patterns and trends - world...
TRANSCRIPT
51
Reduction in childmalnutrition is another MDGrelated to an improvement in child welfare.44 Childmalnutrition significantlyincreases the risk of infant andchild death, with someestimates suggesting that childmalnutrition is responsible forhalf or more of child deaths inthe developing world.45 TheNFHS-2 data analyzed in theprevious chapter showed astrong relationship betweenunder-five child mortality ratesand child underweight ratesacross the various regions ofIndia (see Figure II.22 inChapter II). There is also a large body of evidence from around the world relating under-nutrition inchildhood to lower levels ofschool performance, cognitive
development, health, and,ultimately, to lower levelsof labor productivity inadulthood. Thus, the econo-mic and human costs of childmalnutrition in India are likelyto be very high.46
The millennium develop-ment goal is to reduce thepercentage of underweightchildren by one-half between1990 and 2015.47 For India,this would imply a reduction inthe child underweight ratefrom 54.8% in 1990 to 27.4%in 2015.48
Patterns and TrendsLevels
Child malnutrition rates inIndia are extraordinarily high.The NFHS-2, which is themost recent household survey
containing information onchild nutrition, indicates thatnearly one-half of children aged0-35 months are underweightor stunted,49 which translatesinto approximately 37 millionmalnourished children. About18-23 percent of children areseverely underweight or stuntedin the sense of being more thanthree standard deviations belowthe relevant NCHS standards.This suggests that Indianchildren suffer from shortterm,acute food deficits (as reflectedin low weight-for-age) as well asfrom longer-term, chronicunder-nutrition (as manifestedin high rates of stunting).50
TrendsBoth the NFHS-1,
conducted in 1992-93, and theNFHS-2, conducted in 1998-
��������������
44 Another important form of malnutrition that is not pursued in this report is inadequate consumption of micronutrients, such as Vitamin A,
iron and iodine.
45 For instance, based on worldwide evidence, Pelletier and Frongillo (2002) estimate that a 5 percentage point reduction in the prevalence of
low weight-for-age could reduce child mortality by about 30% and under-5 mortality by 13%.
46 The World Bank (1998) suggests that the cost of undernutrition in India is at least US$10 billion annually in terms of lost productivity,
morbidity and mortality.
47 While the nutrition MDG is based on the weight-for-age indicator, it should be recognized that there are other important indicators of child
malnutrition. Weight-for-age is an indicator of both short- and long-term malnutrition. Height-for-age or stunting is a better indicator of
long-term (cumulative) malnutrition, while weight-for-height or wasting is generally considered the appropriate indicator for tracking short-
term fluctuations in nutritional status.
48 The most reliable estimate is available from the NFHS-1 for 1992-93. The rate shown for 1990 is projected from the change observed between
1992-93 and 1998-99.
49 As in the literature, a child is considered underweight when his or her weight-for-age is more than two standard deviations below the NCHS
reference weight. A child is stunted when his or her height-for-age is more than two standard deviations below the NCHS reference. Severe
underweight and stunting occur when the relevant nutrition indicator is more than three standard deviations below the NCHS reference.
50 Wasting rates (i.e., low weight-for-height) are significantly lower than underweight or stunting rates, but this is typically the case in most low-
income countries.
52
Attaining the Millennium Development Goals in India
99, obtained information onchild anthropo-metry. Acomparison of the estimates ofthese two surveys indicates amodest decline of about 11%(from a rate of 52.7% to 47%)during the 6-year period –amounting to an annual rate ofdecline of 1.9%. In contrast,underweight rates in neighbor-ing Bangladesh fell from 68%in 1992 to 51% in 2000 – anannual rate of decline of 3.6% (World Bank 2003). In Vietnam, the childunderweight rate fell from49% in 1993 to 36% in 1998– an annual rate of decline of6.1% (World Bank 1999)!India thus appears to be an under-performer inreducingchild malnutrition during the 1990s.
Inter-state variationsAn average child under-
weight rate of 47% masks widevariations in child malnutri-tion across states. Child under-weight rates vary from a low of24-28% in the Northeasternstates and Kerala to 51-55% inthe states of Bihar, Rajasthan,Uttar Pradesh, MadhyaPradesh and Orissa (FigureIII.1). Likewise, the decline inchild underweight rates overtime has also varied greatlyacross states. In Punjab, forinstance, the child under-weight rate fell at an annualrate of 7.6% between 1992-93and 1998-99, while Rajasthansaw an increase of 2% per annum in the childunderweight rate during thesame period. Based on 1992-
93 and 1998-99 values, FigureIII.2 shows the 2015 MDgoals for each of the states.
Although the proportion ofchildren aged 0-3 years in Indiawho were under-weightdeclined from 52.7% in 1992-93 to 47% in 1998-99,the absolute number ofunderweight children hardlychanged over this periodbecause of a large increase in
the population of children aged0-3. Indeed, only about aquarter million fewer Indianchildren aged 0-3 wereunderweight in 1998-99 ascompared to 1992-93. Therewere wide variations in theabsolute decline in the numberof underweight children acrossstates. Tamil Nadu and AndhraPradesh each saw declines inthe absolute number ofunderweight children of about
20
25
30
35
40
45
50
55
60
Nag
alan
d
Aru
nach
alPra
desh
Ker
ala
Man
ipur
Miz
oram G
oa
Punja
b
Jam
mu
&K
ashm
ir
Har
yana
Del
hi
Ass
am
Tamil
Nad
u
And
hra
Prade
sh
Meg
hala
ya
Him
acha
l Prade
sh
Kar
nata
ka
Guj
arat
Indi
a
Wes
t Ben
gal
Mah
aras
htra
Raj
asth
an
Utta
r Prade
sh
Bih
ar
Oris
sa
Mad
hya
Prade
sh%
of
chil
dre
nu
nd
erw
eig
ht
in1
99
8-9
9
-10
-8
-6
-4
-2
0
2
4
An
nu
alra
teo
fch
ang
ein
chil
du
nd
erw
eig
ht
rate
(%),
19
92
-98
Child underweight rate (%), 1998-99 Annual rate of change (%), 1992-98
Figure III.1: Percent of children aged 0-35 months who areunderweight, 1998-99, and annual % change in this rate between
1992-93 and 1998-99, by state
0
5
10
15
20
25
30
35
Manipur
KeralaNagaland
Mizoram
Goa
Haryana
Arunachal Pradesh
Delhi
Jammu & Kashmir
Meghalaya
Rajasthan
Himachal PradeshPunjabTamil Nadu
Andhra Pradesh
Gujarat
Assam
Karnataka
Maharashtra
India
Orissa
West Bengal
Uttar Pradesh
Madhya PradeshBihar
Figure III.2 : Child underweight rate MDGs by state, 2015(% of children 0-35 months underweight)
0.5 million, but Rajasthan andBihar saw an increase of 0.5million each in the number ofunderweight 0-3 year olds(Figure III.3). Uttar Pradeshand Madhya Pradesh eachrecorded an increase of about0.2 million underweightchildren aged 0-3 during thesame period.
Figure III.4 suggests thatthere is no systematicassociation between the initial level of childmalnutrition in 1992-93 andthe rate of decline inmalnutrition between 1992-93and 1998-99. Nagaland, whichhad a child underweight rate ofonly 27.7% in 1992-93,experienced an annual rate ofdecline of 2.3% over thefollowing 6 years – the samerate of decline experienced byBihar, which began with a much higher childunderweight rate of 62.6% in 1992-93.
Is child malnutrition relatedto living standards? The cross-state data suggest an inverse,albeit not perfect, associationbetween the child underweightrate and gross state domesticproduct per capita (FigureIII.5). That Kerala emerges as apositive outlier – having amuch lower child underweightrate than would be suggestedby its per capita income – is no
big surprise, but the fact thatGujarat and Maharashtra havesignificantly higher childunderweight rates relative totheir per capita income issurprising. On the other hand,Andhra Pradesh, which is amiddle-income state, has alower child underweight ratethan would be predicted by itsgross state domestic productper capita. This suggests thatcultural and social – not justeconomic – factors have animportant role to play indetermining child malnutritionin India.
Public Spending onNutrition
Much of the publicspending on child nutrition inIndia takes place on theIntegrated Child Develop-ment Services program. This program consists ofanganwadi centers (AWCs) ineach village, typically staffedby a village woman with 5-8years of schooling and an
53
Child Malnutrition
-53
9
-52
0
-31
9
-28
9 -17
5
-15
3
-58
-51
76 1
07 1
82
19
6 24
8
46
1
57
1
-2
-5-6-8
4220
-600
-400
-200
0
200
400
600
Tam
ilN
adu
Andhra
Prades
h
Wes
t Ben
gal
Punjab
Ass
am
Kar
natak
a
Har
yana
Guja
rat
Aru
nachal
Prades
h
New
Del
hi
Him
achal
Prades
h
Miz
oram
Goa
Nag
alan
d
Meg
halay
a
Man
ipur
Oriss
a
Ker
ala
Utta
r Prades
h
Mad
hyaPra
desh
Mah
aras
htra
Raj
asth
an
Bih
ar
Figure III.3 : Change in absolute number of underweight children 0-35months of age between 1992-93 & 1998-99, by state ( '000 children )
-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
25 30 35 40 45 50 55 60 65
% of children 0-35 months who were underweight in 1992-93
An
nu
alra
teo
fd
ecli
ne
inch
ild
un
der
wei
gh
tra
te,
19
92
-98
Figure III.4 : Child underweight rate (%) in 1992-93 andpercent annual decline in underweight rate, 1992-98, by state
assistant. The anganwadiworker receives a cash incomeof Rs. 1,000 per month toprovide growth monitoring,food supplementation, andpre-school education totargeted children aged 0-6years in the village. Althoughthe program covers all thevillages in the country, recentsurveys from a few statessuggest that relatively few(only about 10-30%) childrenaged 0-6 years in states such as Uttar Pradesh, MadhyaPradesh and Rajasthanregularly attend the AWC intheir community (Heywood2003). This may be becausethe amount of foodsupplementation provided tochildren is meager or irregularor both. While the Centralgovernment pays for the
salaries of the anganwadiworker and assistant, the individual states areresponsible for lifting the foodgrains from the stocks of theFood Corporation of Indiaand paying for the cost oftransporting and distributing
these food grains to theAWCs. This is the componentof the ICDS that is typicallyunder-funded (World Bank1998, 2001).
