ArticlePDF Available


Using household survey data from a sample of 810 households, this paper analyses the determinants of children’s nutritional status and evaluates the impacts of improved maize varieties on child malnutrition in eastern Zambia. The paper uses an endogenous switching regression technique, combined with propensity score matching, to assess the determinants of child malnutrition and impacts of improved maize varieties on nutritional status. The study finds that child nutrition worsens with the age of the child and improves with education of household head and female household members, number of adult females in the household, and access to better sanitation. The study also finds a robust and significant impact of improved maize varieties on child malnutrition. The empirical results indicate that adoption of improved maize varieties reduces the probability of stunting by an average of about 26 %. © 2015 Springer Science+Business Media Dordrecht and International Society for Plant Pathology
Determinants of child nutritional status in the eastern province
of Zambia: the role of improved maize varieties
Julius Manda
&Cornelis Gardebroek
&Makaiko G. Khonje
&Arega D. Alene
Munyaradzi Mutenje
&Menale Kassie
Received: 6 January 2015 /Accepted: 27 November 2015
#Springer Science+Business Media Dordrecht and International Society for Plant Pathology 2015
Abstract Using household survey data from a sample of 810
households, this paper analyses the determinants of childrens
nutritional status and evaluates the impacts of improved maize
varieties on child malnutrition in eastern Zambia. The paper
uses an endogenous switching regression technique, com-
bined with propensity score matching, to assess the determi-
nants of child malnutrition and impacts of improved maize
varieties on nutritional status. The study finds that child nutri-
tion worsens with the age of the child and improves with edu-
cation of household head and female household members,
number of adult females in the household, and access to better
sanitation. The study also finds a robust and significant impact
of improved maize varieties on child malnutrition. The
empirical results indicate that adoption of improved maize
varieties reduces the probability of stunting by an average of
about 26 %.
Keywords Childrens nutritional status .Stunting .
Endogenous switching probit .Zambia
Malnutrition remains pervasive in many countries despite sig-
nificant reductions in income poverty in recent years (Horton et
al. 2008). More than 30 % of the developing worlds population
suffers from micronutrient deficiencies and approximately
one-third of the children in developing countries are either
underweight or stunted (World Bank 2008). Malnutrition is
the largest single factor contributing to the global problem of
disease and accounts for about 30 % of infant deaths (Headey
2013). Malnutrition also has adverse effects on the childs phys-
ical development, mental capacity, school performance, and
reduces adult labour productivity and wage earnings, as well
as overall economic growth (Apodaca 2008; Horton et al.
Malnutrition is widespread among children in Zambia and
it is one of the leading contributors to the high burden of
disease in the country (Masiye et al. 2010). According to the
UNDP (2011), about 50 % of children under the age of five are
stunted or too short for their age indicating chronic malnutri-
tion, while about 19 % of Zambian children are underweight
or too thin for their age.
Malnutrition principally results from the independent or
combined effects of three elements: inadequate food availability,
poor access to food by the hungry and poor food utilization
(Staatz 2000). Food availability refers to the supply of food
through adequate production (commercial and home produced),
food aid, or food imports (Apodaca 2008). Food access on the
other hand refers to whether a person has a socially recognized
claim on the available supply of food. It follows therefore that
owning productive assets for producing food and income both
play a role in enabling people to have access to food. Food
utilization depends on having adequate knowledge about how
to prepare food in a way that preserves its nutritional value and
to get it to those in the household who need it most.
*Julius Manda
International Institute of Tropical Agriculture (IITA),
Lilongwe, Malawi
Agricultural Economics and Rural Policy Group, Wageningen
University, Wageningen, The Netherlands
The International Maize and Wheat Improvement Center
(CIMMYT), Harare, Zimbabwe
The International Maize and Wheat Improvement Center
(CIMMYT), Nairobi, Kenya
Food Sec.
DOI 10.1007/s12571-015-0541-y
The above implies that adoption of improved agricultural
technologies can play an important role in reducing malnutrition.
Adoption of modern agricultural technologies such as improved
maize varieties has a positive and significant impact on crop
yields as well as household welfare (Alene et al. 2009; Becerril
and Abdulai 2010). Increased agricultural production through
adoption of improved maize varieties increases the income
earning opportunities for most poor households in rural areas,
thereby improving access to food. According to Headey
(2013), higher incomes raise expenditure levels on food, thereby
increasing the quality and quantity of diets. Furthermore, income
raises expenditure on nutrition-relevant non-food expenditures,
such as health, sanitation, electricity, water, and housing
The purpose of this paper is to analyse the determinants of
chronic malnutrition (stunting) and to evaluate the impacts of
improved maize varieties on stunting in eastern Zambia. The
paper uses household survey data from a sample of 810 house-
holds and applies the endogenous switching probit (ESP)
model to identify the determinants of child nutritional status
and impact of improved maize varieties. We complement our
using semi-parametric propensity score matching (PSM).
The paper adds to existing literature on child nutrition and the
nutritional impacts of improved agricultural technologies on mal-
nutrition. A number of studies have looked at the determinants of
child malnutrition in Africa (e.g. Christiaensen and Alderman
2004; Kabubo-Mariara et al. 2008;Masiyeetal.2010;
Ssewanyana 2003; Asenso-Okyere et al. 1997). However, to
our knowledge, none of the studies have tried to establish a
causal link between improved agricultural technologies such as
improved maize varieties and child malnutrition using rigorous
impact evaluation methods except Zeng et al. (2014). They used
Instrumental Variable (IV) methods to show that adoption of
improved maize varieties improves the nutritional status of chil-
dren in Ethiopia. One of the drawbacks of most IV methods is
that they only assume an intercept effect which may under- or
over-estimate the impacts of adoption. Zeng et al. (2014)also
assumed that the characteristics and resources of adopters and
non-adopters have the same impact on outcome variables (i.e.,
homogenous returns to their characteristics and resources). In this
study, we control for selection and endogeneity biases that may
potentially arise due to correlation between unobserved house-
hold characteristics and observed health outcomes using the ESP
approach. The ESP model estimates two separate equations for
adopters and non-adopters, thus allowing us to explore the dif-
ferential effects of the two groups.
The remainder of the paper is organized as follows. The
next section discusses child malnutrition in Zambia. The third
section outlines the conceptual and empirical frameworks
followed by a section presenting the data and descriptive sta-
tistics. The penultimate section presents the empirical results
and the last section draws conclusions.
Child malnutrition and adoption of improved maize
varieties in Zambia
Child malnutrition in Zambia
Child malnutrition rates in Zambia have long been high, but
there has been a noticeable increase in the past decade.
Although the burden of other infectious and preventable dis-
eases is high and contributes significantly to child morbidity
and mortality, nearly 52 % of all under 5 deaths in Zambia
are attributed to malnutrition (UNICEF 2008). There are several
factors that have been identified as causes of child malnutrition
in Zambia, including household food insecurity, lack of access
to health and other social services, especially among the poor
and rural population, poor nutrition of mothers and frequent
infections (Masiye et al. 2010; Sitko et al. 2011). Poverty
coupled with current rising food and fuel prices, scarcity of food
due to extensive crop loss owing to climate change effects such
as flooding, and in some cases lack of knowledge on proper
infant feeding practices further exacerbates the underlying
chronic nutrition problems (UNICEF 2008).
Tab le 1presents trends in the nutritional status of children in
Zambia using anthropometric data from the Zambia
Demographic Health Surveys (ZDHS) undertaken from 1992
to 2007 and the 2011 Zambia Human Development Report
(ZHDR). Inspection of Table 1shows that there was no consis-
tent trend in the nutritional indices for children under the age of
five over the past four ZHDS surveys (1992, 1996, 2002, and
2007). Wasting remained at roughly the same levels through-
out. During the period between 1992 and 2002, Zambia expe-
rienced an increasing trend in the malnutrition levels as
measured by stunting and underweight, coinciding with the
time that the country experienced some droughts and
unfavourable weather. However, the results of the 2007
ZDHS show a notable improvement in the nutritional status
of children as measured by both the height-for-age and
weight-for-age indices from the 2002 and 2007 ZDHS surveys.
Although there was a significant reduction in stunting (45 %)
and underweight (15 %) levels from 2002 to 2007, the stunting
rates were still high relative to the average prevalence of child
stunting of 39 % for 19 sub-Saharan African countries in the
mid-nineties (Morrisson et al. 2002). The 2007 ZDHS further
reveals that there were slightly more boys (48 %) than girls
(42 %) who were stunted. Results from all the demographic
health surveys show that the rural areas have more children
who are suffering from malnutrition than those in urban areas.
Among the nine provinces, eastern province has one of the
highest rates of malnutrition in Zambia at 50 %, third only to
central and Luapula provinces at 53 % and 59 %, respectively.
Tab le 1further show that the 2009 average stunting and under-
weight rates have started rising again, with stunting going
up from 45 % to 50 % and underweight going up from 15 %
to 19 %.
Manda J. et al.
Adoption of improved maize varieties in Zambia
Improved maize varieties wereintroducedtosmallholder
farmers in Zambia in the 1970s and almost 60 % of the
farmers have adopted these varieties to date (Kumar 1994;
Tem bo and Si t ko 2013). Improved maize varieties consist of
both hybrids and open pollinated varieties (OPVs). In simple
terms, hybrid maize results from the fertilization of one maize
plant by another genetically un-related plant (MacRobert et al.
2014) while OPVs are populations that breeders have selected
for a very specific set of traits and generally they can be
replanted up to three years without a decline in yields
(Becerril and Abdulai 2010). Over the past three decades,
more than 50 improved maize varieties have been developed
by the Zambia Agricultural Research Institute (ZARI) in col-
laboration with the International Maize and Wheat
Improvement Center (CIMMYT) and International Institute
of Tropical Agriculture (IITA) (Kalinda et al. 2014). The east-
ern province of Zambia is one of the largest producers of
maize in the country. For instance in the 20112012 season,
the province accounted for 21 % of the total maize produced
by small and medium scale farmers in Zambia (Tembo and
Sitko 2013), second only to the Southern Province which
contributed about 22 %.
Improved maize varieties have several advantages over local
varieties which include, but are not limited to: higher yields;
early maturation; uniform grain color and resistance to diseases.
Most of the improved varieties in Zambia have an estimated
yield advantage of 2060 % over locals (Howard and
Mungoma 1996). For instance one of the most popular varieties
in the eastern province of Zambia is MRI 634, which was re-
leased in 2000 through the Zambia Agricultural Research
Institute (ZARI). This is a medium maturing hybrid variety, with
dent white grains and a potential yield of 10 tons per ha.
Increased maize yields certainly play an important role in
increasing incomes and reducing poverty through the sale of
surplus maize. For example, recent studies in Zambia show that
improved maize varieties have significantly increased income
for adopters (Khonje et al. 2015; Smale and Mason 2014).
Although there is enough evidence on the productivity and
income effects of improved maize varieties, there is limited
evidence on the nutritional impacts on children under the age
of five.
