourished live in
South Asia and Sub-Saharan Africa. The malnutrition


prevalence rates are consistently and appreciably
higher for these two regions across all AEZs. This
indicates that international policymakers should con-
tinue to adopt important regional foci in their strate-
gies and policy implementation. A balanced agenda,
comprised of agricultural research with a greater em-
phasis on highland area farming systems, particularly
those of Asia and Sub-Saharan Africa, together with
policies designed to promote economic growth in the
poorest regions of the world, may be the best strategy
to achieve sustainable long-term reductions in the
high levels of malnutrition observed in these regions.
It is important that both scientists and policymakers be
cognizant of and attentive to the links between the
natural environment, agricultural productivity, health,
and nutritional status. Analysis at the ecoregional
level suggests that targeting of scarce resources to
address the problems of poverty and malnutrition
could potentially be improved by 2020 by taking into
consideration these linkages between ecoregional
characteristics and poverty.















References


Broca, S., and P. Oram. 1991. Study on the location of
the poor. Paper prepared for the Technical Advi-
sory Committee of the Consultative Group on
International Agriculture Research. International
Food Policy Research Institute, Washington, D.C.

Garcia, M. 1994. Malnutrition andfood insecurity pro-
jections, 2020. 2020 Brief 6. Washington, D.C.:
International Food Policy Research Institute.

CGIAR (Consultative Group on International Agri-
cultural Research). 1994. CGIAR annual report
1993-1994. Washington, D.C.
de Onis, M., and M. Bloessner. 1994. WHO global
database on child growth. World Health Organi-
zation, Geneva.
Hill, A. G., and S. C. Randall. 1983. Infant and child
mortality in rural Mali. Research Paper No. 83-5.
London: London School of Hygiene and Tropical
Medicine.

Hunt, E. D. 1993. Agroecological zones for planning
and research: Using geographic information sys-
tems. Quarterly Journal of International Agri-
culture 32 (4): 388-408.
IITA (International Institute of Tropical Agriculture).
Undated. Agroecological zones in Africa. Map.
Ibadan, Nigeria: Agroecological Studies Unit.

Kassam, A. 1991. Estimate of total arable land by
AEZ by country-1990. Technical Advisory
Committee, Consultative Group on International
Agricultural Research, Rome. Mimeo.
Pardey, P. G., and J. Roseboom. 1991. National agri-
cultural research from a regional and agro-
ecological perspective. The Hague: International
Service for National Agricultural Research.


Payne, P., and M. Lipton. 1994. How Third World
households adapt to dietary energy stress: The
evidence and the issues. Food Policy Review 2.
Washington, D.C.: International Food Policy
Research Institute.

TAC (Technical Advisory Committee, Consultative
Group on International Agricultural Research).
1992. Review of CGIAR priorities and strate-
gies. Part I. TAC Secretariat, Food and Agricul-
ture Organization of the United Nations, Wash-
ington, D.C.

UN ACC/SCN (United Nations Administrative
Committee on Coordination-Sub-Committee
on Nutrition). 1992. Second report on the world
nutrition situation. Vol. I, Global and regional
results. Geneva.

1993. Second report on the world nutrition
situation. Vol. II, Country methods, trends, and
statistics. Geneva.

1994. Update on the nutrition situation,
1994. Geneva.

WHO (World Health Organization). 1983. Measur-
ing change in nutritional status. Geneva.

Wood, S., and P. G. Pardey. 1993. Agroecological
dimensions of evaluating and prioritizing re-
search from a regional perspective: Latin Amer-
ica and the Caribbean. Discussion Paper 93-15.
The Hague: International Service for National
Agricultural Research.

World Bank. 1993. Implementing the World Bank's
strategy to reduce poverty: Progress and chal-
lenges. Washington, D.C.







Food, Agriculture, and the Environment Discussion Papers


1. Alleviating Poverty, Intensifying Agriculture, and Effectively Managing
Natural Resources, by Per Pinstrup-Andersen and Rajul Pandya-Lorch, 1994

2. Sociopolitical Effects of New Biotechnologies in Developing Countries, by
Klaus M. Leisinger, 1995

3. Africa's Changing Agricultural Development Strategies: Past and Present
Paradigms as a Guide to the Future, by Christopher L. Delgado, 1995

4. A 2020 Vision for Food, Agriculture, and the Environment in Sub-Saharan
Africa, edited by Ousmane Badiane and Christopher L. Delgado, 1995

5. Global Food Projections to 2020, by Mark Rosegrant, Mercedita Agcaoili-
Sombilla, and Nicostrato Perez, 1995

6. A 2020 Vision for Food, Agriculture, and the Environment in Latin America,
edited by James L. Garrett, 1995

7. Agriculture, Trade, and Regionalism in South Asia, by Dean A. DeRosa and
Kumaresan Govindan, 1995

8. Major Natural Resource Management Concerns in South Asia, by Gerard J.
Gill, 1995

9. Agriculture, Technological Change, and the Environment in Latin America:
A 2020 Perspective, by Eduardo J. Trigo, 1995




















Manohar Sharma and Lynn Brown are research analysts and Aamir Qureshi was a research assistant
at the International Food Policy Research Institute. Marito Garcia is an economist at the World Bank.














































1200 SEVENTEENTH STREET, N.W. WASHINGTON, D.C. 20036-3006 U.S.A.
1-202/862-5600 FAX 1-202/467-4439 E-MAIL IFPRI@CGNET.COM


IFPRI




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs