Policy Research Working Paper 7794

Synergies in Child Nutrition

Interactions of Food Security, Health and Environment, and Child Care

Emmanuel Skoufias

Poverty and Equity Global Practice GroupAugust 2016

WPS7794P

ublic

Dis

clos

ure

Aut

horiz

edP

ublic

Dis

clos

ure

Aut

horiz

edP

ublic

Dis

clos

ure

Aut

horiz

edP

ublic

Dis

clos

ure

Aut

horiz

ed

Produced by the Research Support Team

Abstract

The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent.

Policy Research Working Paper 7794

This paper is a product of the Poverty and Equity Global Practice Group. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The author may be contacted at eskoufias@worldbank.org.

This paper examines the extent to which the three key under-lying determinants of nutrition—food security; adequate caregiving resources at the maternal, household, and com-munity levels; and access to health services and a safe and hygienic environment—on their own and interactively are correlated with nutrition outcomes, such as height-for-age z-scores. Based on data from different years in eight coun-tries in four regions where malnutrition is high, an indicator

is constructed for each component of the three underlying drivers of nutrition. In spite of the limitations inherent in the available data, the analysis (i) reveals that progress toward improved access to adequate food security and adequate environment and health has been quite limited; and (ii) provides evidence of significant synergies among adequate food, child care, and environment and health.

SynergiesinChildNutrition:InteractionsofFoodSecurity,Healthand

Environment,andChildCare

EmmanuelSkoufias

Poverty&EquityGlobalPractice

TheWorldBankGroup

JELClassification:I14,I15,I18,I38,J13

Keywords:ChildNutrition,Stunting,FoodSecurity,HealthServices,Environment,ChildCare.

Contents

Acknowledgments...................................................................................................................................................................3 

1.Introduction.........................................................................................................................................................................1 

1.1  MethodologicalFramework.............................................................................................................................4 

2  DataandMeasures........................................................................................................................................................7 

2.1  MeasuresofFoodSecurity...............................................................................................................................8 

2.1.1  “Ideal”IndicatorsofFoodSecurity.....................................................................................................8 

2.1.2  AvailableIndicatorsforFood..............................................................................................................12 

2.2  MeasurementofChildCare............................................................................................................................13 

2.2.1  IdealIndicatorsforChildCare............................................................................................................13 

2.2.2  AvailableIndicatorsforchildcare.....................................................................................................14 

2.3  MeasurementofEnvironmentandHealth..............................................................................................15 

2.3.1  IdealIndicatorsforenvironmentandhealth................................................................................15 

2.4  Otherdataissuesandrelatedconsiderations........................................................................................16 

3  Prevalenceofstuntingandaccesstoadequatefoodsecurity,childcare,andhealthandenvironment............................................................................................................................................................................19 

3.1  Prevalenceofstunting......................................................................................................................................19 

3.2  AdequaciesinFoodSecurity,Environment,Health,andCarePractices.....................................21 

3.2.1  Adequatefoodsecurity..........................................................................................................................22 

3.2.2  AdequateCarePractices........................................................................................................................25 

3.2.3  AdequateEnvironmentandHealth...................................................................................................28 

3.3  AdditionalConsiderations:WeakLinksinaChain..............................................................................33 

4  Synergies.........................................................................................................................................................................38 

4.1  Model.......................................................................................................................................................................38 

4.2  Results.....................................................................................................................................................................40 

5  ConcludingRemarksandPolicyConsiderations............................................................................................44 

References................................................................................................................................................................................47 

Annex1:Assetindex............................................................................................................................................................51 

Annex2:Componentsofadequacymeasuresbywealth......................................................................................52 

Annex3:Synergiesbywealthgroup.............................................................................................................................56 

Acknowledgments

TheteamforthisstudywasledbyEmmanuelSkoufias(LeadEconomist,GPVDR)andwascomprisedofJoseAntonioCuesta,(SeniorEconomist,GPVDR),NkosinathiVusizihloboMbuya(SeniorNutritionSpecialist,GHNDR),SaileshTiwari(SeniorEconomist,GPVDR),EkoSetyoPambudiandIndiraMaulaniHapsarifromtheJakartaWorldBankOffice,KatjaVinha(Consultant),LauraMaratou‐Kolias(Consultant,GWASP),SushenjitBandyopadhyay(Consultant,GENDR),ChristineJ.Foreman(Consultant),andLeonardoTornarolli(Consultant).

ThepeerreviewerswereLeslieElder(SeniorNutritionSpecialist,GHNDR),PatrickHoang‐VuEozenou(Economist,GHNDR),CraigKullmann(SeniorWaterSupplyandSanitationSpecialist,GWASP),JohnNewman(LeadStatistician,GPVDR),andHaroldAlderman(SeniorResearchFellowatIFPRI).TheteamisespeciallythankfultotheNationalInstituteofHealthResearchandDevelopment(NIHRD)oftheMinistryofHealthofIndonesiaformakingavailabletherelevantdatafromthe2010RISKESDASsurvey.AdditionalcommentsandsuggestionswerereceivedfromMeeraShekar(LeadHealthSpecialist,GHNDR),ClaireChase(EconomistGWASP)andteam(DeanSpears,,andMariaQuattri),RajaB.Kattan(SectorLeader,AFTHD,onbehalfofVeraSongweandAFCF1HDteam),MartinRama(ChiefEconomist,SARCE),andGladysLopez‐Acevedo(LeadEconomist,SARCE).TheteamhasalsobenefittedfromcommentsreceivedfromparticipantsinseminarsattheWorldBank,IFPRI,andattheUNICEFHeadquartersinNewYork.

1

1.Introduction

TherecentassessmentsoftheglobalprogresstowardstheachievementoftheMDGsshowthatprogressinnutritionhasbeenslowerthanexpected(WorldBank,2013).Malnutritionrateshaveremainedsurprisinglyhighinseveralcountrieswithrobusteconomicandagriculturalgrowth,whichsuggeststhatincreasesinrealGDPandincomeareinsufficientforreductionsinchildmalnutrition.

Whilethereiswidespreadagreementthatreducingchildhoodmalnutritioniscriticallyimportantfordevelopment,thereislessconsensusonhowtoachieveimprovementsinnutrition.Alargebodyofresearchinthefieldofpublichealthandnutritionhasconcentratedonevaluatingtheimpactsofspecificinterventionssuchasfood,mineral,andvitaminsupplementsortrainingprograms.Ontheotherhand,theCopenhagenConsensusrecommendationssuggestthatinterventionsoutsidethenutritionsectorcouldbemorecosteffective(Deolalikar,2008;Hoddinott,etal,2012).However,uptonow,thereislittleempiricalevidenceonthefactorsorcombinationofsector‐specificinterventionsthatareassociatedwithmeasurableimpactsinnutrition.

Recently,anumberofinitiativeshavesurfacedattheinternationalfrontaimingtoscaleupnutritioninterventions.OneprominentexampleistheScalingUpNutrition(SUN)movement,whoseframeworkisendorsedby30developingcountries(Horton,etal.2010).Leadersofthesecountriesareprioritizingnutritionasaninvestmentintheirpeople’sgrowth,andrecognizingnutritionasaninvestmentineconomicandsocialdevelopmenttostrengthentheirnations.Alongparallellines,initiativeswithintheWorldBankandotherdevelopmentagenciesandresearchinstitutions,aimtofosterknowledgeexchangeandcross‐sectoralcollaborationandcoordinationattheprojectlevelforimprovingnutrition.Alltheseinitiativesarebasedonthepremisethatthedeterminantsofmalnutritionaremulti‐sectoralandthatthesolutiontomalnutritionrequiresmulti‐sectoralapproaches.

TheUNICEFframework,firstproposedin1990(UNICEF,1990),wasoneofthefirstattemptsatemphasizingfoodsecurity,environmentandhealth,andchildcarepracticesasunderlyingdeterminantsofchildmalnutritionindevelopingcountries.Oneofthefundamentalideasunderpinningthisframeworkisthattherearesubstantialinteractionsandsynergiesamongfoodsecurity,environment,health,andcare.Thisconceptualframeworkhasbeenguidingoperationalandappliedanalyticalworkformorethan20yearsnow.1

Theempiricalevidencethatexiststodateprovidesaverypartialpictureregardingthedirectionandmagnitudeoftheinterdependenceamongadequate(orinadequate)accesstofoodsecurity,environment,health,andchildcareinchildnutrition.OnefundamentalpremiseoftheUNICEFconceptualframework,isthatincreasesinaccesstoadequateservicesinoneorallofthe

1Attheempiricallevel,the"guidance"providedbytheframeworkhasbeenprimarilyintermsofminimizingomittedvariablebiasor,putdifferently,intermsofsuggestingimportantvariablestobeusedasexplanatoryvariablesinreducedformregressionsattheindividual,householdand/orcommunitylevel(Alderman,HoogeveenandRossi,2005;BehrmanandDeolalikar,1998,ChristiaensenandAlderman,2004;Haddad,Alderman,Appleton,SongandYohannes,2003;ElfindriandGourangaLalDasvarma,1996;Rajaram,ZottarelliandSunil,2007;SahnandAlderman,1997;Skoufias,1999;StraussandThomas,1998).

2

subcomponentsofanyoneoftheseclusters,sayforexample,foodsecurityalone,cannotsubstituteforinadequatelevelsofaccesstotheotherclustersorindividualcomponentsoftheseclusters.

Foranacceleratedprogresstowardsreducingchildmalnutritionwhatisrequiredisamorecompleteanddetailedknowledgeoftheextenttowhichtherearegapsinaccesstoadequatelevelsineachoftheclustersofunderlyingdeterminants.Incontextswherechildmalnutritionisprevalent,adetailedprofileofcommunitiesordistrictsorevencountriesintermsofwhetherthereisadequateorinadequateaccesstoeachofthesubcomponentsoffoodsecurity,environment,healthandchildcare,canprovideasolidbasisfortheidentificationofpotentialbindingconstraintsimpedingprogresstowardsreducingchildmalnutrition.Thisexercisemayalsofacilitatetheprioritizationamonginterventionsintheeventthatlimitedfinancialresourcesorcapacityontheground,donotpermitsimultaneousimprovementsinallofthecriticalfactorsofchildmalnutrition.