Theoretical and empirical approaches
Theoretical framework
Figure 1shows the pathway through which agriculture is ex-
pected to affect child nutritional status. The figure shows that
there are two pathways through which adoption of improved
maize varieties could affect child nutritional status. It is expect-
ed that improved maize adoption will lead to an increase in
yields and consequently availability of more food for the house-
hold. On the other hand, improved maize adoption is expected
to increase household income through the sale of surplus maize,
which in turn translates into increased food expenditure on high
calorie and protein foods, finally leading to improvement in
child nutritional status (solid arrows in Fig. 1).
The other pathway involves the adoption of nutrition enhanc-
ing technologies, e.g. adoption of crops that are high in protein
content. Consumption of such crops is expected to increase the
intake of proteins which will translate into improved child nu-
tritional status. In this study, we envisage that adoption of im-
proved maize varieties will affect child nutritional status through
both the household income pathway and the diet composition
pathway. According to Dorosh et al. (2009), maize accounts for
about 60 % of the national calorie consumption and serves as the
dietary mainstay in central, southern, and eastern Zambia, hence
in addition to income, we believe that adoption of improved
maize varieties also serves as a proxy for food availability, pro-
viding the much needed calories and energy for children. The
supply of child nutrition is a complex process, and it may in-
volve multiple relationships, hence we cannot entirely rule out
the nutrition effects through the diet composition pathway.
The challenge in this study is to estimate the causal effect of
improved maize adoption on child nutrition (Fig. 1). One way
is to compare the stunting levels for children from improved
maize adopting and non-adopting households. However, just
comparing stunting levels between adopters and non-adopters
may be misleading, because there may also be differences in
e.g. access to resources, sanitation and health services. Without
controlling for these other factors the conclusions obtained
from this type of analysis may be false. One way to control
for other factors would be to regress the adoption variable on
the outcome variable (stunting) with variables such as access to
sanitation added as controls. However, because farmers often
Tabl e 1 Trends in the
malnutrition levels of under-five
children in Zambia, 19922009
Indicator 1992 (ZDHS) 1996 (ZDHS) 2002 (ZDHS) 2007 (ZDHS) 2009 (ZHDR)
Stunting 46 49 53 45 50
Wasting6565 -
Underweight 21 19 23 15 19
Note: ZDHS = Zambia Demographic Health Survey; ZHDR = Zambia Human Development Report
Source: UNZA, CSO and MII (1993, 2009), CSO, CBoH and ORC Macro (2003), UNDP (2011)
Determinants of child nutritional status in the eastern province
self-select into the adopter category or some technologies are
targeted to a given group of farmers, endogeneity problems
may arise which may lead to biased estimates (Alene and
Manyong 2007; Rao and Qaim 2011). Other methods such as
instrumental variable (IV) regression can be used to account for
endogeneity; however this method assumes technology adop-
tion has an average impact on child nutrition over the entire
sample of children, by way of an intercept shift in the child
nutrition production function. Other factors such as education
can also lead to an improvement in child nutrition by way of
slope shifts in the nutrition production function but are not
ential effects of the above aspects, two separate equations for
adopters and non-adopters have to be specified (Alene and
Manyong 2007). Interactions between improved maize adop-
tion and a set of explanatory variables at the same time account-
ing for endogeneity can only be effectively examined through
the simultaneous endogenous switching regression model.
The endogenous switching probit model
The modelling of the impact of adopting improved maize
varieties on child nutritional status using the ESP model pro-
ceeds in two stages. The first stage is the decision to adopt
improved maize varieties and it is estimated using a probit
model. In the second stage, a probit regression with selectivity
correction is used to examine the relationship between the
outcome variable (stunting) and a set of explanatory variables
conditional on the adoption decision.
The observed outcome of the improved maize varieties
adoption decision can be modelled in a random utility frame-
work. Following Aakvik et al. (2000), Heckman et al. (2001)
and Alene and Manyong (2007) let the adoption of the im-
proved maize varieties be a binary choice, where a farmer
decides to adopt improved maize varieties if the difference
between the utility of adopting and not adopting improved
maize varieties is positive. Let this difference be denoted
as I
is the utility obtained from adopting
improved maize varieties and U
the utility from not adopting
improved maize varieties. The farmer will adopt improved
maize varieties if I
is not observed, what is
observed is I, a binary indicator that equals one if a farmer
adopts improved and zero otherwise. More formally, the rela-
tionship can be expressed as;
Ii¼1if Ii*>0;
Ii¼0if Ii*0:
where Z is a vector of observed household and farm char-
acteristics determining adoption; αis the vector of unknown
parameters to be estimated; and ε
turbances related with the adoption of improved maize varie-
ties with mean zero and variance σ2
Following Lokshin and Sajaia (2011),thetwooutcomere-
gressions equations, conditional on adoption can be expressed as;
Regime 1 AdoptersðÞ:y1i¼β1X1iþu1iif Ii¼1
Regime 2 NonadoptersðÞ:y2i¼β2X2iþu2iif Ii¼0
where y
and y
is our outcome variable, viz. stunting; X
and X
are vectors of weakly exogenous covariates; β
and β
are vectors of parameters; and u
and u
are random distur-
bance terms.
For the ESP model to be identified, it is important for the Z
variables in the adoption model (eq. 1)tocontainaselection
instrument. We use distance to extension agentsoffice
(minutes) and sources of variety information (government ex-
tension (1 = yes) and non-governmental organization extension
(1 = yes)) as instrumental variables for the identification of the
impact of adoption on child nutrition. We envisage that farmers
are less likely to adopt improved maize varieties if they live far
from the office of the extension agents because the further away,
the more costs are incurred if the farmers are to access extension.
Similarly, information variables affect the decisions to adopt
improved agricultural technologies in Africa (Di Falco and
Ver o ne si 2013; Di Falco et al. 2011). We envisage that these
variables are correlated with the adoption of improved maize
varieties, but are unlikely to directly affect the nutritional status
of children. We follow Di Falco et al. (2011) in establishing the
admissibility of these instruments; if a variable is a valid instru-
ment, it will affect the decision to adopt, but it will not affect the
stunting levels of children among households that did not adopt.
The results
show that three variables can be considered as valid
Since the treatment and outcome variables are both binary, we used a
probit regression model to test validity of the instrumental variables. The
results from these tests are not discussed because of limited space but are
available on request
Adoption of
improved maize
protein and
Fig. 1 Pathways of impact of
agricultural interventions on child
nutritional status (adapted from
Masset et al. (2011))
Manda J. et al.
instruments because they are jointly statistically significant in
explaining the adoption decision [χ2 = 13.17 (p = 0.004)] but
are not statistically significant in explaining the outcome equa-
tion [χ2 = 5.61 (p = 0.133)].
The estimation of β
and β
above using a probit regression
may lead to biased estimates because of self-selection into the
adopter or non-adopter categories resulting from the non-zero
covariance between the error terms of the adoption decision equa-
tion and the outcome equation (Abdulai and Huffman 2014). The
error terms (u
,ε) are assumed to have a joint normal distri-
bution with mean vector zero and correlation matrix;
where ρ
and ρ
are the correlations between the error terms
and u
and ρ
is the correlation between of u
and u
assume that ρ
=1, since αis estimable only up to a scalar factor.
Estimation of average treatment effects
The endogenous switching probit model can be used to estimate
the average treatment effects on the treated (ATT) and the average
treatment effect of the untreated (ATU) by comparing the expect-
ed values of the outcomes of adopters and non-adopters in actual
and counterfactual scenarios. Following Aakvik et al. (2000), and
Lokshin and Sajaia (2011), we calculated the ATT and ATU
based on the expected outcomes, conditional on adoption:
Adopters with adoption (actual expectations observed in
the sample)
1ijI¼1;XðÞ ð4aÞ
Non-adopters without adoption (actual expectations
observed in the sample)
2ijI¼0;XðÞ ð4bÞ
Adopters had they decided not to adopt (counterfactual
expected outcome)
2ijI¼1;XðÞ ð4cÞ
Non-adopters had they decided to adopt (counterfactual
expected outcome)
1ijI¼0;XðÞ ð4dÞ
The average treatment effect on the treated (ATT) is com-
puted as the difference between (4a) and (4c);
2ijI¼1;XðÞ ð5Þ
The average treatment effect on the untreated (ATU) is
given by the difference between (4d) and (4b)
2ijI¼0;XðÞ ð6Þ
Previous studies that have used the ESP model include;
(Ayuya et al. 2015; Gregory and Coleman-Jensen 2013;
Lokshin and Glinskaya 2009).
The propensity score model
The ESP model can sometimes be sensitive to exclusion re-
striction assumptions, hence, to check the robustness of the
ESP results, we also estimated the ATTs using the propensity
score matching approach.
Following Becerril and Abdulai (2010) and Caliendo and
Kopeinig (2008), let Y
and Y
denote child stunting in
household ithat adopts an improved variety and the household
that does not adopt an improved variety, respectively. In real-
ity, only Y
or Y
are observed at one particular time and not
both. Let Trepresent a binary treatment variable that equals
one if a farmer adopts an improved variety and zero otherwise.
The observed stunting can be expressed as;
Yi¼TiYiA þ1Ti
ðÞYiN T¼0;1ðÞ ð7Þ
Furthermore, let Pbe the probability of observing a house-
hold with T= 1. The Average Treatment Effect (ATE) can be
expressed as follows;
The ATE is the weighted average effect of adoption on
the population, which is simply the difference of the ex-
pected outcomes after adoption and non-adoption
(Caliendo and Kopeinig 2008). However since the coun-
terfactual mean E(Y
|T= 1) is not observed, one has to
choose a proper substitute for it in order to estimate ATT
(Caliendo and Kopeinig 2008). According to Caliendo and
Kopeinig (2008), using the mean outcome of untreated
individuals E(Y
|T= 0) in non-experimental studies is usu-
ally not a good idea because it is most likely that compo-
nents which determine the treatment decision also deter-
mine the outcome variable of interest. To address this prob-
lem, the Propensity Score Matching (PSM) approach is
used. The propensity score is defined as the conditional
probability that a farmer adopts the new technology, given
pre-adoption characteristics (Rosenbaum and Rubin,
1983). The PSM employs the unconfoundedness assump-
tion also known as conditional independence assumption
(CIA) or selection on observables assumption. This as-
sumption implies that systematic differences in outcomes
between adopters and comparison individuals with same
values for covariates are attributable to adoption thereby
making adoption random and uncorrelated with the
Determinants of child nutritional status in the eastern province
outcome variables (Ali and Abdulai 2010; Caliendo and
Kopeinig 2008). The propensity score can be expressed as;
pXðÞ¼Pr T¼1jXðÞ¼ETjXðÞ;pXðÞ¼FhX
where Xis the multidimensional vector of pre-treatment
characteristics (same as Z in eq. 1above); and F{.} is the
cumulative distribution function. If the p(X) is known, then the
ATT can be estimated as follows:
iAYiN jT¼1
iAYiN jT¼1;pXðÞ
iN jT¼0;pXðÞ
where the outer expectation is over the distributionof ( p(X)
and Y
are the potential outcomes in the two
counterfactual situations of adoption and no adoption
Data, variable specification and descriptive statistics
Survey design and data collection
The data used in this paper come from a survey of 810 sample
households conducted in January and February 2012 in the
eastern province of Zambia. This was a baseline
survey con-
ducted by the International Institute of Tropical Agriculture
(IITA) and the International Maize and Wheat Improvement
Centre (CIMMYT) in collaboration with the Zambia
Agricultural Research Institute (ZARI) for the project entitled
Sustainable Intensification of Maize-Legume Systems for the
eastern province of Zambia (SIMLEZA). A survey question-
naire was prepared and administered by trained enumerators
who collected data from households through personal inter-
views. The survey was conducted in the same SIMLEZA
project districts in eastern Zambia Chipata, Katete, and
Lundazi which were targeted by the project as the major
maize and legume growing areas. In the first stage, each dis-
trict was stratified into agricultural blocks
(8 in Chipata, 5 in
Katete and 5 in Lundazi) as primary sampling units. In the
second stage, 41 agricultural camps were randomly selected,
with the camps allocated proportionally to the selected blocks
and the camps selected with probability of selection propor-
tional to size. Overall, 17 camps were selected in Chipata, 9 in
Katete and 15 in Lundazi. A total sample of 810 households
was selected randomly from the three districts with the num-
ber of households from each selected camp being proportional
to the size of the camp (Table 2).
The selected sample of 810 households was surveyed using
a semi-structured questionnaire. Of the 810 households, 444
households provided anthropometric data on 752 children
of ages 360 months. The weight of the children was
measured using a standard scale. The standing height as op-
posed to recumbent length was measured using a measuring
ruler, preferred mainly for ease of use. Table 3shows the total
number of 670 children who were considered in the analysis
since extreme or biologically implausible z-scores were re-
moved as recommended by Masiye et al. (2010)). The ex-
treme values for height-for-age z-scores (HAZ) were those
which were below 6orabove6.
Variable specifications in the outcome and selection
The dependent variable in our nutritional status model is
stunting representing children who have low height-for-age
z-score index, i.e. a z-score below 2. Stunting is preferred
to the weight-for-height z-scores (WHZ) and weight-for-age
z-scores (WAZ) indices because it represents the prevalence of
long-term growth failure. The WHZ is a condition that usually
reflects severely inadequate food intake and infection happen-
ing at present and, as such, it is recommended that WHZ
should be regressed on flow and not stock variables
(Christiaensen and Alderman 2004). Weight-for-age on the
other hand is a compound measure of height-for-age and
weight-for-height which reflects body mass relative to age
and thus making interpretation difficult (ODonnell et al.
A follow up survey will be conducted in 2015 where the same house-
hold who were interviewed at baseline will be interviewed.
A camp is a catchment area made up of 8 different zones consisting of
villages and is headed by an agricultural camp officer. A block is made up
of camps and is managed by an agricultural block officer.
Tabl e 2 Distribution of sample households by district and gender
District Number of
Number of
Female-headed Male-headed All
Chipata 8 17 129 205 334
Katete 5 9 63 117 180
Lundazi 5 14 98 198 296
All 18 40 290 520 810
Tabl e 3 Distribution of
sample children by
district and gender
Gender of Child
District Female Male All
Chipata 140 137 277
Katete 51 64 115
Lundazi 161 117 278
All 352 318 670
Manda J. et al.
Child characteristics
The explanatory variables relate to child, household, community
and agricultural characteristics. Child level covariates include
gender, age and whether or not the child had suffered from diar-
rhea in the past year. Some evidence from previous studies in
Sub-Saharan Africa shows that boys are more likely to be stunted
than girls (Ojiako et al. 2009; Sanginga et al. 1999; Svedberg
1990). However, some studies in Asia (e.g. Kumar et al. 2006)
show that girls are more stunted than boys; hence the impact of
gender on stunting is indeterminate. Age of the child is an im-
portant determinant of the physiological characteristics which
convert consumption into nutrition and nutrition into higher pro-
ductivity and, therefore, higher earning potential (Sarmistha
1999). Younger children are expected to have better nutritional
status than the older children following commonly observed pat-
terns in developing countries, explained by better child care and
better feeding practices for younger children and exposure of
older children to relatively harsh environments (Sanginga et al.
1999). Illness of a child is hypothesised to negatively affect child
nutrition. Diarrhea (proxy for illness) is expected to be inversely
related to child nutritional status because it causes nutrients to
flush through the intestinal tract too quickly to be absorbed
(Apodaca 2008). A repeatedly sick child may not consume ade-
quate levels of food, which can result in growth retardation.
Household characteristics
Household characteristics include: age of the household head;
gender of the household head; marital status of the household
head; household size; education of the household head; highest
grade attained by the most educated female of the household;
number of household members above 65 years; number of
household members below 15 years; number of adult females
in the household (1665 years old); household assets; coopera-
tive membership (group membership); kinship and political con-
nections. The gender of the household head is measured by a
dummy variable equal to one for male headed households and
zero for female headed households. Men are generally believed
to be less involved than women in taking care of children and
providing for their familiesfood needs (Onyango et al. 1994).
However, past studies have also shown that female headed
households are usually poor relative to their male counterparts
and therefore expenditure on child related nutrition is expected to
be less than in male headed households. We therefore expect the
sign on the gender of the household head to be either negative or
positive. Similarly we expect the marital status of the household
to have a positive effect on the nutritional status of children
because children will have good care as both parents can take
turns in looking after the child. Parental education is assumed to
have a direct positive link to child nutrition through better child-
care practices and resource allocation in the household.
Education affects care giving practices through the ability to
acquire skills and the ability to model behaviour (Chirwa and
Ngalawa 2008). In addition, to account for potential intra-
household externalities from education, which are especially im-
portant in households at low education levels (Christiaensen and
Alderman 2004), we posit that the presence of educated female
household members will have a positive effect on child nutrition-
al status. It is assumed that household members who at least
completed primary school are in a better position to comprehend
and apply information related to childrenshealth.
Information gleaned from the literature shows that large fam-
ily sizes impact negatively on nutritional status and household
welfare in that the percentage of children under five, relative to
total household size, reflects the burden of care in terms of nutri-
tion finance, and parental time, and thus affects nutrition out-
comes (Ajieroh 2009). Household assets are often used as a
proxy for household wellbeing or resources and some studies
have shown that it is a positive determinant of child nutritional
outcomes (Kabubo-Mariara et al. 2008). Greater assets at house-
hold level allow people to spend more on important aspects of
child nutrition such as health care, hygiene, food and clean water
(Alderman et al. 2005). We also expect the nutritional status to
reduce with an increase in the number of household members
below 15 years and above 65 (dependants) because with an
increase in the number of dependants, we expect a greater burden
on household resources for food consumption.
Group membership, kinship (number of relatives) and the num-
ber of relatives or friends in leadership positions (political connec-
tions) represent the household social networks. Previous studies
have shown that cooperative group membership indicates the in-
tensity of contacts with other farmers (Adegbola and Gardebroek
2007) hence we expect farmers who are members of a group to
have more information on improved maize varieties. Membership
is therefore hypothesized to be positively associated with better
child nutrition. Households with more relatives are more likely to
have children who are better nourished as the household may have
relatives they can rely on for critical support. However, an increase
in the number of relatives may also come at the expense of income
growth, which may negatively affect the the nutritional status of
children. Therefore the sign on kinship is indeterminate. Similarly
we expect households with political connections to have children
who are well nourished as they can obtain support from their
influential relatives/friends in times of problems.
Agricultural characteristics
To capture farm characteristics, we included adoption
of im-
proved maize varieties, total land cultivated and distance to
the nearest market. Adoption of improved maize varieties is
expected to improve the nutritional status of children by
An adopter in this study is defined as any farmer who planted or allo-
cated land to at least one improved maize variety consistently for the past
three years prior to the survey
Determinants of child nutritional status in the eastern province
promoting a link between food security and nutrition security
(World Bank 2008). Adoption of improved maize varieties leads
to higher yields which in turn improves the food security status of
farmers as well as increased income through sale of surplus food.
The demand for productive agricultural land has been growing,
partly due to the growing population in many developing coun-
tries. The more arable land under permanent crops or pastures,
the more food there is and this in turn allows greater access to
nutrition by increasing the availability of food (Apodaca 2008).
Distance to the nearest market reflects the transaction costs that
the household incurs, such that the greater the distance, the higher
the costs. We therefore expect distance to the nearest market to be
negatively related with the nutritional status of the child.
Community characteristics
Sanitary conditions in the community are usually reflected in
the percentage of households using toilets and the percentage
of households who have access to safe drinking water from
taps and deep, well protected wells. Access to good toilet and
safe drinking water facilities is expected to affect nutrition in a
positive way as some studies have shown (Glewwe et al.
2002; Christiaensen and Alderman 2004; Chirwa and
Ngalawa 2008). Access to good sanitation may prevent the
occurrence of infectious diseases such as diarrhea, dysentery
and cholera which can adversely affect child nutrition.
The distance to the health centre approximates the avail-
ability and costs of health services; therefore we expect the
distance to the nearest health centre to be inversely related to
the child nutritional status.
Factors that are hypothesised to affect adoption of im-
proved maize varieties include household and social network
characteristics mentioned above. For a detailed description of
the hypothesized relationships between adoption and the
variables used in the selection equation see Feder et al.
(1985) and Kassie et al. (2013).
Socioeconomic characteristics of the sample households
Tab le 4presents the characteristics of households in eastern
Zambia. Considering all three districts, i.e. Chipata, Katete
and Lundazi, on average 56 % of the children were stunted,
with Katete having slightly more with 57 %. Table 4further
shows that about 23 % were severely stunted, with Lundazi
having the largest percentage of 26 % of the severely stunted
children. The average stunting rate (56 %) for the three dis-
tricts was higher than the average for the eastern province of
Zambia (50 %) partly because we only considered three dis-
tricts out of the 9 districts available in the province. Table 4
further shows that in our sample, about 53 % of the children
were girls with Lundazi having the highest number of almost
60 %. The results also show that the average age of the chil-
dren in the sample was 33 months and at least 60 % of the
children had diarrhea the year preceding the survey. Lundazi
had the highest number of children who had diarrhea with
73 %, followed by Katete with 52 % and this could be one
of the reasons as to why these districts had relatively higher
percentages of stunted children compared to Chipata district.
The average household size was 7.3 persons and across dis-
tricts it ranged from 8 persons in Lundazi to 6.6 persons in Katete
with Chipata having 7.4 persons per household. At national level,
Tabl e 4 Mean values of social economic characteristics of the sample
District Chipata Katete Lundazi All
Child characteristics
Normal stunting (> 2) 0.53 0.57 0.56 0.56
Moderate stunting (3to2) 0.14 0.10 0.14 0.13
Severe stunting (< 3) 0.20 0.22 0.26 0.23
Child age (months) 31.98 33.38 33.98 33.11
Had diarrhea in the past
one year (1 = yes)
0.52 0.54 0.73 0.60
Gender (1 = male) 0.49 0.56 0.42 0.49
Household characteristics
Gender of household head
(1 = male)
0.68 0.70 0.71 0.69
Marital status (1 = married) 0.85 0.87 0.92 0.89
Total household size (number) 7.38 6.56 7.99 7.31
Household completed primary
school (1 = yes)
0.70 0.71 0.52 0.63
Asset per capita (000` ZMK) 1.06 1.34 1.30 1.23
Highest grade completed by most
educated female (years)
6.68 6.30 7.47 6.94
Highest grade of most educated
male (years)
7.55 6.83 8.66 7.67
Number of adult females in the
household (1665 years old)
1.65 1.50 1.89 0.17
Number of household members
above 65 years
0.21 0.14 0.22 0.20
Number of household members
below 15 years
Kinship (number of relatives) 4 4 3 4
Household has political
connections (1 = yes)
0.66 0.60 0.62 0.63
Group membership (cooperative)
(1 = yes)
0.91 0.83 0.96 0.92
Agricultural characteristics
Total cultivated land (ha) 3.22 3.37 5.38 4.14
Adoption of improved maize
varieties (1 = adopted)
0.11 0.14 0.22 0.15
Distance to nearest market
441 237 450 410
Community characteristics
Distance to the nearest health
center (minutes)
61.09 72.62 85.65 73.26
Access to toilets (sanitation)
(1 = yes)
0.21 0.28 0.13 0.21
Access to safe water (1 = yes) 0.18 0.10 0.18 0.15
ZMK = Zambian Kwacha
Manda J. et al.
the average household size in Zambia in 2010 was 5.2 persons
(CSO 2012), lower than the average in Table 4. Inspection of
Tab le 4reveals that most of the household heads completed
primary school education with an average of 63 %. To control
for household resources, we included total household assets per
capita. On average, the value of assets for the households was
about ZMK1.23 million (US$236),
with Katete having the
highest with ZMK1.34 million (US$258). Households in
Chipata on the other hand had the lowest assets per capita with
a total asset value of ZMK1.06 million (US$204). Most of the
farmers belonged to a cooperative group with an average of
about 93 %, with Lundazi having the highest percentage of 96 %.
As stated earlier, agriculture is a major source of livelihood
and a key determinant of food security in rural areas. On average
about 14 % of the households adopted improved maize varieties,
with Lundazi having the largest percentage of 20 %. The own-
ership of land by households is an indicator of the households
ability to withstand economic shocks and is also commonly used
as a proxy for household income. Chipata had the lowest area of
cultivated land per capita (3.22 ha) while Lundazi had the highest
with 5.38 ha. One of the reasons why Chipata had the lowest
cultivated land is that among the three districts, Chipata is the
most densely populated district and hence there is more pressure
on the land. According to CSO (2012), Chipata district contrib-
uted about 27 % to the population of the eastern province of
Zambia, which was the largest amongst the three districts.
Lundazi, on the other hand, is sparsely populated and therefore
most farmers own relatively large pieces of land. However,
owning large pieces of land may not necessarily translate into
higher incomes as in the case of Lundazi, because it may also
have to do with the quality and the capacity to work the land.
Tab le 4also shows that, on average, 28 % of the house-
holds had access to toilet facilities in Katete and only 13 % in
Lundazi. Similarly, Chipata and Lundazi had the highest pro-
portion of farm households who had access to drinking water
with 18 %. This is plausible because Chipata and Lundazi are
relatively more urban than the Katete district.
in Table 5.WHO(1995)recommendsthatatleasttwoage
disaggregations be used, under 24 months and 24 months and
over. The reason is that patterns of growth failure vary with
age and the identification of determinants of malnutrition is
facilitated. More girls (55 %) in the 023 age category were
stunted than boys (36 %). Overall, the results show that the
scourge of malnutrition affects older children (60 %) more
than younger ones (47 %). This finding is consistent with
other studies on the nutritional status of children in Africa
(e.g. Ssewanyana 2003;Kabubo-Mariaraetal.2008).
Table 6shows the relationship between adoption of im-
proved maize varieties and child stunting. Non-adopting house-
holds had more children who were stunted (57 %) than those
who adopted improved maize varieties (51 %). This may imply
that improved maize adoption has an effect on child stunting,
although we may not make a causal inference at this stage.
Empirical results
Determinants of child malnutrition
The estimated parameters for the endogenous switching probit
(ESP) model, revealing the factors that affect child nutritional
status, are presented in Table 7. Estimates for the first stage
regression for the determinants of improved maize adoption
are presented in the Appendix in Table 10 and to conserve
space, the results will not be discussed here.
Child, household and community characteristics have differ-
ential impact among adopters and non-adopters (Table 7).
Among the child characteristics, only age (for non-adopters)
and diarrhea (for adopters) are important determinants of long-
term child malnutrition. Similar to the descriptive results above,
the results in Table 7show that the probability of stunting in-
creases with the age of the child among the non-adopters of
improved maize varieties. As children grow older, weaning and
less breast milk may make them more vulnerable to malnutrition
(Kabubo-Mariara et al. 2008). It may also suggest that as children
grow older, less attention is given to them by their parents in
terms of health care, the food they eat, and the nutritional value of
the food. Similarly, children who suffered from diarrhea the
previous year before the survey were more stunted than those
who did not and this is in line with our theoretical expectations.
Food consumed by children suffering from diarrhea does not
result in any meaningful nutrition for the child as nutrients flush
through the intestinal tract too quickly to be absorbed.
In line with previous studies on child malnutrition (e.g.
Kabubo-Mariara et al. 2008) parental education reduced the
probability of stunting by as much 75 %. Similar to the results
of Christiaensen and Alderman (2004) the presence of educated
Exchange rate at the time of the survey: 1US$ = ZMK5,1974
Tabl e 5 Child stunting
by age and gender in
eastern Zambia
Age (months) Male Female All
023 0.38 0.55 0.47
2460 0.61 0.60 0.60
Tabl e 6 Child stunting by household adoption status and gender of
Gender of child
Adoption status Male Female All
Adopters 0.42 0.61 0.51
Non-adopters 0.58 0.56 0.57
All 0.55 0.56 0.56
Determinants of child nutritional status in the eastern province
female adults in a household also had a significant correlation
with the probability of stunting amongst children from adopting
households. The probability of being stunted reduces by 16 %
with each additional year of schooling for the most educated
female household member among adopters. This shows that ed-
ucated females play an important role in sharing knowledge re-
lated to childrens health such as good child care practices and the
ability to recognize illness. Presence of adult females in the
household has a negative effect on the probability of stunting
amongst non-adopters, implying that there is knowledge transfer
related to child care from elderly to young mothers which in turn
benefits the nutrition of the children. Contrary to theoretical ex-
pectations, the results also show that marriage was not beneficial
to the nutritional status of children among non-adopters. This
may have to do with the age at which the mothers got married.
Early marriages and age of the mother have been linked with
reduced nutritional outcomes for children (Raj et al. 2010;
Kabubo-Mariara et al. 2008). This is so because young mothers
may have low educational attainment and may be physically
immature, and socially and economically unstable (Bwalya et
al. 2015), all of which are associated with child malnutrition.
Consistent with our theoretical expectations, household heads
who were members of a cooperative were associated with better
child nutrition. Similarly, the probability of stunting increased by
between 5 and 2 % among adopters and non-adopters, respec-
tively, with kinship. This is so because the more relatives a
household has, the more the pressure on household resources
which may in turn result in poor nutrition especially among
Amongst the community variables, access to sanitation had
a negative and significant effect on stunting among non-
adopters. This can be partly attributed to the fact that with an
improvement in sanitation, the elimination of parasites that
cause infections such as diarrhea and dysentery is facilitated.
Impact of improved maize adoption on child malnutrition
The estimates for the average treatments effects (ATT), which
show the impact of adoption on stunting after accounting for
both observable and unobservable characteristics, are presented
in Table 8. Both adopters and non-adopters benefit from adop-
tion. Specifically, the probability of stunting for children from
adopting households would be 26 % greater had the households
not adopted improved maize varieties. This is the average treat-
ment effect on the treated (ATT) which is statistically significant
at the 1 % confidence level. Similarly, the probability of
stunting for children from non-adopting households would be
33 % less had the household adopted improved maize varieties,
implying that non-adopting households would have realized
lower rates of stunting from switching to improved maize vari-
eties under the given conditions. This is the average treatment
effect on the untreated (ATU) which is also statistically signif-
icant and implies that children from non-adopting households
would be better off if their parents were to adopt improved
maize varieties (as opposed to local varieties).
The results from the ESP model above may be sensitive to
the exclusion restriction assumption; hence we also used the
PSM approach to check the robustness of the estimated effects
Tabl e 7 Determinants of child malnutrition in eastern Zambia
Variable Adopters
Coefficient Coefficient
Age in months 0.01 (0.89) 0.01 (3.51)***
Gender of child 0.65 (0.82) 0.04 (0.30)
Child had diarrhea 0.87 (2.17)** 0.12 (0.98)
Ln distance to health center 0.18 (0.59) 0.07 (1.17)
Age of household head 0.00 (0.15) 0.00 (0.79)
Number of elderly (>65 years) 0.25 (0.20) 0.16 (0.99)
Number of children (<15 years) 0.14 (0.31) 0.05 (0.61)
Household completed primary
school (1 = yes)
0.75 (1.86)* 0.14 (1.05)
Gender of household head 0.35 (0.91) 0.07 (0.49)
Household size 0.04 (0.12) 0.08 (1.16)
Ln assets per capita 0.25 (0.72) 0.05 (0.79)
Highest grade completed by most
educated female
0.16 (1.67)* 0.01 (0.51)
Number of adult females
(1665 years old)
0.44 (1.18) 0.23 (2.38)**
Married 1.47 (1.42) 0.42 (2.09)**
Group membership 2.29 (2.17)** 0.09 (0.44)
Kinship 0.05 (2.04)** 0.02 (1.87)*
Political connections 0.28 (0.39) 0.02 (0.11)
Total land cultivated 0.03 (0.47) 0.01 (0.51)
Ln distance to nearest
village market
0.20 (0.90) 0.02 (0.48)
Access to sanitation 0.99 (1.59) 0.70 (4.78)***
Access to safe water 0.09 (0.21) 0.04 (0.35)
Chipata district dummy 0.23 (0.38) 0.06 (0.38)
Lundazi district dummy 0.13 (0.17) 0.05 (0.28)
Constant 1.40 (0.13) 1.03 (1.19)
Diagnostic tests
Wald test χ
Note *, ** and *** denotes significance level at 10 %, 5 % and 1 % (t-
ratio in parenthesis); Ln =Natural logarithm
Table 8 Impact of improved maize varieties on child malnutrition
(endogenous switching probit results)
Mean of outcome variable Treatment effect Average treatment
effects (ATE)
Stunting Farm households that
adopted (ATT)
0.26 (4.52)***
Farm households that
did not adopt (ATU)
0.33 (17.10)***
Note *** denotes significance level at 1 % (t-ratio in parenthesis)
Manda J. et al.
obtained from the ESP model. The same variables were used
in the estimation of propensity scores as those reported in
Table 10. We followed the rule of Augurzky and Schmidt
(2001), and Brookhart et al. (2006), for quality implementa-
tion of propensity score estimation.
A visual inspection (Fig. 2) of the density distributions of
the estimated propensity scores for the two groups indicates
that the common support condition is satisfied: there was a
substantial overlap in the distribution of the propensity scores
of both adopter and non-adopter groups. The bottom half of
the graph shows the distribution of propensity scores for the
non-adopters and the upper half refers to the adopters.
Tab le 9provides the ATTestimates from the PSM approach.
The effect of improved maize varieties on stunting was estimat-
ed with the Nearest Neighbour (NNM) and the bias-adjusted
NNM estimator developed by Abadie and Imbens (2011).
Similar to the ESP results, adoption of improved maize varieties
significantly reduces the probability of stunting. The causal
effects from NNM approaches generally indicate that adoption
of improved maize varieties exerts a negative and significant
effect on stunting. Table 9shows that on average, children from
non-adopting households were relatively more stunted (62
63 %) than those from adopting households (51 %).
Consistent with the ESP results reported in Table 8,thePSM
results suggest that adoption of improved maize varieties sig-
nificantly reduces the probability of stunting in the range of 11
12 %. Compared to the ESP results, the estimated effects from
the PSM approach are relatively lower, probably because the
latter does not take into account the selection on unobservables.
Conclusion and implications
This paper analyses the factors that affect the nutritional status
of under-five children as well as the impact of improved maize
varieties on child stunting in Zambia using household survey
data from a sample of 810 households in the eastern province
of Zambia. Given the non-experimental nature of the data
used in the analysis, a combination of parametric and non-
parametric econometric methods was used to mitigate biases
resulting from both observed and unobserved characteristics.
Empirical results show that child malnutrition is a function of
the childs age and gender, gender of the household head, edu-
cation of female household members, number of adult females in
the household, and access to sanitation. The results are largely
consistent with findings from other malnutrition studies (e.g.
Christiaensen and Alderman 2004; Kabubo-Mariara et al. 2008).
Average treatment effects from both the ESP and PSM anal-
ysis show that adoption of improved maize varieties significantly
reduced the prevalence of stunting. The ESP results show that
farm households that adopted benefited more from adoption.
Probability of stunting for children from adopting households
was reduced by as much as 26 %. The probability of stunting
would have also reduced by about 33 % for children from non-
adopting households, if the households had adopted improved
maize varieties, suggesting that non-adopting households would
have realized lower rates of stunting from switching from grow-
ing local to improved maize varieties. Results from the matching
estimates show that the probability of stunting also reduced
among children from adopting households.
The results stress the key role of adoption of improved maize
varieties in improving the income earning opportunities for
rural households in order to fight the scourge of malnutrition.
However, realizing the full benefits of improved technologies
such as improved maize varieties in terms of improved income
earning opportunities and food security will require increased
investment and policy support aimed at enhancing technology
0 .2 .4 .6 .8
Propensity Score
Treated: On supportUntreated
Treated: Off support
Fig. 2 Propensity score distribution and common support for propensity
score estimation. Note:BTreated: on support^indicates the observations
in the adoption group that have a suitable comparison. BTreated: off
support^indicates the observations in the adoption group that do not
have a suitable comparison
Tabl e 9 Impact of improved maize varieties on child malnutrition (matching results)
Matching Algorithm Outcome variable Means of outcome variables ATT difference
Adopters Non Adopters
Nearest Neighbor Matching Stunting 0.51 0.63 12 (1.74) *
Bias adjusted Nearest Neighbor Matching Stunting 0.51 0.62 11 (2.03)**
Note * and ** denotes significance level at 10 % and 5 % (t-ratio in parenthesis)
Determinants of child nutritional status in the eastern province
adoption by farmers. Secondly, the significance of education in
reducing child stunting suggests that the assimilation of nutri-
tional messages may require morethanbasiceducationtobe
more effective. Promoting education among females is thus
critical for nutrition-enhancing child care practices.
Acknowledgments The authors gratefully acknowledge financial sup-
port from the USAID Zambia Mission (USAID/Zambia). The household
survey was conducted in collaboration with the Ministry of Agriculture
and Livestock of Zambia and the Zambia Agricultural Research Institute
(ZARI). We thank Bernadette Chimai of the University of Zambia who
ably supervised the data collectionprocess. We are grateful to three anon-
ymous referees and the Editor-in-Chief of this journal for their comments
on an earlier draft of the paper.
Aakvik, A., Heckman, J., and Vytlacil, E. (2000). Treatment effect for
discrete outcomes when responses to treatment vary among obser-
vationally identical persons: An application to Norwegian vocation-
al rehabilitation programs., in: Research, N.B.o.E. (Ed.), Technical
Paper 262, Cambridge
Abadie, A., & Imbens, G. W. (2011). Bias-corrected matching estimators
for average treatment effects. Journal of Business and Economic
Statistics,29(1), 111.
Abdulai, A., & Huffman, W. (2014). The adoption and impact of soil and
water conservation technology: An endogenous switching regres-
sion application. Land Economics,90(1), 2643.
Adegbola, P., & Gardebroek, C. (2007). The effect of information sources
on technology adoption and modification decisions. Agricultural
Ajieroh, V. (2009). A quantitative analysis of determinants of child and
maternal malnutrition in Nigeria. Background Paper No. NSSP10.
Alderman, H., Hoogeveen, H., and Rossi, M. (2005). Reducing child
malnutrition in Tanzania: Combined effects of income growth and
program interventions. World Bank Policy Research Working Paper
Alene, A. D., & Manyong, V. M. (2007). The effects of education on
agricultural productivity under traditional and improved technology
in northern Nigeria: An endogenous switching regression analysis.
Empirical Economics,32(1), 141159.
Alene, A. D., Menkir, A., Ajala, S. O., Badu-Apraku, B., Olanrewaju, A.
S., Manyong, V. M., et al. (2009). The economic and poverty im-
pacts of maize research in west and central Africa. Agricultural
Economics,40(5), 535550. doi:10.1111/j.1574-0862.2009.00396.
Ali, A., Abdulai, A. (2010). The adoption of genetically modified cotton
and poverty reduction in Pakistan. Journal of Agricultural
Economics, 61,175192
Apodaca, C. (2008). Preventing child malnutrition: Health and agricul-
ture as determinants of child malnutrition. Journal of Children and
Poverty,14(1), 2140.
Asenso-Okyere, W. K., Asante, F. A., & Nube, M. (1997). Nutritional
status of children in Ghana. Agricultural Economics,17,5974.
Augurzky, B., & Schmidt, C.M. (2001). The propensity score: a means to
an end. Discussion Paper No. 271, IZA.
Ayuya, O. I., Gido, E. O., Bett, H. K., Lagat, J. K., Kahi, A. K., & Bauer,
S. (2015). Effect of certified organic production systems on poverty
among smallholder farmers: empirical evidence from Kenya. Worl d
Becerril, J., & Abdulai, A. (2010). The impact of improved maize varie-
ties on poverty in Mexico: A propensity score-matching approach.
World Development,38(7), 10241035.
Brookhart, M. A., Schneeweiss, S., Rothman, K. J., Glynn, R. J., Avorn,
J., & Stürmer, T. (2006). Variable selection for propensity score
models. American Journal of Epidemiology,163, 11491156.
Bwalya, B. B., Lemba, M., Mapoma, C. C., & Mutombo, N. (2015).
Factors associated with stunting among children aged 623 months
in zambian: evidence from the 2007 Zambia demographic and
health survey. International Journal of Advanced Nutritional and
Health Science,3,116131.
Caliendo, M., Kopeinig, S. (2008). Some practical guidance for the im-
plementation of propensity score matching. Journal of Economic
Surveys, 22,3172.
Chirwa, E. W., & Ngalawa, H. (2008). Determinants of child nutrition in
Malawi. South African Journal of Economics,76(4), 628640.
Christiaensen, L., & Alderman, H. (2004). Child malnutrition in Ethiopia:
Can maternal knowledge augment the role of income. Economic
Development and Cultural Change,52(2), 287312.
Tabl e 1 0 Probit estimates of determinants of adoption of improved
maize varieties in eastern Zambia
Variable Coefficient
Age in months 0.01 (1.68)*
Gender of child 0.29 (2.10)**
Child had diarrhea 0.19 (1.30)
Ln distance to health center 0.1 (153)
Age of household head 0.00 (0.45)
Number of elderly (>65 years) 0.47 (2.77)**
Number of children (<15 years) 0.23 (2.46)**
Household head completed primary school (1 = yes) 0.12 (0.77)
Gender of household head 0.06 (0.37)
Household size 0.17 (2.40)**
Ln assets per capita 0.19 (3.25)***
Highest grade completed by most educated female adult 0.02 (0.84)
Number of adult females in the household (1665 years
0.14 (1.30)
Married 0.01 (0.02)
Group membership 0.02 (0.08)
Kinship 0.01 (1.26)
Political connections 0.37 (2.39)**
Total land cultivated 0.04 (2.23)**
Ln distance to nearest village market 0.02 (0.35)
Access to sanitation 0.07 (0.35)
Access to safe water 0.05 (0.34)
Chipata district dummy 0.08 (0.37)
Lundazi district dummy 0.11 (0.56)
Access to NGO extension 0.13 (0.96)
Access to government extension 0.22 (1.30)
Distance to extension office 0.00 (2.92)**
Constant 4.60
Note *,**, and *** denotes significance level at 10 %, 5 % and 1 %
Manda J. et al.
CSO. (2012). The monthly. Zambia Central Statistics Office. Lusaka
Di Falco, S., & Veronesi, M. (2013). How can African agriculture adapt to
climate change? A counterfactual analysis from Ethiopia. Land
Economics,89(4), 743766.
Di Falco, S., Veronesi, M., & Yesuf, M. (2011). Does adaptation to cli-
mate change provide food security? A micro-perspective from
Ethiopia. American Journal of Agricultural Economics,93(3),
Dorosh, P. A., Dradri, S., & Haggblade, S. (2009). Regional trade, gov-
ernment policy and food security: Recent evidence from Zambia.
Food Policy,34(4), 350366.
Feder, G.R., Just, R. E., & Zilberman, D. (1985). Adoption of agricultural
innovations in developing countries: a survey. Economic
Development and Cultural Change,33,255298.
Glewwe, P., Koch, S., & Nguyen, B. L. (2002). Child nutrition, economic
growth and the provision of health care services in Vietnam in the
1990s. In Policy research working paper 2838. Washington, DC:
World Bank.
Gregory, C. A., & Coleman-Jensen, A. (2013). Do high food prices in-
crease food insecurity in the United States? Applied Economic
Perspectives and Policy,35,679707.
Headey, D. D. (2013). Developmental drivers of nutritional change: a
cross-country analysis. Wo r l d D e v e lop me nt,42,7688.
Heckman, J., Tobias, J. L., & Vytlacil, E. (2001). Four parameters of
interest in the evaluation of social programs. Southern Economic
Horton, S., Alderman, H., and Rivera, J. A. (2008). The challenge of
hunger and malnutrition. Copenhagen Consensus.Copenhagen
Howard, J., & Mungoma, C. (1996). Zambiasstopandgo revolution:
the impact of policies and organizations on the development and
spread of maize technology. Department of Agricultural
Economics, East Lansing: Michigan State University.
Determinants of childrens nutritional status in Kenya: Evidence
from demographic and health surveys. Journal of African
Economies,18(3), 363387.
Kalinda, T., Tembo, G., & Kuntashula, E. (2014). Adoption of improved
maize seed varieties in southern Zambia. Asian Journal of
Agricultural Sciences,6,3339.
Kassie, M., Jaleta, M., Shiferaw, B., Mmbando, F., & Mekuria, M.
(2013). Adoption of interrelated sustainable agricultural practices
in smallholder systems: evidence from rural Tanzania.
Technological Forecasting and Social Change,80,525540.
Khonje, M., Manda, J., Alene, A. D., & Kassie, M. (2015). Analysis of
adoption and impacts of improved maize varieties in eastern
Zambia. Wo rld Devel o p m e n t,66,695706.
Kumar, S.K. (1994). Adoption of hybrid maize in Zambia: effects on
gender roles, food consumption, and nutrition (Research Report
100). Washington, DC: International Food Policy Research Institute.
Kumar, D., Goel, N., Mittal, P. C., & Misra, P. (2006). Influence of infant-
feeding practices on nutritional status of under-five children. The
Indian Journal of Pediatrics,73(5), 417421.
Lokshin, M., & Glinskaya, E. (2009). The effect of male migration on
employment patterns of women in Nepal. The World Bank
Economic Review,23,481507.
Lokshin, M., & Sajaia, Z. (2011). Impact of interventions on discrete
outcomes: maximum likelihood estimation of the binary choice
models with binary endogenous regressors. Stata Journal,11,
MacRobert, J., Setimela, P., Gethi, J., and Regasa, M.W. (2014). Maize
hybrid seed production manual
Masiye, F., Chama, C., Chitah, B., and Jonsson, D. (2010). Determinants
of child nutritional status in Zambia: An analysis of a national sur-
vey. Zambia Social Science Journal,1.No.1,Article4
Masset, E., Haddad, L., Cornelius, A., and Isaza-Castro, J. (2011). A
systematic review of agricultural interventions that aim to improve
nutritional status of children
Morrisson, C., Thorbecke, E., Guilmeau, H., and Linkens, C. (2002).
Poverty and malnutrition in Sub-Saharan Africa. Cornell
ODonnell, O., Doorslaer, E.V., Wagstaff, A., and Lindelow, M. (2008).
Analyzing health equity using household survey data: a guide to
techniques and their implementation. WBI Learning Resources
Ojiako, I. A., Manyong, V. M., & Ikpi, A. E. (2009). Determinants of
nutritional status of preschool children from rural households in
Kaduna and Kano States, Nigeria. Pakistan Journal of Nutrition,
8(9), 14971505.
Onyango, A., Tucker, K., & Eisemon, T. (1994). Household headshipand
child nutrition: a case study in western Kenya. Social Science and
Medicine,39(12), 16331639.
Raj, A., Saggurti, N., Winter, M., Labonte, A., Decker, M. R., Balaiah,
D., et al. (2010). The effect of maternal child marriage on morbidity
and mortality of children under 5 in India: cross sectional study of a
nationally representative sample. BMJ,340, b4258.
Rao, E. J. O.,& Qaim, M. (2011). Supermarkets, farm household income,
and poverty: insights from Kenya. World Development,39(5), 784
Rosenbaum, P.R., Rubin, D.B. (1983). The central role of the propensity
score in observational studies for causal effects. Biometrika, 70,41
Sanginga, P. C., Adesina, A. A., Manyong, V. M., Otite, O., and Dashiell,
K. E. (1999). Social impact of soybean in Nigerias southern Guinea
savanna. International Institute of Tropical Agriculture (IITA),
Sarmistha, P. (1999). An analysis of childhoodmalnutrition in rural India:
Role of gender, income and other household characteristics. Wo rld
Development,27(7), 11511171.
Sitko, N., Chapoto A, Kabwe S, Tembo S, Hichaambwa M, Lubinda R,
Chiwawa, H., Mataa, M., Heck, S., and Nthani, D. (2011). Technical
Compendium: Descriptive Agricultural Statistics and Analysis for
Zambia in Support of the USAID MissionsFeedtheFuture
Strategic Review. Working Paper No. 52 Food Security Research
Smale, M., & Mason, N. (2014). Hybrid seed and the economic well-
being of smallholder maize farmers in Zambia. Journal of
Development Studies,116.
Ssewanyana, N. S. (2003). Food security and child nutrition status among
urban poor households in Uganda: Implications for poverty allevia-
tion. AERC Research Paper 130
Staatz, J. (2000). Strategic pathways and interactions to cutting hunger in
half in Africa. Paper presented at the future opportunities for rural
Africa workshop, November 26, hosted by IFPRI and USAID
Svedberg, P. (1990). Undernutrition in sub-saharan Africa: is there a
gender bias? Journal of Development Studies,26(3), 469486.
Tembo, S., & Sitko, N. (2013). Technical compendium: descriptive agri-
cultural statistics and analysis for Zambia indaba agricultural pol-
icy research institute (IAPRI). Zambia: Lusaka.
UNDP. (2011). Zambia Human Development Report 2011: Service de-
livery for sustainable human development. In G. Luchsinger (Ed.).
UNICEF. (2008). Zambia situation analysis of children and women 2008.
Lusaka: UNICEF Zambia
WHO (World Health Organization). (1995). BPhysical Status: The use
and interpretation of anthropometry. Report of a WHO Expert
Committee.^World Health Organ Tech Rep Ser 854: 1452
World Bank (2008). World development report 2008: agriculture for
Zeng, D., Alwang, J., Norton, G. W., Shiferaw, B., Jaleta, M., and Yirga,
C. (2014). Agricultural technology adoption and child nutrition:
Determinants of child nutritional status in the eastern province
Improved maize varieties in rural Ethiopia. In Selected Paper pre-
pared for presentation at the Agricultural and Applied Economics
Associations 2014 AAEA Annual Meeting. Minneapolis
Julius Manda is an agricultural
economist at the International In-
stitute of Tropical Agriculture
(IITA), based in Malawi and a
PhD candidate at Wageningen
University under the Agricultural
Economics and Rural Policy
group. He received his MSc with
a major in Development Econom-
ics from Wageningen University
in 2012. His research primarily
focuses on adoption and impacts
of new agricultural technologies
on small holder farmerswelfare
in Africa.
Cornelis Gardebroek is an asso-
ciate professor at the Agricultural
Economics and Rural Policy
group at Wageningen University,
The Netherlands. He received his
PhD in Agricultural Economics at
Wageningen University in 2001.
His research and teaching focus
on applied econometric analysis
of agricultural and rural problems.
He has published in the major ag-
ricultural economic journals in-
cluding American Journal of Ag-
ricultural Economics,European
Review of Agricultural
Economics,Journal of Agricultural Economics,Land Economics,Agri-
cultural Systems,andAgricultural Economics. He has published on is-
sues related to risk and uncertainty, adoption of innovations,
microfinance, and agricultural markets. In recent years he received four
awards for his publications and six awards for teaching. He is currently an
editorial board member of the European Review of Agricultural
Makaiko G. Khonje is a research
associate for impact assessment
for the International Institute of
Tropical Agriculture (IITA) based
in Malawi. He provides support to
impact economists responsible for
impact analysis of different pro-
grams and projects for IITA in
Southern Africa and other regions
in Africa. He obtained his Master
of Science in Agricultural and
Applied Economics jointly from
Makerere University in Uganda
and University of Pretoria in the
Republic of South Africa in 2009.
He has worked in areas of agricultural policy, economics, development
economics, natural resources management, climate smart agriculture,
monitoring and evaluation.
Arega D. Alene is an agricultural
economist with the International
Institute of Tropical Agriculture
(IITA) based in Malawi and leads
research programs on impact
evaluation and strategic analysis
of R&D investments and priori-
ties. He started his career in 1992
at Alemaya University of Agricul-
ture in Ethiopia as a lecturer in ag-
ricultural economics upon earning
his BS in agricultural economics
(with distinction) from the same
university. Arega joined IITA in
October 2003 as a postdoctoral fel-
low with the impact, policy, and systems analysis program. He has pub-
lished a number of peer-reviewed journal articles in Agricultural
Economics,Food Policy,World Development,Empirical Economics,Jour-
nal of African Economies,Journal of Developing Areas,Journal of Agri-
cultural Economics,Agricultural Systems,Agricultural Economics Review,
Quarterly Journal of International Agriculture,Outlook on Agriculture,
Journal of Agricultural and Food Economics,and South African Journal
of Agricultural Economics. An Ethiopian national, Dr. Alene holds a PhD
in Agricultural Economics from the University of Pretoria, South Africa.
His research interests include R&D impact evaluation, productivity analy-
sis, agricultural policy, and international development.
Manda J. et al.
Munyaradzi Mutenje is an Ag-
riculture Economist, working
with CIMMYT based in Harare,
Zimbabwe. She received her
PhD from the University of Kwa-
Zulu-Natal, South Africa in 2011
and joined CIMMYT thereafter.
Her professional and research in-
terests focus on food security,
poverty and livelihood analyses,
impact assessments and sustain-
able development. She possesses
vast experience as an extension
officer, monitoring and evaluation
specialist, lecturer and researcher
and is involved in four projects on sustainable intensification in southern
Africa. She has authored and co-authored 10 peer-reviewed publications.
Menale Kassie is a Development
Economist working at the Interna-
tional Maize and Wheat Improve-
ment Center (CIMMYT). Since
he joined CIMMYT in 2010, he
has been coordinating the socio-
economic components of the pro-
gram BSustainable intensification
of maize-legume cropping sys-
tems for food security in eastern
and southern Africa (SIMLESA)^
supported by the Australian gov-
ernment through the Australian
Center for International Agricul-
tural Research (ACIAR). He is al-
so currently a project leader of the Adoption Pathways Project funded by
the Australian government through the Australian International Food Se-
curity Research Center (AIFSRC). This project focuses on establishing
panel data in five African countries in order to understand the drivers of
adoption of technologies and their impacts from a dynamic perspective.
Menales research focuses on adoption and impact of crop and natural
resource management technologies on rural household welfare, using
advanced cross-section-panel econometrics and mathematical program-
ming models. He has analysed the contribution of sustainable land man-
agement technologies on agricultural productivity and the production
risks of technologies on crops such as maize, wheat, groundnuts and
pigeon peas and their effects on poverty and food security.
