Child malnutrition and poverty: the case of Pakistan.
Arif, G.M. ; Farooq, Shujaat ; Nazir, Saman 等
1. INTRODUCTION
The role of economic factors, particularly income and consumption,
in the wellbeing of a population is well documented. The well-being,
however, does not depend solely on these factors, social indicators such
as life expectancy, health, education and nutrition serve an important
complementary function [Linnemayr, et al. (2008)]. The most significant
social problems in many developing countries including Pakistan are
widespread child malnutrition, high infant mortality and low literacy.
Child malnutrition is considered as the key risk factor for illness and
death, contributing to more than half the deaths of children globally
[Cheah, et al. (2010)]. It also affects the child morbidity rate and
poses threat to their physical and mental development, which results in
lower level of educational attainment [Chirwa and Ngalawa (2008)]. The
recent literature therefore considers the nutrition status as an
important dimension of individual wellbeing [Babatunde, Olagunju, and
Fakayode (2011)].
Although the causes of child malnutrition are interrelated and
multi-sectoral involving many different aspects of life [Cheah, et al.
(2010)], food insecurity, poor nutritional status of mothers, frequent
infections, lower utilisation of health services and care provided to
children are considered the most important correlates of malnourishment
[Linnemayr, et al. (2008)]. There is, however, no consensus in the
literature regarding the role of poverty in child malnutrition. Results
are rather mixed. Several studies have shown malnutrition as a
reflection of poverty, with people not having enough income to buy food,
while many other empirical studies have found no association between
poverty and child malnutrition [Chirwa and Ngalawa (2008)].
The performance of Pakistan in social indicators including the
nutritional status of children is not satisfactory. Although the
proportion of underweight children has declined during the last one and
a half decade, approximately one-third of young children are still
counted as underweight, according to the 2011 National Nutrition Survey
(NNS). Stunting and wasting, the other two measures of children's
nutritional status have deteriorated. Thus, child malnutrition in
Pakistan can be considered as a widespread phenomenon.
The question is how this phenomenon can be explained? Is the
malnutrition of children related to poverty status of their households
or are other factors particularly child illness, health status of their
mothers and access to health care services the major determinants? An
investigation of this question is vital in view of both poor health
indicators (particularly high infant and child mortality) and
instability in poverty reduction in the past. The findings of earlier
studies are not conclusive. Alderman and Garcia (1993) found that
illness and diarrhea are strongly related to the poor nutrition among
young children in Pakistan. Arif (2004) found a significant relationship
between poverty and weight-for-age, but no association of poverty with
stunting or wasting. He, however, did not take into account the
endogeneity of the welfare index (poverty) in the nutritional status
equation. As Chirwa and Ngalawa (2008) argue:
The poor nutritional status of children in a household may reflect
the lack of adequate calorie intake that may in turn affect the health
status of adults. The poor health of adults may negatively affect their
income earning potential and demand for calories that may adversely
affect the nutritional status of children and members of the household.
The major objectives of this paper are: first, to examine the
trends in child malnutrition during the last decade using three-round
data of a longitudinal household survey; and, second, to find its
correlates, focusing on household poverty. It has used individual
(child), household and community level variables to understand
variations in child malnutrition. Poverty status of households is the
key factor used in this study to understand the malnutrition phenomenon.
The rest of the study is organised as follows. The conceptual
framework, data sources and methodology used in the study are discussed
in Section 2, followed by a presentation of the trends in child
malnutrition and poverty in Section 3. The socio-demographic
determinants of child malnutrition are presented in Section 4, which
include gender and age of children, mother and household
characteristics. The determinants of child nutrition are examined in a
multivariate analysis in Section 5 while the penultimate section
presents the discussion on poverty and child malnutrition nexus in
Pakistan, followed by conclusions in the final section.
2. CONCEPTUAL FRAMEWORK, DATA SOURCES AND METHODOLOGY
The nutritional status of children is determined by factors that
can be divided into three main categories; immediate, underlying and
basic causes [UNCIEF (1990)]. Immediate causes are linked with the
dietary intake and occurrence of diseases in children while the
underlying causes encompass the access to food for children and mothers,
their health care and the environmental conditions. Basic causes include
economic, political and institutional structure of the country and
availability of resources. Poverty can affect child nutrition through
dietary intake or inability of a household to buy sufficient food. Food
inadequacy increases the chances of infections and frequent infections
cause nutritional deficiencies. Although many studies have explored the
poverty and child malnutrition nexus, its robustness is not established
[Pal (1999)]. As Sununtar (2005) shows:
Malnutrition is the result of marginal dietary intake compounded by
infection. In turn, marginal dietary intake is caused by household food
insecurity, lack of clean water, lack of knowledge about good
sanitation, and lack of alternative sources of income. It is also
compounded by inadequate care, gender inequality, poor health services,
and poor environment. While income is not the sum total of people's
lives, health status as reflected by level of malnutrition is.
The conceptual framework, which this study has used to examine the
determinants of children's nutritional status, is based on the
household utility maximising model by specifying a household production
function [Becker (1965); Behrman and Deolalikar (1988); Strauss and
Thomas (1995)]. In this model, it is assumed that a household has
preferences that can be characterised by the utility function, U which
depends on consumption of a vector of commodities, A, leisure, L, and
quality of children represented by their nutritional status, N:
U = u(X, L, N) ... (1)
Household utility is maximised subject to several constraints,
including a time specific nutrition production function and income
constraints. The nutritional status of children is determined by food
availability, morbidity, access to health services and the quality of
care at home. The nutritional outcome of each child measured by standard
anthropometric measures can be derived as:
Ni = n(C, W, H, Z, e) ... (2)
Where C is consumption, W is a vector of child-specific
characteristics, H is a vector of household specific characteristics, Z
is a vector of health variables and e is child-specific disturbance
term. In equation 2, N is measured by standardised anthropometric
measures of height-for-age, z-score (HAZ), weight-for-age, z-score (WAZ)
and weight-for-height, z-scores (WHZ). The z-scores are computed by
using the World Health Organisation recommended reference population
[WHO (2006)]. The WAZ of a child, for example, is the difference between
the weight of the child and the median weight of the reference
population of the same age and sex, divided by the standard deviation
(SD) of the weight of same group of children:
WAZ = Wi - Wr/SD ... (3)
Three anthropometric measures, WAZ, HAZ and WHZ, provide different
information about the nutritional status of children. HAZ measures
stunting, a condition that reflects chronic malnutrition. WHZ measures
the current nutritional status of a child while WAZ captures aspects
covered in both HAZ and WHZ [Chirwa and Ngalawa (2008)].
Pakistan Institute of Development Economics (PIDE) has carried out
three rounds of a longitudinal (panel) survey in 2001, 2004 and 2010.
The first (2001) and third (2010) rounds of the survey collected data on
age, weight and height of children, necessary for anthropometric
measurement. This study has used these two rounds of data to see changes
in child nutritional status during the last decade; whereas, to study
the determinants of child nutrition, all the three rounds data (2001,
2004 and 2010) have been used. The sample of the first two rounds of the
panel survey (2001, 2004) consisted of only rural areas of 16 districts
located in four provinces of the country, and it was named as the
Pakistan Rural Household Survey (PRHS). The third round sample survey,
carried out in 2010, was named as the Pakistan Panel Household Survey
(PPHS) since it includes both rural and urban areas of these 16
districts [for more detail, see Arif and Farooq (2012)]. The total rural
sample of the 2010 PPHS consisted of 2800 households while the urban
sample comprised of 1342 households, leading to the total sample of 4142
households. In the PPHS-2010, data on weight and height of all children
less than 6 years old were obtained. However, this study has included in
the analysis 6-59 months old children. The study has identified in this
age group 3218 children, about half of them (48.2 percent) female (Table
1). The data on weight is available for 80 percent of the children while
the data on height is available for approximately two-third of the
sampled children.
Following the WHO recommendations, for WAZ analysis, children with
-6 to 5 z-scores are included. For HAZ and WHZ, the children with -6 to
6 and -5 to 5 scores are included [WHO (2008); WFP and CDC (2005)].
Outliers or children out of the given ranges were found more in HAZ
z-scores than in WAZ and WHZ scores. A child is characterised as
malnourished if s/he is more than two standard deviations below the
standard reference population. While these anthropometric measures are
important indicators of child malnutrition, child health itself could be
considered an extreme form of child malnutrition. Selective child
mortality could then lead to biased estimates if children who have died
by 2010 and are missing from the sample. These children were more likely
to be from households that are extremely poor. This selective attrition
has been checked with no evidence of higher child morality in the
poorest households.
Equation 2 has been used to examine the determinants of child
nutritional status in 2010. Individual characteristics of children,
household level characteristics and community variables are included
into this equation. Individual child characteristics include age and
gender of the child. Parental characteristics include the level of
educational attainment of mother. Two housing related variables included
in the equation are the structure of dwelling units (pacca/katcha) and
availability of toilet--a village level variable. Per capita consumption
expenditure is used in the equation for poverty status of the household.
Availability of lady health workers at the village level represents the
health care services while the region of residence (urban/rural) is a
community variable.
Per capita expenditure, a household level variable, is likely to be
determined, as reported earlier, by the anthropometric outcomes through
its effect on the health status of adults and their earnings [Chirwa and
Ngalawa (2008)]. In order to account for the endogeneity problem, the
following methodology has been adopted:
(i) The analysis in the first stage is limited to rural panel
households covered in 2001 and 2010. To get robust estimates, per capita
expenditure in 2001 is used in equation 2 to explain variation in the
2010 child nutrition status. As the sample is limited to children below
5 years old in 2010, who were not born in 2001 therefore their nutrition
outcomes are less likely to affect 2001 poverty status (or consumption
expenditure). Both the OLS and two-stage least square (2SLS) techniques
are used in the analysis: in OLS, the actual per capita expenditure in
2001 has been used to explain the child nutritional status in 2010,
while in 2SLS, per capita expenditure in 2001 is instrumented by 2001
household variables including landholding, ownership of livestock, work
status of the head of households and household size.
(ii) In the second stage, per capita expenditure is replaced by
change in poverty status between 2004 and 2010. The change in poverty
status has four categories: poor in two rounds (2004 and 2010); non-poor
in two rounds; moved out of poverty; and moved into poverty. The last
two categories are combined to represent transitory poverty. The
analysis is carried out only for the 2004 and 2010 rural panel
households. The official poverty line has been used for poverty
estimation [for details, see Arif and Shujaat (2012)].
(iii) In the final stage, the analysis has used the 2010 PPHS full
sample (rural and urban), and per capita expenditure is replaced by the
perceived household food security. The OLS technique has been applied in
this stage, where perceived food security indicators are used as
independent variables instead of per capita expenditure.
3. TRENDS IN CHILD NUTRITION AND POVERTY
Pakistan has a long history of data collection on socio-economic
and demographic issues through household surveys, but information on
child nutrition is generally missing in these surveys. It is, thus,
difficult to analyse the trends in nutritional status of children for a
long period of time. However, the NNS carried out in 1985-87, 2001 and
2011 has to some extent filled the gap. Some other surveys, though
relatively smaller in their sample sizes, such as Pakistan
Socio-economic Survey (PSES) 2001, Pakistan Demographic and Health
Survey (PDHS) 1990, PRHS 2001 and PPHS-2010, have also gathered data on
height and weight of children to determine their nutritional status.
Table 2 has pulled together information from these sources on three well
known anthropometric measures; underweight, stunting and wasting for
rural and urban areas. According to the NNS series, the incidence of
underweight among children aged 6-59 months old has gradually declined
from around 48 percent in 1985-87 to about 32 percent in 2011. This
decline has been observed in both rural and urban areas. The two rounds
of the panel dataset, PRHS-2001 and PPHS-2010 also support the NNS data
and show a decline in underweight children during the last decade,
although the NNS and the panel data show different magnitudes of
underweight children. However, despite this decline in the proportion of
underweight children overtime, at present more than one-third of
children (32 percent in NNS-2011 and 39 percent in PPHS-2010) are
underweight.
The situation of other two anthropometric measures, stunting and
wasting, is different and alarming. The stunting, which reflects chronic
malnutrition, has increased between 2001 and 2011. According to the
NNS-2011 data, around 44 percent of children were stunted. This
proportion is about 2 percentage points higher than the stunting in
1985-87 (Table 2). The panel data, however, show no major change in
stunting between 2001 and 2010. Overall, the magnitude of stunting is
much higher in the panel datasets (PRHS-2001 and PPHS-2010) than in the
NNS dataset. According to the NNS series, the incidence of wasting has
also increased from 11 percent in 1985-87 to 15 percent in 2011. The
panel series, however, shows a mild decline in wasting, from 18 percent
in 2011 to 17 percent in 2010. The deterioration in stunting overtime,
with the high prevalence of underweight (more than one-third), reflects
the weak performance of Pakistan in improving the nutritional status of
children.
The data in Table 2 are also presented separately for rural and
urban areas. All data sources indicate higher prevalence of underweight
and stunted children in rural areas than in urban areas. However, in
contrast, wasting appears to be moderately higher in urban areas than in
rural areas. Majority of malnourished children in urban as well as rural
areas are in the 'severe' category (Table 3). The proportion
of children in this category is very high in case of stunting. Thus not
only is the overall prevalence of stunting high, but also children are
severely malnourished.
The available data on the poverty levels and trends in Pakistan for
the last five decades show that poverty reduction has not been
sustainable; rather that it has fluctuated remarkably. In late 1980s,
when approximately half of the children were malnourished (underweight),
poverty was at a very low level, only 17 percent. There is a consensus
in the poverty literature about a sharp rise in poverty in the 1990s.
The incidence of poverty, as estimated from the three rounds of panel
survey (2001, 2004 and 2010), also illustrates that poverty has
fluctuated during 2001-2010 (Table 4). First poverty declined from 31.3
percent in 2001 to 24.1 percent in 2004 and then increased to 27 percent
in 2010 in two major provinces of Pakistan, Punjab and Sindh. In rural
Pakistan, poverty declined by 5 percentage points, from 27.5 percent in
2001 to 22.4 percent in 2010. In 2010, the overall poverty was estimated
at 20.7 percent with a higher incidence of poverty in rural areas (22.4
percent) than in urban areas (16.6 percent).
Poverty estimates based on the three rounds of data show that
during the last decade, more than half of the rural population (51
percent) in two largest provinces, Punjab and Sindh, remained in the
state of poverty at least for one point in time. Within this poor group,
the majority was categorised as 1-wave poor (31 percent), although
considerable proportion, around 17 percent, is found to be poor in
2-periods. Chronic poor, those who remained poor in all three waves is
only 4 percent, which is less than half of the population who remained
poor in two waves. The three-wave data are spread over 10 years period,
2001 to 2010. During this decade, only a small proportion of households
remained continuously in the state of poverty. Movement into and out of
poverty is a common phenomenon in Pakistan, particularly in its rural
areas.
4. SOCIO-DEMOGRAPHIC DIFFERENTIALS OF CHILD MALNUTRITION
Figures 1-3 present data on three anthropometric measures by gender
for the total sample as well as rural-urban areas, while Figure 4
presents data on the nutritional status of children by their age.
Overall there is no major gender difference in the three measures.
However, gender differences are more profound within the rural and urban
areas. In rural areas, for example, more males are underweight and
wasted than females while in urban areas the prevalence of malnutrition
(under weight and wasting) is higher among females than among males. It
is not easy to explain these gender differentials in rural and urban
areas. However, it appears from the available studies in Pakistan and
elsewhere in subcontinent that evidence on gender differentials in child
nutritional status is inconclusive. As Shah, et al. (2003) while
studying child nutrition in 64 villages of Sindh (Pakistan) found no
difference in stunting between male and female children. In rural
Bangladesh, Choudhury, et al. (2000) found that female children were
more likely to be severely malnourished than male children. However, the
nutritional survey in 2005-06 in India reveals no significant difference
in nutrition status (stunting and wasting) between boys and girls (1). A
recent study in urban slum of India found the prevalence of malnutrition
higher among female children than among male [Damor, et al. (2013)].
Figure 4 shows a nonlinear relationship between the age of child
and the three measures of his nutritional status. In the case of
underweight, it is highest for the 6-11 months old children. It
decreases for the next age group (12-21 months), but it increases for
the 2-3 years old children. The lowest prevalence is found for children
in age group 48-59 months. Despite these variations across the age
groups, the minimum prevalence of underweight stands at 36 percent,
suggesting widespread malnutrition in all age groups of the sampled
children. The situation is not different for stunting and wasting
(Figure 4).
[FIGURE 1 OMITTED]
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
[FIGURE 4 OMITTED]
5. DETERMINANTS OF CHILD NUTRITION
The determinants of child nutritional status are examined by
estimating the Equation 2, where WAZ z-scores, WHZ z-scores and HAZ
z-scores are used as the dependent variables. Independent variables
include child characteristics (gender and age), child illness (incidence
of diarrhea), education of mother, per capita expenditure as an
indicator of household poverty, number of siblings, environmental
factors (structure of the dwelling unit), access to toilet (a village
level variable), availability of LHWs at village level and the region
(rural-urban) of residence. As noted in Section 2, because of the
endogeneity problem, per capita expenditure of 2001 are instrumented by
2001 household ownership of land and livestock, work status of the head
of household and household size. The 2SLS regression has been used. In
addition, 2004 poverty status and change in poverty status between 2004
and 2010 has also been used to predict the nutritional status of child.
Table 6 provides the summary statistics of the 2010 dependent and
independent variables.
The mean values for the z-scores of WAZ, HAZ and WHZ are -1.55,
-2.38 and 0.12 respectively. Per capita expenditure is computed at
Rs.1167 per month. About half of the sampled children are female and
their mean age is about 31 months (Table 6). About 9 percent of the
children had diarrhea during the month preceding the survey. More than
half of the housing units where children live are pacca (cemented) and
more than 50 percent of the households have been visited by LHWs and
have a toilet with flush.
Results based on 2001-2010 panel households for the three equations
(WAZ, HAZ and WHZ) are presented in Table 7 where OLS and 2SLS methods
have been applied. The First Stage regression results of 2SLS have been
presented in Appendix Table 1 which suggests that all the six excluded
instruments are highly correlated with per capita expenditure. The
question is whether the instruments for per capita expenditure are
uncorrelated with the disturbance process. To answer that, we computed
the test for over identifying restrictions and results are presented in
Appendix Table 2 where both the Sargan and Basmann test shows that
specification of the equation is satisfied.
Overall, the results of both techniques (OLS and 2SLS) are similar
except for per capita expenditure, which shows a significant association
with child malnutrition in the OLS model while it turns out to be
insignificant in the 2SLS model (Table 7). It supports the existing
literature that impact of poverty on the nutritional status of children
is ambiguous.
The gender variable has significant and negative relation with WAZ
and WHZ, showing that boys are more likely than girls to be underweight
and wasted. Age has a positive impact on WAZ while [Age.sup.2] has also
a significant and positive association with the WAZ scores, suggesting a
non-linear relationship, which implies that boys gradually improve their
weight/age score. The coefficient of [Age.sup.2] is not significant in
2SLS.
The number of siblings does not have a significant effect on all
the three anthropometric measures of nutritional status. The incidence
of diarrhoea had a statistically significant negative association with
the three anthropometric measures. It appears that morbidity adversely
affects the growth of children by reducing the ability of a body to
convert food into energy. Surprisingly mothers' education effect
turned out to be statistically significant only on WAZ, and not on HAZ
and WHZ.
An environmental factor represented by the availability of flush
toilet at village level has a statistically significant relationship
with WAZ and WHZ scores, but the relationship is insignificant for the
HAZ scores. It appears from this association that the village level
environmental factors such as toilet with a flush system affect the
current health status more than impacting the chronic malnutrition
(HAZ).
The role of LHWs in improving the nutritional status of children is
positive with statistically significant association with WAZ, HAZ and
WHZ scores. It means that the availability of health services at village
level help to improve not only the current nutritional status but also
affect child growth in the long term through improving the HAZ and WHZ.
To explore further the relationship between poverty and the
nutritional status of children (weight for age), per capita expenditure,
which represents the 2001 poverty status, has been replaced by the
poverty status in 2004 and change in poverty status between 2004 and
2010 in two models, as given in Table 8. The hypothesis is that the
poverty of a household in recent past and movement in poverty status
affect the nutritional status of children. As noted earlier, the sampled
children included in the nutritional status equation were 6-59 months
old. The PPHS was carried out in the last quarter of the year 2010, as
part of the panel survey. Its earlier round was carried out in 2004, but
only in rural areas of Punjab and Sindh, the two largest provinces of
the country. Poverty in 2004 or a change in the poverty status of
households between 2004 and 2010 (2), when the sampled children were
born, may have an impact on their nutritional status. Table 8 shows the
results of OLS for WAZ, where two models have been estimated. In
model-1, per capita household expenditures are replaced by the household
poverty status in 2004; poor in 2004 are given the value 1, zero
otherwise. In model-2, two dummies of poverty dynamics are used;
transitory poor and chronic poor while the third category, remained
non-poor in 2004 and 2010, is used as the reference category.
The model-1 examines the effect of poverty status in 2004 on the
child nutritional status in 2010 while model-2 deals with the effects of
poverty movements on the child nutritional status. No single category of
poverty or poverty dynamics turned out to be statistically significant
(Table 8). It shows that the recent past poverty status, as well as
household's movement into or out of poverty even the chronic
poverty is not relevant to the nutritional status of children in
Pakistan. It is noteworthy that age, age-square and education of mother
that were statistically significant in the WAZ models shown in Table 7
did not turn out to be significant in the models shown in Table 8. There
is no change in the significance of other variables.
In the PPHS-2010, the sampled households in both rural and urban
areas were asked if they faced food shortage during the last 12 months.
In another similar question, they were asked whether the food during
last 12 months has been insufficient for the household members. These
two questions show the perception of households about the food security.
This type of household perception may not reflect a true picture of the
household food security because it does not determine for how many days
they have faced food shortage and what is the nature of the food
shortage. However, it does provide information about the households that
have faced food shortage for some time during the 12 months preceding
the survey. The PPHS-2010 shows that about one-third of the households
reported such shortage.
In the final stage of analysis, the Equation 2 is estimated by
replacing 2001 per capita expenditure with the household's
perceived food security variables, as discussed above. If a household
faced food shortage or food was insufficient during the last 12 months,
it was coded 1, otherwise zero. Two models (for WAZ only) have been
estimated. In model-3, the variable food shortage is used while in
model-4, it is replaced by the perceived food insufficiency. Table 9
presents the findings of the OLS regression. The variables representing
food security or food shortage also did not turn out to be statistically
significant. Like poverty, the perceived food shortage is not related to
the nutritional status of children. The regional dummy (rural-urban) was
entered into the models to examine the effects of community factor on
the nutritional status and it appears from negative sign of this
variable that the nutritional status of urban children is lower than
their rural counterparts. Since the difference in child malnutrition is
significant between the rural and urban areas, the determinants of
malnutrition are also estimated separately for these two sub-samples and
are reported in Table 10. Age of the child, which has significant
positive association with the malnutrition in full sample models, lost
its significance in rural/urban separate models. Mother's education
that was insignificant in full model, turned out to be significant in
the rural model, showing the importance of women education for child
welfare in rural settings. No major difference could be found in the
magnitude and significance of other variables used into these two
separate models.
6. DISCUSSION: EXPLANATION OF POVERTY--CHILD MALNUTRITION NEXUS IN
PAKISTAN
A major finding of this study is that the nutritional status of
children in Pakistan is predominantly related to their exposure to
illness (diarrhoea), provision of health care services and environmental
factors. The recent past poverty status of a household or change in
poverty status over time as well as the perceived food shortage are not
significantly associated with child malnutrition. Now the question is
how to explain this lack of association between the poverty and child
nutritional status. As noted earlier, there is no consensus in the
literature regarding the role of poverty in child malnutrition. Several
studies have shown malnutrition as the reflection of poverty, while
other empirical studies have found no association between poverty and
child malnutrition [Chirwa and Ngalawa (2008)]. As NEPAD (2004) notes,
"[the] availability and access to sufficient quantity and quality
of affordable food is necessary but not sufficient to ensure adequate
nutrition". Alone the food security and low poverty cannot make a
household nutritionally secured. Beside poverty, other basic
determinants of nutrition are social, economic, political, cultural and
non-food factors i.e. care and health [ACC/SCN-IFPRI (2000)]. A
nutrition secure society is a society that achieves the adequacy of
food, adequate maternal and child care, and good health and
environmental services [Gillespie and Haddad (2003)].
In the case of Pakistan, based on the PSES-2001, Arif (2004) has
earlier found a positive impact of per capita expenditure (or poverty)
only on weight-for-age, but no association with stunting or wasting.
But, he did not account for the endogeneity problem. When endogeneity
problem is addressed in the present study, poverty has shown no
association with all three anthropometric measures (underweight,
stunting and wasting). As shown earlier, Pakistan has not experienced a
sustained reduction in poverty during the last five decades, it has
fluctuated. In the 1990s, poverty increased, but the prevalence rate of
underweight declined. Poverty during the first half of the last decade
declined, but it increased in its second half. Although the proportion
of underweight children declined during the last decade, stunting and
wasting remained unchanged or even increased.
Poverty in Pakistan is largely considered a rural phenomenon, but
there is no major difference between urban and rural areas in child
malnutrition (see Table 2). This can be partially explained by the rural
economy dynamics. Despite highly unequal land distribution, about
two-thirds of the rural households are engaged in production of some
food items from agriculture or/and livestock related activities,
ensuring necessary dietary intake of household members. Moreover, social
and financial support is deeply embedded in Pakistani culture, where the
vulnerable households get support from their neighbours, relatives and
well-off families and thus maintain their subsistence nutritional
intake. Such support is even enhanced when some households or group of
society face some natural or unnatural negative shocks. The Government
of Pakistan also provides a number of direct and indirect transfers and
subsides to the poor to protect them from both the short and long-term
social and financial insecurity. A number of targeted direct transfers
in the public sector such as zakat, Baitulmal and Benazir Income Support
Programme (BISP) help in the provision of food. Nayab and Farooq (2012)
have found a positive impact of the BISP on food consumption.
Evidence from other countries like India shows that the issue is
not about having enough food; there is a need to look beyond income
levels, poverty and food availability [Mendelson (2011)]. The episodes
of illness, particularly diarrhoea, reduce the ability of body to
convert food into energy, leading to high levels of malnutrition among
children. Children who suffer from illnesses, even though their dietary
requirements are met, cannot grow robustly as excessive nutrition losses
occur during the frequent episodes of disease [Rosenberg, Soloman, and
Schneider (1977)]. The frequent episodes of diarrhoea account for high
neonatal and infant mortality, which is the second most killing disease
among children in world [UNCIEF (2011)]. Pneumonia is also one of the
leading killers
of Pakistani children [UNICEF (2012)]. (3) There is a strong
association between the incidence of diarrhoea and lack of access to
safe drinking water. The access to clean water is another major concern
in both urban and rural areas of the country. For example, in Karachi,
the largest city of the country, the 22 percent water samples as
provided by the government were found to be either non-chlorinated or
containing insufficient amount of chlorine. (4) While the reduction in
poverty is vastly dependent on private household consumption
expenditures, the improvements in child malnutrition are largely driven
by public expenditures. Improved sanitation and access to clean water,
usually invested by the government, can have significant impact on
malnutrition [IFPRI (2005)].
Similarly, the significance of LHWs in the present analysis shows
the importance of child care services in improving the nutritional
status of children. In Pakistan, the health facilities are very poor as
the country has been spending only 0.6 percent of its GDP on health
services over the last two decades. The pervasive and troubling
weaknesses in the health system have caused high mortality and diseases
among women and children. (5)
7. CONCLUSIONS
The high prevalence of malnourishment among children in Pakistan
remains a critical issue in policy debate. This study has examined the
trends in child malnutrition and assessed its linkages with the
characteristics of children, provision of health care services and the
poverty status of households. The study found very high levels of
malnutrition among children and no significant association between
poverty and child malnourishment. No association could be found between
the perceived food shortage and child malnutrition. Child malnutrition
is deeply rooted in child illness, environmental factors and weak health
system.
Several policy suggestions emerge from the findings of this study.
First, Pakistan should not assume that economic growth or poverty
reduction will automatically translate into improved child nutrition.
Measures for enhancing actions about social determinants of health, and
specific programs for improved early life nutrition are needed to reduce
child malnourishment.
Second, the existing child and maternal health care services in the
country are inadequate for improving child nutritional status. Many
developing countries, some with even more limited resources than
Pakistan, are 'on the track' to improve maternal and child
health. The key weaknesses in Pakistan, which hold back the
country's progress in this regard, are insufficient financing, poor
governance, lack of skilled health workers, and inequalities in access
to healthcare. (6) Thus, direct investments in appropriate health
interventions, focusing on women and children, are necessary to improve
child health and nutrition.
Third, the high incidence of child illness, particularly diarrhoea,
needs to be overcome by preventive measures, including the awareness
about hygienic environment and specific dietary intake during illnesses
that compensate nutrient losses. Finally, the positive contribution of
LHWs to child nutrition shows the importance of the provision of door to
door health care services in Pakistan. The LHW program should be
universalised, particularly in rural areas.
Appendix Table 1
The Determinants of Child Malnutrition-First Stage Results
of 2SLS Estimates
WAZ HAZ WHZ
Determinants Coeff. Coeff. Coeff.
Per Capita Expenditure (sq) 0.001 * 0.001 *** 0.001 ***
Sex (male = 1) -3.850 -14.513 -9.549
Child age (months) -13.175 *** -17.540 *** -16.327 ***
Child [age.sup.2] 0.235 *** 0.276 *** 0.263 ***
Number of Siblings
(<2 as reference)
2-3 -65.278 -80.815 -14.428
4-6 -275.538 * -276.345 * -249.020 *
7+ -266.647 * -208.969 *** -274.388 **
Diarrhoea (yes = 1) -109.193 -165.675 *** -125.630
Mother's Education (no
education as reference)
Primary -19.636 -23.394 31.325
Secondary 267.604 * 305.398 ** 221.665
College 410.926 * 434.357 ** 340.260
Housing Type (Pacca = 1) 79.325 102.945 107.340
Toilet Facility (% at
village level) 3.676 * 4.117 * 3.635
LHW visited (% at village
level) 1.202 -0.348 0.747 *
Education of Head of
Household in 2001 (up
to primary as ref.)
6-10 421.998 * 382.563 * 572.025 *
11 and above 356.412 * 234.755 275.157 ***
Work status of head O1
(yes = 1) 111.699 *** 154.156 *** 169.502 **
Household size_01 (numbers -46.771 * -47.740 * -50.111 *
Land_01 (in acres) 2.000 *** 3.520 *** 2.962
Large animals_01 (in numbers) 48.971 * 41.336 * 39.993 *
Constant 1410.542 * 1523.364 * 1560.652 *
F-stat 13.26 9.16 9.70
R-square 0.1705 0.1596 0.1659
Adjusted R-square 0.1577 0.1422 0.1488
N 1,311 986 977
Source: Authors' estimation from the micro-data of PRHS
2001 and PPHS 2010.
Note: * significant at 1 percent, ** significant at 5
percent, *** significant at 10 percent.
Appendix Table 2
Over Identification Test
WAZ HAZ WHZ
Sargan (score) chi2(5) 4.79804 1.45916 2.28048
(p = 0.4410) (p = 0.9177) (p = 0.8091)
Basmann chi2(5) 4.73853 1.43019 2.23757
(p = 0.4486) (p = 0.9210) (p = 0.8154)
Source: Authors' estimation from the micro-data of
PRHS 2001 and PPHS 2010.
REFERENCES
Aber, J. Lawrence and Bennett Neil G. (1997) The Effects of Poverty
on Child Health and Development. Annu. Rev. Public Health.
ACC/SCN-IFPRI (United Nations Administrative Committee on
Coordination, Standing Committee on Nutrition and International Food
Policy Research Institute) (2000) Fourth Report on the World Nutrition
Situation. Geneva: ACC/SCN in collaboration with IFPRI.
Alderman Harold and Garcia Marito (1993) Poverty Household Food
Security and Nutrition in Rural Pakistan. International Policy Research
Institute, Washington, DC. (Research Report 96).
Arif, G. M. (2004) Child Health and Poverty in Pakistan. The
Pakistan Development Review 43:3, 211-238.
Arif, G. M. and Shujaat Farooq (2012) Dynamics of Rural Poverty in
Pakistan: Evidence from Three Waves of the Panel Survey. Pakistan
Institute of Development Economics, Islamabad. (P1DE Working Paper).
Babatunde, R. O., F. I. Olagunju, S. B. Fakayode, and F. E.
Sola-Ojo (2011) Prevalence and Determinants of Malnutrition among
Under-five Children of Farming Households in Kwara State, N igeria.
Journal of Agricultural Science 3:3, 173-181.
Becker, G. S. (1965) A Model of the Allocation of Time. The
Economic Journal 75:299, 493-517.
Behrman, J. B. and A. B. Deolalikar (1988) Health and Nutrition. In
H. B. Chenery and T. N. Srinivasan (eds.) Handbook of Development
Economics 1, 631-711.
Cheah, W. L., W. W. Muda, and Z-H Zamh (2010) A Structural Equation
Model of the Determinants of Malnutrition among Children in Rural
Kelantan, Malaysia. The International Electronic Journal of Rural and
Remote Health 10, 1248.
Chirwa, E. W. and N. Ngalawa (2008) Determinants of Child Nutrition
in Malawi. South African Journal of Economics 76:4, 628-640.
Gillespie, S., and L. J. Haddad (2003) The Double Burden of
Malnutrition in Asia: Causes, Consequences and Solutions. New Delhi,
India: SAGE Publications.
IFPRI (2005) An Assessment of the Causes of Malnutrition in
Ethiopia. Washington, DC: International Food Policy Research Institute.
Linnemayr, S., H. Alderman, and K. Abdoulaye (2008) Determinants of
Malnutrition in Senegal: Household, Community Variables, and Their
Interaction. Economics and Human Biology 6:2, 252-263.
Mendelson, Sam (2011) Child Malnutrition in India: Why Does It
Persist?
National Institute of Population Studies (1990-91) Pakistan
Demographic and Health Survey. Islamabad: Pakistan and Macro
International Inc.
Nayab and Frooq (2012) Effectiveness of Cash Transfer Programmes
for Household Welfare in Pakistan: The Case of the Benazir Income
Support Programme. Pakistan Institute of Development Economics,
Islamabad: (Working Paper).
NEPAD (2004) Micronutrient Initiative. Draft NEPAD Nutrition
Concept Note. Ottawa, Canada: The Micronutritient Initiative. Photocopy.
Opara, A. Jacinta, E. Helen Adebola, S. Nkasiobi Oguzor, and A.
Sodienye Abere (2011) Malnutrition during Pregnancy among Child Bearing
Mothers in Mbaitolu of South-Eastern Nigeria. Advances in Biological
Research 5:2.
Pakistan Institute of Development Economics, UNICEF and Planning
Commission (2001). National Nutrition Survey. Islamabad: Pakistan
Institute of Development Economics, UNICEF and Planning Commission.
Pakistan, Government of (1985-87) National Nutrition Survey.
Islamabad: Nutrition Division, National Institute of Health.
Pakistan, Government of (2011) National Nutrition Survey. Karachi:
Agha Khan University, Pakistan Medical Research Council and Nutrition
Wing, Cabinet Division.
Pal, Sarmistha (1999) An Analysis of Childhood Malnutrition in
Rural India: Role of Gender, Income and other Household Characteristics.
World Development 27:7.
Rosenberg, I. H., N. W. Soloman, and R. E. Schneider (1977)
Malabsorption Associated with Diarrhea and Intestinal Infections. Am J
ClinNutr, 1248-1253.
Strauss, J. and D. Thoman (1995) Human Resources: Empirical
Modelling of Household and Family Decisions. In J. B. Behrman and T. N.
Srinivasan (eds.) Handbook of Development Economics 3.
Sununtar, Setboonsarng (2005) Child Malnutrition as a Poverty
Indicator: An Evaluation in the Context of Different Development
Interventions in Indonesia. ADB. (ADB Institute Discussion Paper No.
21).
UNICEF (1990) Strategy for Improved Nutrition Status of Children
and Women in Developing Countries. New York. (Policy Review Paper,
E/ICEF/1990/1.6).
WFP and CDC (2005) A Manual: Measuring and Interpreting
Malnutrition and Mortality. World Food Program and Centre for Disease
Control and Prevention.
WHO (2006) Child Growth Standards: Length/Height-for-age,
Weight-for-age, Weight-for-length, Weight-for-height and Body Mass
Index-for-age: Methods and Development. Department of Nutrition for
Health and Development, World Health Organisation.
WHO (2008) Interpreting Growth Indicators. Training Course on Child
Growth, Department of Nutrition for Health and Development, World Health
Organisation.
(1) www.ifpri.org/sites/default/files/publications/oc63ch04.pd
(2) Based on this panel data, Arif and Farooq (2012) have estimated
that between 2004 and 2010, 15 percent of the households moved out of
poverty while 18 percent fell into poverty. Another 9 percent households
were identified as chronic poor, remaining in poverty in two rounds,
2004 and 2010.
(3) UNICEF (2012).DAWN newspaper, October 10, 2012.
(4) DAWN newspaper, October 10,2012.
(5) UN Report titled "Every Women, Every Child: From
Commitment to Action" DAWN newspaper, October 10,2012.
(6) DAWN newspaper, October 10, 2012.
G. M. Arif <
[email protected]> is Joint Director at the
Pakistan Institute of Development Economics, Islamabad. Shujaat Farooq
<
[email protected] > is Assistant Professor at the Pakistan
Institute of Development Economics, Islamabad. Saman Nazir
<
[email protected]> is Staff Demographer, and Maryam Satti
<
[email protected]> is MPhil student at the Pakistan Institute
of Development Economics, Islamabad.
Table 1
Sampled Children by Region and Gender, PPHS-2010
Region Both Sexes Male Female
Total 3218 1666 1552
Urban 844 440 404
Rural 2374 1226 1148
Table 2
Trends in Child Nutrition in Pakistan
% Underweight % Stunted
Data Source Total Rural Urban Total Rural Urban
NNS 1985-7 47.9 -- -- 41.8 -- --
NNS 2001 41.5 42.3 38.7 31 32.5 24.5
NNS 2011 31.5 33.3 26.6 43.7 46.3 36.9
PDHS 1990 40.4 -- -- 50 -- --
PSES 2001 48.2 51.4 41.7 49.7 52.7 43.5
PRHS 2001 -- 56.6 -- -- 64.4 --
PPHS 2010 39.4 39.8 38.1 63.9 64.5 62.1
% Wasted
Data Source Total Rural Urban
NNS 1985-7 10.8 -- --
NNS 2001 11.6 11.2 12.1
NNS 2011 15.1 12.7 16.1
PDHS 1990 9.2 -- --
PSES 2001 -- -- --
PRHS 2001 -- 18.4 --
PPHS 2010 17.9 17.2 19.9
Note: The differences between figures may be due to
methodological variations among these surveys. PDHS 1990-1
used NCHS standard with reference population of children (0-
59) months. The figures reported for NNS 2001 are percent
median with reference population (6-59) months. PRHS, PSES,
PPHS-2010, NNS-2011 are using reference population of 6-59,
0-59, 6-59 and 0-59 months respectively.
Table 3
Child Nutrition Status (Moderate/Severe) by Region, 2010
% Underweight %Stunted
Nutritional
Status of Children Total Urban Rural Total Urban Rural
Normal 56.9 57.7 56.7 31.2 32.6 30.7
Moderate 15.7 15.0 15.9 20.2 23.2 19.2
Severe 23.7 23.1 23.9 43.7 38.9 45.4
Over Weight/Height 3.7 4.2 3.5 4.9 5.3 4.8
Total 100 100 100 100 100 100
%Wasted
Nutritional
Status of Children Total Urban Rural
Normal 61.8 61.9 61.8
Moderate 8.9 9.4 8.7
Severe 9.0 10.5 8.5
Over Weight/Height 20.3 18.2 21.0
Total 100 100 100
Source: Authors' computation from the micro-data of PPHS-2010.
Note: Normal children are healthy children having Z-scores
between -2 and +2 SD, while Z-scores for moderate
malnourished child are below -2 SD and severe malnourished
child are below -3 SD.
Table 4
Incidence of Poverty: A Cross-sectional Analysis of
Three Waves of the Panel Survey (2001, 2004 and 2010)
Survey Year All Provinces Punjab and Sindh
2001--Rural only 27.5 31.3
2004--Rural only -- 24.1
2010-Rural 22.4 27.0
Urban 16.6 18.5
All 20.7 24.4
Source: Arif and Sliujaat (2012).
Table 5
Poverty Dynamics by Region (Rural only) Using Three Waves
(2001, 2004 and 2010)
Punjab
Total Central--
sample North
Change in Poverty (Sindh and (excluding
Status Punjab) Total South) South Sindh
3 Period Poor
(Chronic) 4.01 3.71 1.06 6.46 4.32
2 Period Poor 16.60 10.34 6.17 14.65 23.12
1 Period Poor 30.90 23.97 17.41 30.76 38.12
Never Poor 48.48 61.98 75.36 48.14 34.44
All 100.0 100.0 100.0 100.0 100.0
N (1395) (792) (417) (375) (603)
Source: Arif and Shujaat (2012).
Table 6
Summary Statistics for Dependent and Independent Variables
Determinants Mean Minimum Maximum
WAZ -1.55 -5.98 4.94
HAZ -2.38 -6.01 6.00
WHZ 0.12 -4.99 5.00
Per Capita Expenditure in 2001 (Rs) 1166.68 1048.76 148.88
Child Characteristics
Sex (male =1) 0.53 0 1
Age (in months) 31.36 6 59
Number of Siblings (< 2) 0.21 0 1
2-3 0.35 0 1
4-6 0.26 0 1
7+ 0.06 0 1
Incidence of Diarrhea last 30
days (yes=l) 0.09 0 1
Mother Characteristics
Mother Education (No education) 0.81 0 1
Primary (yes=1) 0.08 0 1
Secondary(yes=1) 0.07 0 1
College(yes=1) 0.04 0 1
Housing and Hygiene
Housing type (Pacca= 1) 0.33 0 1
Community Factor
Toilet in village (in %) 54.87 0 100
LHW presence in village (in %) 57.79 0 100
Determinants S.D N
WAZ 1.96 3540
HAZ 2.20 2742
WHZ 2.22 2280
Per Capita Expenditure in 2001 (Rs) 22102.39
Child Characteristics
Sex (male = 1) 0.50 4604
Age (in months) 14.97 3218
Number of Siblings (< 2) 0.415 6509
2-3 0.489 4214
4-6 0.449 4214
7+ 0.24 4214
Incidence of Diarrhea last 30
days (yes=1) 0.295 4635
Mother Characteristics
Mother Education (No education) 0.49 4635
Primary (yes=1) 0.27 4635
Secondary(yes=1) 0.25 4635
College(yes=1) 0.19 4635
Housing and Hygiene
Housing type (Pacca= 1) 0.47 4616
Community Factor
Toilet in village (in %) 34.02 4604
LHW presence in village (in %) 30.35 4604
Table 7
The Determinants of Child Malnutrition-OLS and 2SLS
Estimates (only 2001 and 2010 Panel Households)
OLS
WAZ HAZ WHZ
Determinants Coeff Coeff. Coe IT.
Per capita exp._200l (Rs) 0.001 ** 0.002 0.001 ***
Per capita exp._2001 (sq) 0.001 0.001 0.001
Sex (male=l) -0.312 * -0.100 -0.233 ***
Child age (months) 0.034 ** 0.023 -0.013
Child [age.sup.2] 0.001 *** 0.001 0.001
Number of Siblings (<2 as
reference)
2-3 0.040 0.077 0.044
4-6 -0.066 0.208 -0.195
7+ 0.154 0.168 -0.292
Diarrhea (yes=l) -0.420 * 0.185 -0.295 ***
Mother's education (no
education as reference)
Primary 0.032 0.187 0.158
Secondary 0.434 * 0.052 0.456
College 0.620 * 0.744 -0.338
Housing Type (Pacca=1) -0.049 -0.032 -0.253
Toilet Facility (% at village
level) 0.005 * 0.001 0.011 **
LHW visited (% at village
level) 0.014 * -0.005 0.012 *
Constant -1.505 * -2.883 0.523 *
N 1,328 998 1,010
2SLS
WAZ HAZ WHZ
Determinants Coe IT. CoefT. CoefT.