Figure III.6 shows the totalamount (excluding training)spent by various states in thecountry on the ICDS program.The amounts are expressed interms of spending per child aged0-6 years, since that is the targetgroup of the ICDS.51 Twoobservations can be made fromthis figure; first, the amountsspent by most states are low –typically below Rs. 200 per childper annum. Second, there arelarge inter-state disparities inspending on nutrition. Thepoor, high-malnutrition states,such as Bihar, Uttar Pradesh,Madhya Pradesh and Rajasthan,spend only Rs. 30-50 per child,while Gujarat, Punjab, and
54
Attaining the Millennium Development Goals in India
25
30
35
40
45
50
55
4,500 7,500 10,500 13,500 16,500
Gross state domestic product per capita
%o
fch
ild
ren
0-3
5m
on
ths
wh
oar
eu
nd
erw
eig
ht
Kerala
Maharashtra
Gujarat
Andhra Pradesh
Figure III.5 : Child underweight rate (%) and gross state domesticproduct per capita across states, 1998-99
31404451525961646573
8389909199
150167170
182208
225231234
502618
707
0 100 200 300 400 500 600 700 800
BiharMadhya Pradesh
Uttar PradeshNew DelhiRajasthan
MaharashtraAndhra Pradesh
West BengalAssamKerala
KarnatakaGujaratOrissaPunjab
HaryanaJ & KashmirTamil NaduMeghalaya
GoaHimachal Pradesh
TripuraManipur
SikkimMizoram
ArunachalPradeshNagaland
Figure III.6 : Government expenditure on ICDS (child nutrition) program(excluding training) per child aged 0-6 years, 1999-2000 (nominal Rs.)
51 Even though the target group is 0-6 years, the ICDS has historically focused on children aged 3-6 years. The focus on 3-6 yearolds is a major design flaw of the ICDS, since malnutrition typically sets in much earlier in childhood. Experience from othercountries, as well as from other nutritional interventions in India (e.g., Tamil Nadu Integrated Nutrition Project, discussed inBox III.1), has shown that child malnutrition can be addressed much better by targeting nutritional supplementation tochildren in the younger age groups.
Haryana spend Rs. 90-100.Tamil Nadu’s expenditure isabout Rs. 170, while spendingin the Northeastern states isabove Rs. 500.
Is there an associationbetween child underweightrates and per-child spendingon the ICDS? Pooled data on14 states for two years – 1992-93 and 1998-99 – suggestan inverse association (FigureIII.7). However, since there isno control for other variables,such as per capita income, the association does notnecessarily indicate a positiveeffect of public spending on nutrition. Indeed, theredoes not appear to be anyassociation between changes inthe level of per-child spendingon ICDS and changes in thechild under-weight rate(Figure III.8).
However, the data dosuggest an inverse associationbetween changes in the childunderweight rate and changesin gross state domestic productper capita (Figure III.9). Statesthat experienced greater growthin real gross state domestic
product per capita between1992-93 and 1998-99experienced larger declines inthe child under-weight rateduring the same period.
Intra-State VariationsMap III.1 shows the child
underweight rate in 1998-99for different regions inthe country.52 The childunderweight rate ranges from alow of 11% in the Hills regionof Assam to a high of 60% in Chhattisgarh. More than a third of the regions have child underweight rates of 50%or greater, confirming theubiquity of child malnutritionin the country.
What is even moreworrying is that more than aquarter of the regions in thecountry for which under-weight data are available
55
Child Malnutrition
25
30
35
40
45
50
55
60
65
15 25 35 45 55 65 75 85 95 105
Real public spending on ICDS per child 0-6 (1993-94 Rs.)
Un
de
rwe
igh
tra
tea
mo
ng
0-3
ye
ar
old
s(%
)
Figure III.7 : Underweight rate among 0-3 year olds (%) and realexpenditure on the ICDS program per child 0-6 years,
across states, 1992-93 & 1998-99
-40
-30
-20
-10
0
10
-25 0 25 50 75 100 125 150 175 200 225
% change in gov't exp on the ICDS program (nutrition) per child 0-6
%c
ha
ng
ein
un
de
rwe
igh
tra
tea
mo
ng
0-3
ye
ar
old
s
Figure III.8 : Changes in child (0-3) underweight rates (%) and inreal government expenditure on the ICDS program per child
0-6 across states, 1992-93 to 1998-99
52 It is not possible to obtain reliable district-level estimates of child malnutrition in India. Regional-level estimates are morereliable, given the sample size of the NFHS-2 survey. Data from the NFHS-2 are not available for a few of the regions.
experienced an increase inchild underweight ratesbetween 1992-93 and 1998-99 (Map III.2). These regionsare scattered around thecountry – in the poor states(Jharkhand, Bihar, Orissa,and Uttar Pradesh) but also inthe more prosperous states,such as Gujarat, Maharashtra,and Haryana.
Concentration ofChild Malnutrition
Child malnutrition inIndia is not as heavilyconcentrated geographically
as, say, infant deaths. Forinstance, the four states –Uttar Pradesh, MadhyaPradesh, Bihar and Rajasthan– account for 51% of allinfant deaths in the country, but for 43% of all under-weight children under the age of 3 (FigureIII.10). Nevertheless, inabsolute terms, these are large numbers.
Disaggregating further,one finds that childmalnutrition is relativelyconcentrated across districtsand villages in the country. Amere 10% of districts andvillages account for 27-28% –and a quarter of districts andvillages account for more thanhalf – of all the underweightchildren in the country(Figure III.11).53
56
Attaining the Millennium Development Goals in India
-40
-35
-30
-25
-20
-15
-10
-5
0
5
10
15
10 15 20 25 30 35 40 45
% change in real per capita GSDP (1993-94 Rs.)
%c
ha
ng
ein
ch
ild
un
de
rwe
igh
tra
tea
mo
ng
0-3
ye
ar
old
s
Figure III.9 : Changes in child underweight rates and in real per capitagross state domestic product across states, 1992-93 to 1998-99
% of children 0-35 months who
are underweight, 1998-99
% Children
>= 5035 to 4925 to 34 0 to 34 Missing
Map III.1 : Regional estimates of underweight childrenaged 0-35 months, 1998-99
53 Since, as noted in chapter II, the NFHS-2 covered only a fraction of all the villages in the country and the number of sampledhouseholds in each village is too small to be representative, these numbers are merely indicative of possible patterns. It wouldbe worthwhile to explore the use of promising new methodologies available to more accurately identify villages with the largestnumber of underweight children in the country, so that nutritional interventions could be better targeted to these villages.
Proximate Causes ofChild MalnutritionInfant feeding practices
An important correlate ofchild nutritional status isnutrient intake, which in turndepends on the nature andduration of feeding (includingbreastfeeding) practices. Feed-ing practices are especiallycritical during the first fewdays and months of an infant’slife, since growth is faster andprotection against illnessesand infections is most neededduring this crucial period.Ideally, a baby should be putto the mother’s breastimmediately after birth.However, NFHS-2 dataindicate that nearly one-halfof Indian babies have to waitto be breastfed for more thana day after they are born(Figure III.12). In the poorstates of Bihar, Uttar Pradesh,Madhya Pradesh, Orissa andRajasthan, this ratio is evenhigher (62%). This is how the
cycle of child malnutritionbegins very early in an Indian
child’s life. The delay inbreastfeeding is often relatedto an incorrect perceptionthat the first breast milk(colostrum) is an inferiorfood, when in fact colostrumis rich in antibodies andhighly beneficial to the new-born infant.
Another common feedingpractice in India that has adverseimplications for child malnutri-tion is the early termination ofexclusive breast-feeding andintroduction of supplementaryfeeding. One reason whymothers give up exclusive breast-feeding early is their perceptionthat they are producing
57
Child Malnutrition
������������������ ���������
��������������� ���������
�����������������
��������
���� ����
���� �����������������
Map III.2 : Regional estimates of the change in proportion ofunderweight children aged 0-35 months, 1992-93 to 1998-99
50
33
61
69
77
8287
91
117 6 5 4 4 4 3 3 3 2 2 2 2 2 1 1
11 1011
0
10
20
30
40
50
60
70
80
90
100
Utta
r Prade
sh
Mad
hya
Prade
sh
Raj
asth
an
Bih
ar
Mah
aras
htra
Oris
sa
W. B
enga
l
Kar
nata
ka
Guj
arat
Tamil
Nad
u
And
hra
Prade
sh
Him
acha
l Pra
desh
Har
yana
Jam
mu
and
Kas
hmir
Ass
am
Punja
b
New
Del
hi
Man
ipur
Ker
ala
Meg
hala
ya
Cu
mu
lati
ve
co
ntr
ibu
tio
n(%
)
Cumulative share in nationwide number of underweight 0-35 month
oldsShare in nationwide number of underweight 0-35 month olds
94
Figure III.10 : Contribution of 20 states to the national numberof underweight 0-35 month olds, 1998-99
insufficient quantities of milkdue to their poor nutrition andheavy workload. Prematureintroduction of foods other thanbreast milk greatly increases therisk of infection in the smallinfant, and this sets in motion theprocess of malnutrition. It alsoputs the infant at greater risk ofmalnutrition, since weaning dietsare often inadequate in India.Supplementary feeding beginswith a thin gruel of rice, oftenheavily diluted with water andwith some vegetables or legumesadded as a relish depending onseason and availability, butgenerally in very small quantities.The consequent low energydensity of this weaning food leadsto a reduced intake of caloriesand protein, and is an importantcause of growth faltering duringthe weaning period, from sixmonths to two years of age. The
NFHS-2 data indicate thatnearly 49% of children aged 0-4months and 58% of childrenaged 0-6 months are not exclusiv-ely breastfed (i.e., supplementaryfeeding is introduced), which is not in line with the
recommendations of WHO and UNICEF that exclusivebreastfeeding continue for thefirst six months of a child’s life.
InfectionsIllness and infection,
especially diarrheal infections,are strongly associated with childmalnutrition. Infections reducethe ability of the body to absorbcritical nutrients from food,which in turn leads tomalnutrition.54 The NFHS-2data indicate that an averageinfant begins suffering fromdiarrheal diseases very early inhis or her life; by the age of 6months, he or she has alreadyexperienced an average of 2.2diarrheal episodes, and by the age of 12 months, 5.2 illness episodes (Figure III.13).Diarrhea is even more prevalentamong infants in the poor states
58
Attaining the Millennium Development Goals in India
0
10
20
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60 70 80 90 100
Cumulative % of villages or districts (ranked by number of underweight children)
Cu
mu
lati
ve
%o
fall
un
derw
eig
ht
ch
ild
ren
inth
e
co
un
try
Districts
Villages
Figure III.11 : Cumulative distribution of all underweight 0-35month old children in India across villages and districts, 1998-99
8
2517
18
29
2311
12
12
54
30
42
84 6
0%
20%
40%
60%
80%
100%
Bihar, Madhya Pradesh, Orissa,
Rajasthan, Uttar Pradesh
Other States All India
%o
fch
ild
ren
by
tim
eo
ffi
rst
bre
ast-
feed
ing
afte
rb
irth
Immediately at birth < 1 day 1 day 2-3 days > 3 days
Figure III.12 : Initiation of breast-feeding after birth,by groups of states, 1998-99
54 The association reflects two-way causality; while infections lead to malnutrition, malnourished children are more susceptibleto infections.
of Bihar, Madhya Pradesh, UttarPradesh, Orissa and Rajasthan.The NFHS-2 data also showthat children who have suffereda diarrheal episode are 15%more likely to be underweightthan children who have notexperienced diarrhea.55
Maternal weight andlow birth weight
For a large number ofIndian children, malnutritionbegins very early in life – whenthey are born with low birthweight. As Figure III.14 shows,children born with a weight ofless than 2.5 kgs are atsignificantly greater risk ofsubsequent malnutrition thanchildren whose birth weight isabove 2.5 kgs. Nationally,about 22% of births classify as
low birth weights, with widevariations across states. InUttar Pradesh and MadhyaPradesh, more than a third ofall children weigh less than 2.5
kgs at birth. Even in Kerala, theproportion of low birth weightsis as high as 18%.
Low birth weight in turn isdetermined by a number offactors, but important amongthem is maternal nutrition.Malnourished or low-weightmothers are more likely to givebirth to low-weight babies,which implies that children oflow-weight mothers are morelikely to be malnourished thanchildren of heavier mothers.Figure III.14 shows this trend;nearly three-quarters of childrenunder 3 born to mothers whoseweight is less than 35 kgs. arelikely to be underweight. Thisratio drops to just over a quarterfor mothers weighing 50 kgsor more.
59
Child Malnutrition
55 It is not just diarrhea that reduces nutrient absorption of the body; even repeated bouts of fever (indicating infection) andacute respiratory infections can slow down weight gain and lead to malnutrition. It is unlikely however that a single bout ofdiarrhea, fever or cough could lead to severe malnutrition. What is more likely is that the children who report being ill duringthe two-week reference period are the ones who repeatedly come down with diarrhea, fever or cough. It is this chronic feverand cough that is associated with a higher risk of malnutrition. Indeed, malnutrition also increases a child’s susceptibility toinfections, and so the ‘causality’ can go in the reverse direction.
0
1
2
3
4
5
6
0 1 2 3 4 5 6 7 8 9 10 11 12
Age (months)
0
1
2
3
4
5
6
Bihar, Madhya Pradesh, Orissa,Rajasthan, Uttar Pradesh
All India
Other States
Figure III.13 : Cumulative number of diarrheal infectionsexperienced by infants, by age (months), 1998-99
74
61
46
28
47
36
26
0
10
20
30
40
50
60
70
80
<35 35-39 40-49 >=50 < 2,500 gms 2,500 - 2,999gms
>= 3,000 gms
Mother's weight (kgs.) Birth weight (gms.)
Figure III.14 : Percent of children under 3 who are underweight, bymother's weight and birth weight of child, 1998-99
Socio-economicCorrelates of ChildMalnutritionLiving standards
There is remarkably littlevariation in child under-weight rates across economic groups. Indeed, theprevalence of child malnutri-tion is virtually identicalacross the bottom fourconsumption quintiles, withonly the top quintile(representing the richest 20%of individuals) showingsignificantly lower childunderweight rates (FigureIII.15).56 The fact that nearly a third of the topconsumption quintile in thecountry – a group that is likely to have goodeconomic access to food – ismalnourished suggests thatcultural and social factorshave an important role to
play in determining childmalnutrition in India. Thisis also consistent with theevidence, cited earlier, thatthe child underweight rateis relatively high inprosperous states like Gujaratand Maharashtra.
Social groupsThe NFHS-2 data also
indicate somewhat higher childunderweight rates for schedul-ed castes and tribes and otherbackward castes relative to theforward castes (Figure III.16).Scheduled castes and otherbackward castes appear to be ata greater disadvantage relativeto the forward castes in thepoor states than in the non-poor states.
Maternal schoolingThere is a large literature
from around the world thatdocuments the many benefitsof maternal schooling for childoutcomes – infant and childmortality, child nutrition, andchild schooling. The NFHS-2data show a sharp decline inthe incidence of childmalnutrition with mother’sschooling (Figure III.17).
60
Attaining the Millennium Development Goals in India
56 Consumption quintiles are based on predicted household consumption expenditure per capita, the estimation of which isdescribed in footnote 24 in chapter II.
5757
52
44
54
48
39
37
56
54
48
42
35
40
45
50
55
60
Scheduled Tribe Scheduled Caste Other Backward Caste Forward caste
Poor States Other States All States
Figure III.16 : Child underweight rates (%) of various socialgroups in poor and other states, 1998-99
57 5
8
52
48
31
55
54
48
44
29
56 5
7
50
47
31
55
53
51
44
29
25
30
35
40
45
50
55
60
Poorest Second Third Fourth Richest
Per capita expenditure quintile
Females Males Poor states Other states
Figure III.15 : Child underweight rates (%) by per capitaexpenditure quintile, 1998-99
Children of mothers with noschooling are nearly two timesmore likely to be underweightthan children of mothers withmore than 8 years of schooling.Children of mothers with 6-8years of schooling are morethan two times more likely tobe underweight than childrenof mothers with more than 12years of schooling. Interestin-gly, unlike the case of infantmortality, the gender disparityin child underweight rates doesnot seem to narrow withmother’s schooling.
Birth orderAs with infant mortality,
the birth order of a child has asignificant association with theprobability of him or her beingunderweight. The disparityacross birth order is greater for girls than for boys. For instance, while theunderweight rate is 42% forfirst-born girls under 3, it is as high as 58% for birth
order four or higher girls(Figure III.18).
InfrastructureGiven that repeated bouts
of infection, especially gastro-intestinal and diarrheal, are animportant reason for childmalnutrition in India, access tosafe drinking water andsanitation can improve child
nutrition by reducing a child’sexposure to water- and vector-borne diseases. Likewise, accessto electricity can also improvenutritional status by improvingthe hygiene, cooking andhealth practices in thehousehold and in thecommunity. Rural roads enableeasier access to marketsand health workers andthereby better information tochild nutrition-improvinginformation.