Determinants of child nutritional status in the eastern province
... Following previous studies, the endogenous switching probit model (ESP) can well addressed above issues (Ayuya et al., 2015;Gregory & Coleman-Jensen, 2013;Manda et al., 2016;Min, Waibel, & Huang, 2017;). The ESP model takes into account unobserved household characteristics that could simultaneously affect households' decisions to use smartphones and to participate in off-farm work or to transform agricultural structures (Lokshin & Glinskaya, 2009). ...
... It is vital for the Z variables in the adoption model (Eq. 2) to contain a selection instrument (Manda et al., 2016). Additionally, for the model to be robust, we need exclusion restrictions. ...
... Additionally, for the model to be robust, we need exclusion restrictions. Following previous studies (e.g., Di Falco et al., 2011;Ma & Abdulai, 2016;Manda et al., 2016;Min et al., 2017;, the inclusion of variables as exclusion restrictions can be validated using a falsification test. According to this test, a variable is 1 A robustness check in the appendix further reports the estimate results which make use of continuous variables of crop diversification. ...
This paper investigates the role of information communication technologies (ICTs) in the transformation of rural economies by evaluating the use of smartphones among farmers in China. We use unique three-wave panel data to document the transformation path of rural economies in recent years. An endogenous switching probit model and a counterfactual analysis are applied to estimate the effects of smartphone use. The results show that from 2008 to 2015, rural economies in China could be characterized by the following three aspects: a) increased off-farm employment, b) expanded grain cultivation, and c) decreased crop diversification. The estimation results indicate that the use of smartphones among farmers had significant impacts on the transformation of rural economies by facilitating the off-farm employment of the farmers' family members, the cultivation of nongrain crops and crop specialization. These findings complement the empirical evidence on the role of ICTs, particularly smartphones, in the development of rural economies in China and other developing countries.
... To our knowledge, only two previous studies have examined the relationship between SI of maize production and child nutrition (Manda et al. (2016a) and Zeng et al. (2017)), and both focus on adoption of improved maize varieties. Yet there are numerous other agricultural practices that can contribute to the SI of maize production and potentially affect child nutrition. ...
... Second, we explore whether these effects operate through the crop productivity and/or income pathways. Third, we use household-level panel data, whereas Manda et al. (2016a) and Zeng et al. (2017) use crosssectional data. This enables us to control for time-constant unobserved heterogeneity, which should improve the internal validity of our estimates. ...
... Because farmers often self-select into agricultural technology adopter groups or some technologies are targeted to certain groups of farmers, selection bias and endogeneity may arise (Manda et al., 2016a;Kassie et al., 2015). In the context of this paper, these problems occur if unobserved factors affecting a household's SI category adoption decision are correlated with children's HAZ and WAZ. ...
Full-text available
Food insecurity, child malnutrition, and land degradation remain persistent problems in sub‐Saharan Africa. Agricultural sustainable intensification (SI) has been proposed as a possible solution to simultaneously address these challenges. Yet there is little empirical evidence on if agricultural management practices and inputs that contribute to SI from an environmental standpoint do indeed improve food security or child nutrition. We use three waves of data from the nationally‐representative Tanzania National Panel Survey to analyze the child nutrition effects of rural households’ adoption of farming practices that can contribute to the SI of maize production. We group households into four categories based on their use of three soil fertility management practices on their maize plots: “Non‐adoption”; “Intensification” (use of inorganic fertilizer only); “Sustainable” (use of organic fertilizer, maize‐legume intercropping, or both); and “SI” (joint use of inorganic fertilizer with organic fertilizer and/or maize‐legume intercropping). The results from multinomial endogenous treatment effects models with the Mundlak‐Chamberlain device suggest that use of practices in the “SI” category is associated with improvements in children's height‐for‐age and weight‐for‐age z‐scores relative to “Non‐adoption”, particularly for children aged 25–59 months. These effects appear to come through improvements in both crop income and productivity. This article is protected by copyright. All rights reserved
... In this study, HAZ, WAZ and WHZ are taken as the outcome variables to imply the children's nutritional status. 6,10,12 While standardizing into Z scores, all three indicators HAZ, WAZ, and WHZ follow the basic formula below: ...
Full-text available
Background: Since mothers are the primary caregivers of children under five, their nutritional status depends on their mothers' capacity to feed and nurture them properly. However, mothers' poor child-feeding practices can also lead to child malnutrition. Mothers' nutritional education and childcare habits can improve children's health through child nutrition training. This study examines how child nutrition training for mother affects children's nutritional status in the impoverished Northern Bangladesh.Methods: In this cross-sectional study, total of 300 mothers have been interviewed and data on demographic, socioeconomic, and child-specific related issues are gathered using simple random sampling from the study areas. The data comprise both treatment and control groups. Propensity score matching (PSM) method is applied to examine impact of child nutrition training for mothers on nutritional status of children in terms of stunting, wasting, and underweight.Results: Empirical results of PSM revealed that the children whose mothers have received trainings have lower prevalence of stunting (0.357 SD), wasting (0.646 SD), and underweight (0.935 SD) as suggested by the average treatment effect on the treated.Conclusions: In summary, this study found positive impact of child nutrition training programs. Therefore, it suggests that the government and NGOs should formulate better and expanded programs focusing on training mothers for the betterment of children’s nutritional status.
... The present study also found that children who were born as LBW babies are at substantially increased risk of remaining undernourished during their early years even after controlling for other children, maternal/household and community characteristics. The pragmatic relationship between birth weight and child undernutrition is in line with previous studies carried out in numerous settings including Zambia and India (Manda et al., 2016;Rahman et al., 2016;Ramakrishnan, 2004). The present study also confirmed the strong association between LBW and child stunting. ...
Sri Lanka has been able to achieve satisfactory progress in health and developmental goals, characterized by substantial declines in infant, child and maternal mortality rates. Despite low mortality levels, better access to healthcare, food security and economic growth, there is little improvement in child nutrition – a paradox and a critical policy challenge that remain unresolved for over the last two decades. Low Birth Weight (LBW), child stunting, wasting and underweight have remained high at constant levels for past 10 years, with increasing health inequalities across different social and ethnic groups. On the other hand, rapid socioeconomic and nutrition transitions can lead to the emergence of a double burden of malnutrition (DBM). This has not been systematically investigated at the national level. Using the data from most recent Sri Lankan Demographic and Health Survey (SDHS), this research investigates the bio-behavioural, socioeconomic and demographic factors underlying inequalities in child nutritional outcomes in Sri Lanka. Further, it examines the risk factors associated with the prevalence of DBM at the household level. The first paper investigates social inequalities underlying LBW outcomes using fixed and random intercept logistic regression models and inequality measures. The results show that LBW is linked to socioeconomic disadvantage, as it is highly concentrated among poor households and in rural and the estate sector; in particular Indian Tamils in the estate sector have the highest risk of LBW of any comparable sub-group of the population. There was substantial unobserved variation in LBW outcomes between mothers. Regression models confirmed that LBW is more closely associated with maternal biological factors, including maternal depletion, than it is with socioeconomic factors. The second paper examines the extent of inequalities in child stunting, wasting and underweight and how these are distributed across different socioeconomic groups, residential sectors and geographical regions. The results show that LBW and BMI are associated with all three outcomes. The effect of child immunisation and feeding practices was not strong for child undernutrition outcomes. Results also suggested that characteristics of the children, their mothers and the households in which they live explain most of the variance in child undernutrition. There is relatively little variation between communities that is not accounted by the composition of those communities. The third paper assesses the driving factors associated with coexistence of child stunting and maternal overweight and obesity at the same household. The results confirm that Sri Lanka is facing a DBM at the household level, with the coexistence of child stunting and maternal overweight. LBW status, maternal age, number of household members, delivery mode, wealth status, ethnicity and province are significantly associated with DBM. Overall, the survey evidence demonstrates that LBW and undernutrition among children are clearly interlinked with socioeconomic disadvantages. The findings of this study suggest that Sri Lanka is facing a dual nutrition challenge of reducing both child undernutrition and maternal overweight and obesity, which are intertwined. The study recommends that child health policies and interventions in Sri Lanka should address both under-nutrition as well as preventing obesity and obesity-related chronic disease risks of malnourished children and their mothers.
... On the other hand, the broadband penetration rate in a village also can reflect the use of internet of other farmers in village, due to the existence of peer effects among farmers in a village, internet use of other farmers in the village may directly affect a farmer's use of internet and thereby further indirectly affect her\his FSB. Following previous studies (Di Falco et al., 2011;Ma & Abdulai, 2016;Manda et al., 2015;Min et al., 2017;Shiferaw et al., 2014), the instrumental variable can be verified by a falsification test. According to this test, the broadband penetration rate can be used as a selection instrument if it affects the use of the internet but does not affect the FSB of rural households that do not use the internet. ...
The rapid development of the internet has changed people's lives and daily behavior. This paper assesses the impact of internet use on food safety behavior among rural residents by examining the case of China, which has been experiencing the rapid spread of the internet. Based on data collected from 1080 rural households in three provinces in Central China, an endogenous switching regression model and counterfactual analysis are used to evaluate the impact of internet use on the food safety behavior of rural residents and to explore the possible channels of impact. The results show that internet use can significantly improve the food safety behavior of rural residents in China, and the potential impact channels include the acquisition of food safety information and the level of food safety knowledge. This paper supplements the literature on the impacts of internet use in rural China by providing empirical evidence on the role of internet use in rural residents' food safety behaviors. This study has important implications for those designing policy to improve residents' food safety behaviors in the information age.
... In southern Africa, maize occupies 50-80% of the arable land area with limited or unstructured incorporation of legumes in the farming systems either as rotations or intercrops (Snapp et al., 2002;Waddington, 2003;Kassie et al., 2012). This strong focus on maize in the diet limits dietary diversity with associated negative side effects of nutritional deficiencies, stunting and wasting especially with children (Manda et al., 2016;Nyakurwa et al., 2017;Chakona and Shackleton, 2018). ...
... In southern Africa, maize occupies 50-80% of the arable land area with limited or unstructured incorporation of legumes in the farming systems either as rotations or intercrops (Snapp et al., 2002;Waddington, 2003;Kassie et al., 2012). This strong focus on maize in the diet limits dietary diversity with associated negative side effects of nutritional deficiencies, stunting and wasting especially with children (Manda et al., 2016;Nyakurwa et al., 2017;Chakona and Shackleton, 2018). ...