Per capita exp._2001 (Rs) 0.001 0.001 0.000
Per capita exp._2001 (sq) 0.001 0.001 0.000
Sex (male=l) -0.328 * -0.090 -0.237 ***
Child age (months) 0.032 *** 0.028 -0.010
Child [age.sup.2] 0.001 0.001 0.000
Number of Siblings (<2 as
reference)
2-3 0.059 0.103 0.026
4-6 -0.056 0.287 -0.218
7+ 0.176 0.267 -0.313
Diarrhea (yes=l) -0.429 * 0.219 -0.307 ***
Mother's education (no
education as reference)
Primary -0.053 0.189 0.129
Secondary 0.394 * -0.020 0.475
College 0.630 * 0.614 -0.292
Housing Type (Pacca=1) -0.063 -0.074 -0.255
Toilet Facility (% at village
level) 0.005 * 0.000 0.012 *
LHW visited (% at village
level) 0.014 * 0.005 * -0.012 *
Constant -1.571 * -3.208 * 0.589 *
N 1.311 986 1,873
Source: Authors' estimation from the micro-data of PRHS 2001
and PPHS 2010.
Note: * significant at 1 percent, ** significant at 5
percent, *** significant at 10 percent.
Note: Per capita expenditure of 2001 is instrumented.
Table 8
The Impact of Poverty and Poverty Dynamics on Child
Underweight--OLS Regression
Model-1
Determinants Coeff. Std. Error
Poverty status in 2004 (poor=1) -0.257 0.172
Poverty dynamics (non-poor as reference)
Chronic (poor in 2-periods) -- --
Transitory (moved into or out of poverty) -- --
Sex (male=1) -0.287 ** 0.118
Child age (months) 0.025 0.019
Child age2 0.001 0.001
Number of Siblings (<2 as reference)
2-3 0.086 0.153
4-6 -0.090 0.160
7+ 0.043 0.251
Diarrhea (yes=l) -0.604 * 0.173
Mother's education (no education as
reference)
Primary 0.281 0.226
Secondary 0.399 0.295
College -0.483 0.457
Housing Type (Pacca=1) 0.104 0.140
Toilet Facility (% at village level) 0.009 * 0.002
LHW visited (% at village level) 0.012 * 0.003
Constant 1.536 * 0.341
N 966
Model-2
Determinants Coeff. Std. Error
Poverty status in 2004 (poor=1) -- --
Poverty dynamics (non-poor as reference)
Chronic (poor in 2-periods) -0.109 0.207
Transitory (moved into or out of poverty) -0.141 0.132
Sex (male=1) -0.292 ** 0.119
Child age (months) 0.027 0.019
Child [age.sup.2] 0.001 0.001
Number of Siblings (<2 as reference)
2-3 0.079 0.155
4-6 -0.094 0.162
7+ 0.026 0.254
Diarrhea (yes=1) -0.614 * 0.175
Mother's education (no education as
reference)
Primary 0.261 0.228
Secondary 0.443 0.300
College -0.493 0.460
Housing Type (Pacca=1) 0.087 0.142
Toilet Facility (% at village level) 0.010 * 0.002
LHW visited (% at village level) 0.012 * 0.003
Constant -1.538 * 0.348
N 954
Source: Authors' estimation from the micro-data of PRHS 2004
and PPHS 2010.
Note: * significant at 1 percent, ** significant at 5
percent, *** significant at 10 percent.
Table 9
OLS for Underweight Children (Perceived Food Security)
Model-3
Determinants Coeff. Std. Error
Food Shortage (yes=l) 0.1790 0.788
Sufficient Food (yes=l) -- --
Sex (male=l) -0.2707 * 0.0764
Child age (months) 0.0251 ** 0.0123
Child [age.sup.2] -0.0002 0.0002
Number of Siblings
2-3 -0.0667 0.0929
4-6 -0.2056 ** 0.1036
7+ 0.0322 0.1762
Diarrhea -0.3954 * 0.1210
Mother's Education
Primary 0.1087 0.1410
Secondary 0.1226 0.1566
College 0.2596 0.2060
Housing Type (Pacca=1) -0.0987 0.0905
Toilet Facility (% at village level) 0.0057 * 0.0014
LHW visited (% at village level) 0.0085 * 0.0013
Region -0.2373 ** 0.1050
Constant -1.4245 * 0.2130
N 2,479
Model-4
Determinants Coeff. Std. Error
Food Shortage (yes=1) -- --
Sufficient Food (yes=1) 0.094 0.079
Sex (male=1) -0.268 * 0.076
Child age (months) 0.024 ** 0.012
Child [age.sup.2] 0.000 0.000
Number of Siblings
2-3 -0.066 0.093
4-6 -0.221 ** 0.104
7+ 0.032 0.176
Diarrhea -0.397 * 0.121
Mother's Education
Primary 0.110 0.141
Secondary 0.134 0.157
College 0.263 0.207
Housing Type (Pacca=1) -0.094 0.090
Toilet Facility (% at village level) 0.006 * 0.001
LHW visited (% at village level) 0.009 * 0.001
Region -0.246 ** 0.105
Constant -1.307 * 0.215
N 2,476
Source: Authors' estimation from the micro-data of PPHS 2010.
Note: * significant at 1 percent, ** significant at 5
percent, *** significant at 10 percent.
Table 10
OLS for Underweight Children (Perceived Food Security)
Determinants Rural Only
Coeff. Coeff.
Food shortage (yes=l) 0.174 --
Sufficient Food (yes=l) -- -0.024
Sex (male=l) -0.331 * -0.328 *
Child age (months) 0.017 0.016
Child [age.sup.2] 0.000 0.000
2-3 -0.002 -0.002
4-6 -0.215 *** -0.233 **
7+ 0.055 0.055
Diarrhea -0.444 * -0.443 *
Primary -0.096 -0.074
Secondary 0.374 ** 0.383 *
College 0.301 0.326 **
Housing Type (Pacca=1) -0.067 -0.055
Toilet Facility (% at village level) 0.005 * 0.006 *
LHW visited (% at village level) 0.009 * 0.010 *
Constant -1.350 * -1.261 *
N 1,849 1,847
Determinants Urban only
Coeff. Coeff.
Food shortage (yes=1) 0.185 --
Sufficient Food (yes=1) -- -0.389
Sex (male=l) -0.082 -0.088
Child age (months) 0.023 0.025
Child [age.sup.2] 0.000 0.000
2-3 -0.192 -0.214
4-6 -0.122 -0.116
7+ 0.001 -0.039
Diarrhea -0.214 -0.163
Primary -0.497 * -0.550 *
Secondary -0.122 -0.174
College 0.395 0.315
Housing Type (Pacca=1) -0.137 -0.147
Toilet Facility (% at village level) 0.008 ** 0.007 **
LHW visited (% at village level) 0.007 * 0.007 *
Constant -1.910 * -1.634 *
N 630 629
Source: Authors' estimation from the micro-data of PPHS 2010.
Note: * significant at I percent, ** significant at 5
percent, *** significant at 10 percent.