The NFHS-2 data show significant associationsbetween child underweightrates and the infrastructuralvariables (Figure III.19). Thechild underweight rate is abouta third lower in villages havingregular electricity supply thanin villages with no electricity.Likewise, access to piped waterand some toilet facilities isassociated with significantlylower child underweight rates,although piped water access
61
Child Malnutrition
58
50
42
3231
20
54
48
37
32
27
18
15
20
25
30
35
40
45
50
55
60
None 1-5 years 6-8 years 9-10 years 11-12 years >12 years
Mother's schooling
Females Males
Figure III.17 : Child (0-3 years) underweight rates (%) by sexand by mother's schooling, 1998-99
42
4141
48
44
46
58
53
55
35
40
45
50
55
60
Females Males Both
1 2-3 4 & above
Birth order
Figure III.18 : Child underweight rates (%), by birthorder and sex, 1998-99
appears to be less important tochild malnutrition in the non-poor states.
Multivariate Analysisof Child Malnutrition
To examine the likelihood ofthe various states in Indiaattaining the child underweightMD goal, we have estimated amultivariate model of childunderweight rates, using theNFHS-2 unit record data (at the child level).57 Themultivariate model has theadvantage of controlling forseveral variables that may besimultaneously associated withchild malnutrition. The estima-tion results are reported inAnnex Table 2, while only thebroad findings of the empiricalanalysis are discussed here.
The multivariate modelconfirms most of the bivariate
relationships discussed earlier.Older children are observed to have a higher risk ofmalnutrition, but at a decreas-ing rate, such that the risk ofmalnutrition peaks at age 24months. Mother’s age reducesthe probability of a child being underweight, but at a decreasing rate. Controllingfor other factors, childrenbelonging to scheduled castes,scheduled tribes and otherbackward castes are significan-tly more likely to beunderweight than childrenbelonging to forward castes. Aswith infant mortality, girls per se are not at significantlygreater risk of malnutritionthan boys, but sex interactswith birth order such thathigher birth order girls have asignificantly greater risk ofbeing underweight than higherorder boys.
Infrastructure generally hasstrong associations with childmalnutrition. Children inhouseholds having no access toa toilet are, on average, 8.6%more likely to be underweightthan children in householdshaving access to a toilet.Surprisingly, piped wateraccess has no significant(independent) association withthe probability of malnutri-tion, probably reflecting thefact that it is highly correlatedwith toilet access. As in the caseinfant mortality, access toelectricity is strongly associatedwith child malnutrition, butthe association is much weaker with irregular electricitysupply. Finally, proximityto sealed (pucca) roads isalso associated with asharp reduction in childunderweight rates.
Other variables that aresignificantly associated withchild malnutrition are themother’s schooling, whetherthere was a medical professionalat the time of the child’s birth,log of predicted monthlyconsump-tion expenditure percapita (proxying for thehousehold’s income and livingstandard), and log ofgovernment expenditure onnutritional programs per childaged 0-6 years in the child’s state of residence. Thesignificant association betweengovernment nutritionalexpenditures per child and
62
Attaining the Millennium Development Goals in India
57 Since the dependent variable in the model is a dichotomous variable (i.e., whether or not a child is underweight), the modelhas been estimated by the maximum-likelihood probit method.
57
52
50
53
40
55
36
56
47
40
43
40
48
32
30
35
40
45
50
55
60
Not available Irregular Regular None Yes None Some
Village electricity supply Piped water access Toilet access
Poor states Other states
Figure III.19 : Child underweight rates (%), byinfrastructure access, 1998-99
child malnutrition is supportedby the finding, discussedbelow, that the risk of a childbeing underweight is inverselyassociated with the presenceof an Integrated ChildDevelopment Services (ICDS)center in the child’s village of residence. Much of the government nutritionalexpenditure in India is on theICDS program. The empiricalresults further indicate that the(inverse) association betweenchild malnutrition and levelsof government nutritionalexpenditure is stronger in thepoor states than in the non-poor states. However, neitherpublic spending on health andfamily welfare per capita norper capita GDP in the child’sstate of residence has asignificant association withchild underweight rates.58
Simulations to 2015Based on the multivariate
probit model estimated earlier,we have undertakensimulations of the childunderweight rate for the poorand the non-poor states under different interventionscenarios. These are shown inTable III.1. If the poor stateswere simply brought up to the national average in terms of coverage of sanitation,road access, electricity, medical
attention at time of delivery,female schooling, householdincome (consumption) andpublic spending on nutritionper child, the cumulativereduction in the childunderweight rate would be ofthe order of about 8 percentagepoints (or 15%). If themagnitude of the proposedinterventions were scaled up so as to bring the poor states to the average level prevailingin the non-poor states, the cumulative reduction in the child underweight ratewould be 21 percentage pointsor 38%.
Below we use the probitresults discussed above tosimulate cumulative changes inthe child underweight rate inthe poor states from 2000
to 2015 based on certainintervention scenarios. Thenature and magnitude of theinterventions are detailed inTable III.2. As noted already inchapter II, the scope andmagnitude of the assumedinterventions are only meant toillustrate the likely reduction inchild malnutrition underone possible scenario. It isobviously not possible topredict whether the assumedinterventions will indeed takeplace, and, even if they do,whether they will proceed
as the pace assumed inTable III.2.
Figure III.20 shows theprojected changes in the childunderweight rate in the poorstates when the seveninterventions shown in TableIII.2 are pursued simultane-ously. It is obvious that, whileeach of the interventionscontributes to the reduction inchild malnutrition, the onesthat are associated with thelargest declines in childmalnutrition are increased per-child public spending onnutritional programs, increasesin household consumptionexpenditure per capita, and expansion of adult female schooling. What isencouraging is that, together,the seven interventions areassociated with a reduction of25 percentage points in thechild underweight rate in thepoor states – enough for themto reach their MD goal of27.3% of children beingunderweight. This suggeststhat while attaining the childnutrition MDG will bechallenging in the poor states,it is clearly feasible with apackage of interventions thatinclude economic growth,increased public spendingon nutritional programs,improved physical infrastruct-
63
Child Malnutrition
58 As an alternative, the probit equation was estimated with a set of 23 state dummy variables, which replaced all the state-levelvariables (e.g., log of gross state domestic product per capita, log of state government nutrition per child, interaction betweenthe log of government nutrition expenditure per child and a dummy variable for the poor states, and log of health expenditureper capita). Although the full set of state dummy variables was significant at the 5% level, the explanatory power of theregression, as measured by a pseudo R-squared measure, actually declined from 0.172 to 0.170 with the substitution of thestate fixed effects for the four state-level variables. This suggests that the state fixed-effects model is not a superior model tothe one reported here in terms of goodness-of-fit.