Food and nutrition insecurity in southern Africa call for improvements in traditional agriculture systems. Conservation Agriculture (CA) based on minimum soil disturbance, permanent soil cover and crop diversification has been implemented as a strategy to maintain yields while safeguarding the environment. However, less focus has been placed on potential synergistic benefits on nutrition security. Maize-based systems may increase household income through selling but may not lead to proportionate reduction in malnutrition. Crop diversification in CA systems can have a direct impact on the nutritional status of farm households due to improved dietary diversity. Here we assess how the integration of grain legumes, cowpeas and soybeans, in maize-based CA systems either as intercrops or rotational crops affects maize grain yield and stability, total energy yield, protein yield and surplus calories after satisfying the daily requirement per household. The experiments were carried out from 2012 to 2020 (nine consecutive cropping seasons) in six eastern Zambian on-farm communities using 966 observations. Results show that intercropping compromises maize yields with marginal yield penalties of −5% compared to no-till monocropping. However, intercropped yields were more stable across environments. Total system caloric energy and protein yield were highest in intercropping systems due to higher productivity per unit land area owing to the additive contribution of both maize and legumes. Total system caloric energy and protein yield reached yearly averages of 60 GJ ha ⁻¹ and 517 kg ha ⁻¹ , respectively, for the intercropping system as compared to 48 GJ ha ⁻¹ and 263 kg ha ⁻¹ in monocropped maize systems. Tillage-based monocrop resulted in the least stable yields. Our results suggest that intercropping maize with grain legumes in CA systems is a promising option for smallholder farming households to improve dietary diversity, dietary quality and stability of yields thus contributing to sustainable agriculture intensification while maintaining food and nutrition security.
Full-text available
Background Wasting is perhaps one of the signs of malnutrition that has been linked to the deaths of children suffering from malnutrition. As a result, understanding its correlations and drivers is critical. Using quantile regression analysis, this research aims to contribute to the discussion on under-5 malnutrition by analyzing the predictors of wasting in Bangladesh. Methods and materials The dataset was extracted from the 2017–18 Bangladesh demographic and health survey (BDHS) data. The weight-for-height (WHZ) z-score based anthropometric indicator was used in the study as the target variable. The weighted sample constitutes 8,334 children of under-5 years. However, after cleaning the missing values, the analysis is based on 8,321 children. Sequential quantile regression was used for finding the contributing factors. Results The findings of this study depict that the prevalence of wasting in children is about 8 percent and only approximately one percent of children are severely wasted in Bangladesh. Age, mother’s BMI, and parental educational qualification, are all major factors of the WHZ score of a child. The coefficient of the female child increased from 0.1 to 0.2 quantiles before dropping to 0.75 quantile. For a child aged up to three years, the coefficients have a declining tendency up to the 0.5 quantile, then an increasing trend. Children who come from the richest households had 16.3%, 3.6%, and 15.7% higher WHZ scores respectively than children come from the poorest households suggesting that the risk of severe wasting in children under the age of five was lower in children from the wealthiest families than in children from the poorest families. The long-term malnutrition indicator (wasting) will be influenced by the presence of various childhood infections and vaccinations. Furthermore, a family’s economic position is a key determinant in influencing a child’s WHZ score. Conclusions It is concluded that socioeconomic characteristics are correlated with the wasting status of a child. Maternal characteristics also played an important role to reduce the burden of malnutrition. Thus, maternal nutritional awareness might reduce the risk of malnutrition in children. Moreover, the findings disclose that to enrich the nutritional status of children along with achieving Sustainable Development Goal (SDG)-3 by 2030, a collaborative approach should necessarily be taken by the government of Bangladesh, and non-governmental organizations (NGOs) at the community level in Bangladesh.
Full-text available
Wasting is one of the symptoms of malnutrition that has been connected to the deaths of malnourished children. This study was intended to explain the effect of socio-demographic and economic factors on under-5 wasting by evaluating their conditional effect across the distribution of weight-for-height Z (WHZ) scores using the quantile regression (QR) model. The weighted sample which included 13,680 children under 5 years was taken from the countrywide Egyptian DHS 2014 survey. The results depicted that about 2% of Egyptian children were severely wasted, with the prevalence of wasting being around 8%. It was discovered that across the WHZ distribution, the child's features, maternal characteristics, father's education, and social factors had significant but varied contributions in explaining the wasting status of under-5 children. It was revealed that female children had a significant weight advantage, notably 0.21 standard deviation (SD) higher weight at the 95th quantile over their male counterparts. The WHZ score was also found to be significantly positively associated with both age and household's wealth status at the lower and upper tails of the WHZ distribution, respectively. Moreover, in comparison with children whose mothers were underweight, those whose mothers were normal or overweight had higher WHZ scores, with a 1.45 SD increase in WHZ scores at the 95th quantile for mothers who were normal weights. Furthermore, the children who were breastfed, whose mothers received antenatal care (ANC) services, and/or who had educated parents had an advantage in terms of WHZ scores than their counterparts. In addition, the children with higher birth order and/or who resided in urban areas had weight disadvantages compared to their counterparts. Therefore, in order to improve children's nutritional status and achieve the Sustainable Development Goals (SDGs) by 2030, the government and public–private owner organizations must work together at the community level focusing on vulnerable groups.
Full-text available
An adequate food supply is widely recognized as a necessary condition for social development as well as a basic human right. Food deficits are especially common among semi-subsistence farming households in eastern and southern Africa and farm productivity is widely regarded as the locus for enhancing household food outcomes. However, knowledge gaps surrounding benefits associated with climate smart, productivity-enhancing technologies require attention. This study evaluates benefits associated with sustainable intensification farm management practices (crop residue retention, minimum tillage, manure application and use of herbicides, pesticides, fertilizer, and improved seeds) for household calorie and protein supplies and demonstrates their scope across households with high-, moderate- and low- likelihoods of calorie and protein deficits. Household-level calorie and protein deficits were estimated from survey data on food production, acquisition and consumption for households in Ethiopia and Mozambique. Multinomial logistic models were used to identify drivers of household food deficit status and logistic model trees established “rules of thumb” to classify households by food deficit status as low, moderate or high likelihood. In Ethiopia, especially wet seasons were associated with a high likelihood of a food deficit while especially dry seasons were associated with a high likelihood of food deficit in Mozambique. The practices associated with sustainable intensification and related technologies substantially enhanced food outcomes in groups with a high- and a low-likelihood of food deficit, and associated benefits were high for the best-off households. Benefits associated with sustainable intensification technologies were not observed for households with a moderate likelihood of a food deficit and some technologies even increased risk. The sustainable intensification practices assessed here were associated with improved food outcomes yet benefits were limited in scope for households of intermediate status. Thus, there is a need to expand the technical options available to reduce food deficit.
Full-text available
This book shows how to implement a variety of analytic tools that allow health equity - along different dimensions and in different spheres - to be quantified. Questions that the techniques can help provide answers for include the following: Have gaps in health outcomes between the poor and the better-off grown in specific countries or in the developing world as a whole? Are they larger in one country than in another? Are health sector subsidies more equally distributed in some countries than in others? Is health care utilization equitably distributed in the sense that people in equal need receive similar amounts of health care irrespective of their income? Are health care payments more progressive in one health care financing system than in another? What are catastrophic payments? How can they be measured? How far do health care payments impoverish households? This volume has a simple aim: to provide researchers and analysts with a step-by-step practical guide to the measurement of a variety of aspects of health equity. Each chapter includes worked examples and computer code. The authors hope that these guides, and the easy-to-implement computer routines contained in them, will stimulate yet more analysis in the field of health equity, especially in developing countries. They hope this, in turn, will lead to more comprehensive monitoring of trends in health equity, a better understanding of the causes of these inequities, more extensive evaluation of the impacts of development programs on health equity, and more effective policies and programs to reduce inequities in the health sector.
Full-text available
The reduction of child stunting requires an understanding of the major factors that are associated with it most especially before and during infancy of the child. This is because, the velocity of linear growth is highest during first months of life for most infants, and especially in less developed countries like Zambia. Children aged 6–23 months are usually vulnerable to stunting because of various factors such as lack of complementary foods containing the necessary nutrients during the early stages of life which leaves them vulnerable to opportunistic infections resulting in poor health outcomes and outmately stuntedness.
Full-text available
While research establishing the link between high food prices and increased food insecurity in developing countries is robust, similar research about the United States has been lacking. This has been due in part to a lack of suitable price data, but it has also been due to the assumption that prices matter less in the United States, where households spend a relatively small fraction of their income on food. In this article we examine the role that local food prices play in determining food insecurity in the United States by using newly-developed price data. We examine whether low-income households participating in the Supplemental Nutrition Assistance Program (SNAP, formerly Food Stamps) are more likely to be food insecure in areas where food prices are higher. We find that the average effect of food prices on the probability of food insecurity is positive and significant: a one-standard deviation increase in food prices is associated with increases of 2.7, 2.6, and 3.1 percentage points in household, adult, and child food insecurity, respectively. These marginal effects amount to 5.0%, 5.1%, and 12.4% increases in the prevalence of food insecurity for SNAP households, adults, and children, respectively. These results suggest that indexing SNAP benefits to local food prices could improve the ability of the program to reduce food insecurity and economic hardship more generally in areas with high food prices.
Public agricultural research has been conducted in Africa for decades. While many studies have examined its aggregate impacts, few have investigated how it affects the poor. This paper helps fill this gap by applying a new procedure to explore the ex post impacts of improved maize varieties on poverty in rural Ethiopia. Plot-level yield and cost changes due to adoption are first estimated using instrumental variable and marginal treatment effect techniques where possible heterogeneity is carefully accounted for. A backward derivation procedure is then developed to link treatment effect estimates with an economic surplus model to identify the counterfactual household income that would have existed without improved maize varieties. Poverty impacts are finally estimated by exploiting the differences between observed and counterfactual income distributions. Improved maize varieties have led to a 0.8–1.3 percentage drop of poverty headcount ratio and relative reductions of poverty depth and severity. However, poor producers benefit the least from adoption due to the smallness of their land holdings.
In this article, we describe the switch probit command, which implements the maximum likelihood method to fit the model of the binary choice with binary endogenous regressors.
This paper reviews four treatment parameters that have become commonly used in the program evaluation literature: the average treatment effect, the effect of treatment on the treated, the local average treatment effect, and the marginal treatment effect. We derive simply computed closed-form expressions for these treatment parameters in a latent variable framework with Gaussian error terms. These parameters can be estimated using nothing more than output from a standard two-step procedure. We also briefly describe recent work that seeks to go beyond mean effects and estimate the distributions associated with various outcome gains. The techniques presented in the paper are applied to estimate the return to some form of college education for various populations using data from the National Longitudinal Survey of Youth.
This study evaluates the effect of certified organic production on poverty in smallholder production systems. Data was collected from cross sectional survey of local market-oriented peri-urban vegetable and rural honey producers in Kenya. Poverty was measured using the multidimensional poverty methodology and endogenous switching probit model used to assess the effect of certified organic production on multidimensional poverty. Findings were that certified producers were less likely to be multidimensional poor compared to their counterfactual case of not participating in organic certification schemes. Additionally, noncertified producers would be less likely to be poor if they were to participate in organic certification production.
This paper analyzes the adoption and welfare impacts of improved maize varieties in eastern Zambia using data obtained from a sample of over 800 farm households. Using both propensity score matching and endogenous switching regression models, the paper shows that adoption of improved maize leads to significant gains in crop incomes, consumption expenditure, and food security. Results further show that improved maize varieties have significant poverty-reducing impacts in eastern Zambia. The paper concludes with implications for policies to promote adoption and impacts of modern varieties in Zambia