Starting Assumed EndingIntervention value in change per value in
1998-99 year 2015
Female schooling (years) 2.7 0.3 7.8
Per child government expenditure onnutrition programs (Rs.) 51 4% 98
Consumption expenditure per capita (Rs.) 422 3% 698
Population coverage of regular electricity supply (%) 27.7 1% point 44.7
Population coverage of professionallyassisted deliveries (%) 32.3 1.5% points 57.8
Village access to pucca roads (%) 59.5 1% point 76.5
Population with no access to toilets (%) 76.5 -2% point 42.5
64
Attaining the Millennium Development Goals in India
Table III.1 : Projected decline in child underweight rates (percentage points) with variousinterventions in the poor and non-poor states
Bringing the poor states to the level of:
Non-poorNational states’ 50% 100%
Type of Intervention average average coverage coverage
No access to toilet (%) -0.5 -1.4 -2.0 -5.7(-0.6) (-4.3)
Access to regular electricity supply (%) -0.4 -0.8 -1.6 -5.3(-0.8) (-4.5)
Access to irregular -0.3 -1.1 -1.3 -3.1electricity supply (%) (-0.2) (-2.0)
Female schooling (years) -1.3 -3.9 -5.8 -8.0(-1.8)a (-4.1)b
Monthly consumptionexpenditure per capita (Rs.) -1.0 -2.9
Government expenditure onnutrition per child aged 0-6 years (Rs.) -2.9 -6.9
Percent of villages in district -1.0 -2.9 -6.2connected by a pucca road (%) (-3.3)
Medical attention at time -0.3 -1.1 -0.5 -2.1of child’s delivery (%) (-1.0)
Table III.2 : Assumptions about various interventions to reduce the childunderweight rate in the poor states, 1998-99 to 2015
Notes: Figures in parentheses refer to the underweight reduction obtained in the non-poor states.a The assumed coverage level in this case is 6.5 years of schooling. b The assumed coverage level in this case is 8 years of schooling.Empty cells indicate no substantial reduction in infant mortality due to the specific intervention being considered.
ure (electricity, sanitationaccess, medical attendanceat birth and rural roads), and expansions in femaleschooling.
The caveats noted inchapter II apply here as well.The fact that public spendingon nutritional programs isassociated with lower rates of child malnutrition doesnot mean that increasinggovernment nutritional expen-diture is sufficient. There isconsiderable evidence thatnutritional programs in India,such as the ICDS, NationalMid-Day Meal Program, andthe various micronutrientprograms, have poor coverage,targeting and implementation.The ICDS, for example,mostly focuses on childrenaged 3-6 years, but theconsensus among nutritionistsis that it is critical for direct
nutritional interventions toreach 6-24 month olds andpregnant women to preventmalnutrition (World Bank1998, 2001).59 Further, theICDS anganwadi center healthworker – one per center – istypically over-burdened, as shehas to manage pre-schooleducation, supplementaryfeeding, and outreachactivities. Another problemwith the ICDS program is thefrequent disruptions in foodsupplies that take place at theanganwadi center. Theresponsibility for the foodcomponent of the program lieswith the state governments,which typically under-financethis component owing to costand logistical difficulties. Oneevaluation of the ICDS found that disruptions in food distribution were verycommon at most anganwadicenters, with the average center
going without any food rationsfor 64 days per year (out of anintended 300 feeding days)(National Institute of Public Cooperation and ChildDevelopment 1992).
But it is also the case thatspending on direct nutritionalprograms is very low in India;it amounts to only 0.19percent of GNP. In contrast,neighboring Sri Lanka spendsabout one percent of GNP ondirect nutritional programs(World Bank 1998). A majorWorld Bank report onmalnutrition argues that, giventhe magnitude of themalnutrition crisis, Indiashould be prepared to spend aminimum of 0.5 per cent ofGNP on direct nutritionprograms – more than doublewhat it currently spends(World Bank 1998). It goeswithout saying that thescaling-up of direct nutritionalinterventions would have to gohand in hand with revampeddesign, greater devolution, andbetter implementation of suchinterventions. The highly-successful Tamil NaduIntegrated Nutrition Program(TINP), which has been inoperation for more thantwenty years (and preceded theICDS) in the state of TamilNadu, offers important lessonsfor the design of directnutritional interventions (seeBox III.1).
65
Child Malnutrition
-35
-30
-25
-20
-15
-10
-5
0
19
98
19
99
20
00
20
01
20
02
20
03
20
04
20
05
20
06
20
07
20
08
20
09
20
10
20
11
20
12
20
13
20
14
20
15
-35
-30
-25
-20
-15
-10
-5
0
Expanding medical attention at birthIncreasing real govt exp on nutrition per childExpanding access to sanitationReal income growthExpansion in regular electricity supplyExpansion of female schoolingIncreasing access to rural roads
Intervention
Decline needed to attain MDG
Figure III.20 : Projected decline in % of children 0-3 who are underweightin the poor states, 1998-2015, under different intervention scenarios
(graph shows cumulative effect of each additional intervention)
59 There is also some anecdotal evidence that the ICDS has done better (in terms of addressing child malnutrition) in states suchas Tamil Nadu and Rajasthan which have better targeted the program to younger age groups.
ICDS and ChildMalnutrition
Since the Integrated ChildDevelopment Scheme (ICDS)accounts for much of the publicspending on nutrition in India,it may be instructive to analyzethe impact of that scheme onchild nutritional outcomes.Unfortunately, such anevaluation is stymied by the lackof availability of relevant dataand by the fact that, by now, thescheme extends to virtuallyevery village in the country. As such, there are no ‘control’ villages that offer acounterfactual – viz., theprevalence of child malnutritionin the absence of the program.
However, at the time theNFHS-1 data were collected in 1992-93, ICDS did not fully cover all the villages in the country. Fortunately, the NFHS-1 village /communityquestionnaire obtained infor-mation on whether a sampledvillage had an ICDS anganwadicenter (AWC).60 By merging thehousehold information on childanthropometry and the villageinformation on the existence ofan AWC, it is possible to examinehow the location of an AWC in avillage is associated with childmalnutrition levels in that village.
The NFHS-1 data indicatethat 34.5% of villages in theNFHS sample had an AWC.
Overall, the data show thechild underweight rate (forchildren under 4) to besomewhat lower in the villageshaving an AWC than in villagenot having one (51% versus55%).61 However, upondisaggregating the numbers bysex, it is found that thepresence of an ICDS angawadicenter is associated with amuch larger reduction inmalnutrition for boys than forgirls (Figure III.21).
Since it is important tocontrol for other variables,such as household livingstandards and maternalschooling, in comparing thechild underweight rate acrossAWC and non-AWC villages,we have estimated amultivariate probit model of
child malnutrition with theNFHS-1 data, exactly alongthe lines of the modelestimated with the 1998-99NFHS-2 data (and shown inAnnex Table 2) but with theaddition of an explanatoryvariable indicating the presenceof an ICDS anganwadi centerin a child’s village of residence.The empirical results, shownin Annex Table 3, confirm thepattern observed in FigureIII.21; the presence of anICDS anganwadi center in avillage is associated with areduction of about 5% in thechild underweight rate, butthis association is observedonly for boys. There is nosignificant association of anICDS anganwadi center withthe prevalence of malnutritionamong girls aged 0-3. This
66
Attaining the Millennium Development Goals in India
60 This information was not collected in the NFHS-2 data, presumably because ICDS coverage (of villages) was near universalby the time of that survey.
61 Note that in the NFHS-1 (1992-93), anthropometric data were obtained for children under the age of 4 years, while the cut-off was 3 years in the case of the NFHS-2 (1998-99) data.
56
54
55
50
53
51
46
48
50
52
54
56
58
Males Females Both sexes
No anganwadi center in village Anganwadi center in village
Figure III.21 : Percent of children under 4 years who are underweight,by sex and presence of ICDS anganwadi center in village, 1992-93
67
Child Malnutrition
The Tamil Nadu Integrated Nutrition Project, which was started in 1980 by the state government of Tamil Nadu,was one of the first projects in the world to make large-scale use of growth monitoring of children aged 6-36months old as a means to target the neediest children and monitor their progress. TINP has been hailed as themost successful nutrition intervention program in India (and probably the world) (Berg 1987). It began as anarea-targeted (to the rural areas of six districts having the lowest caloric consumption in the state but subsequentlyextended to the entire state), age-targeted (concentrating exclusively on children 6-36 months of age, whoaccounted for 90 per cent of the pre-school deaths in the state, and pregnant and lactating women), and need-targeted program. The latter was achieved by monitoring the weights of all children 6-36 months old in theproject villages, and enrolling only those children whose weight gain over a certain period fell below standard ina 90-day supplementation program that included daily feeding and counseling of mothers. Since the childrenwere on the supplementation program only for the duration of time their weight gain was below standard, theproject was basically seen as a short-term intervention that sought to reduce long-term dependence ofbeneficiaries on public assistance. To this day, TINP relies heavily on local nutrition workers, working inconjunction with local women’s and girls’ groups. The groups are taught behavior-change strategies. They learnto promote birth weight recording, regular monthly weighing, and spot feeding, while participating incommunity assessment, analysis, and problem-solving.
TINP links the delivery of health and nutrition services. Children who do not respond to the nutritionsupplementation are provided health services, which include check-ups and referrals, treatment of diarrhea,deworming, and immunization. These services are also available to pregnant and lactating women. In addition, theprogram includes intensive counseling of mothers in nutrition and hygiene education.
Evaluation studies of the TINP have indicated significant effects of the program. Severe malnutrition fellsignificantly, by 44 percent between 1992 to 1997, although moderate malnutrition was still quite widespread.Beneficiary children were able to maintain their weight advantage for two years or longer after they completed theprogram, indicating long-term effects. Costs per beneficiary were lower than for less targeted nutrition programs.Indeed, one study estimated that the annual recurrent costs of the TINP were less than one-half of the ICDSprogram operating in Tamil Nadu, while it had an impact on severe malnutrition that was two times as much asthe ICDS. The cost difference between the TINP and the ICDS arises almost entirely from the fact that the ICDSis a mass feeding program, while the TINP is highly-selective supplementary feeding program.
Sources: Dapice (1987), Berg (1987), Chatterjee (1996).
Box III.1: The Tamil Nadu Integrated Nutrition Project
surprising finding could reflectthat parents tend to selectivelybring their boys, but not their girls, for supplementary
feeding at the center. Or itcould indicate that anganwadiworkers or helpers provide alarger allocation of food to
boys than girls. This is an issuethat merits further exploration.