Inspiteofthefundamentalroleofinteractionsamongthethreepillarsof(i)foodsecurity,(ii)environmentandhealth,and(iii)childcare,thereisapaucityofevidenceontheextenttowhichthereisadequateorinadequateaccesstooneormoreofthesethreepillars.2Toalargeextentthelacunaofsuchevidence,maybeattributedtotherelativescarcityofnationallyrepresentativedatasetswithallthenecessarydetailedinformation(inthesamesurvey)onchildnutrition,andallthevariablesthatcouldcaptureinasatisfactorymannerthedifferentdimensionsoffoodsecurity,environmentandhealthandcarepractices.Untilnow,therehasbeennosystematiceffortatdocumentingthegapsinthenecessarydata.ThemostcommonlyusedsourcesofdatasuchastheDemographicandHealthSurveys(DHS)ortheMultipleIndicatorClusterSurvey(MICS),forhistoricorbudgetaryreasonscollectdataforsomekeyvariablessuchaschildcareandenvironmentandhealthwithoutmuchusefulinformationcollectedonfoodsecurity.Thesesurveysarecharacterizedbyastrongpathdependence,inthesensethattheoriginalsurveyscollectedinformationaboutspecificcomponentsofoneortwoofthethreepillarsmentionedabovewiththeadditionalquestionsaddedovertimeconstrainedbythe“straightjacket”ofmaintainingcompatibilitywiththeearliersurveys.Incontrast,otherspecializedsurveysendedupcollectinginformationondifferentdimensionsoffoodsecuritybutnoinformationatallchildnutritionandanthropometricmeasures,oronchildcareorhealthandenvironment.3

Withtheseconsiderationsinmind,thisreportcontributestotheexistingliteratureinthreeways.First,itprovidesoneofthefirstcomprehensiveinvestigationsofthedataavailabilityanddataconstraintsassociatedwithamoresystematicapplicationoftheUNICEFconceptualframeworkemphasizingtheinterrelationshipamongaccesstoadequatefoodsecurity,environmentandhealth,andchildcarepracticesintheprevalenceofmalnutritionratesamongchildrenindevelopingcountries.UsingdetaileddemographicandhealthsurveydataareusedfromBangladesh,Bolivia,

2NotableexceptionsareChongetal.(2003)andAldermanetal(2003).3Interestinglytheempiricalliteratureonthedeterminantsofchildhealthdoesnotappeartobeaffectedmuchbythescarcityofrelevantdataonthedeterminantsofchildmalnutrition.Thepracticeprevalentinthemajorityofstudiesistofocustheanalysisonthecontributionofafewkeyvariablesonchildnutritionaloutcomes.Examples,ofsuchvariablesincludetheeconomicstatusofthehousehold,measuredbyincome,consumptionexpenditures,orhouseholdassets(BehrmanandDeolalikar,1987),andthelevelofeducationofthemotherofthechild(Barrera,1990;BehrmanandWolfe,1984;Skoufias,1999;WebbandBlock,2004).Inmanystudiesvillage‐level(orsometimeshousehold‐level)fixedeffectsareusedtocontrolforthepotentialinfluenceofunobservedorunmeasuredfactorssuchastheenvironmentalhealthandthecarepractices.Unfortunately,fromtheperspectiveoflearningabouttheinteractionsamongthesefactorsthisistantamountto“throwingoutthebabywiththebathwater"(BellandJones,2012;BeckandKatz,1995,2001).

3

Cambodia,Ethiopia,Indonesia,NepalandZimbabwe,acomparisoniscarriedoutoftheidealvariablessummarizingthevariousdimensionsorcomponentsoffoodsecurity,environmentandhealth,andchildcareagainstthemeasuresavailablefromcurrenthouseholdsurveys.Thesecomparisonsservetohighlightthelimitationsofmostdatasetsandthepotentialgainsassociatedwiththecollectionandavailabilityofadditionalinformation.

Second,bearinginmindthelimitationsimposedbythedataavailable,thereportalsoprovidesapracticaldiagnosticframeworkofthemaincorrelatesofchildmalnutritionthatcouldbeappliedtoidentifypotential“bindingconstraints”towardstheefforttoreducechildmalnutrition.Specifically,theUNICEFconceptualframeworkis“operationalized”byservingasaguideforinvestigatingtherelationshipbetweentheprevalenceofmalnutritioninthecountryandinadequatelevelsandaccesstothethreepillarssummarizingtheunderlyingcausesofmalnutrition.Next,foreachindicatoravailablefromthesurveyusedincountry,adefinitionof“adequacy”isconstructedusingthresholdsbasedonacceptedinternationalstandards.Thereportaimstoprovidea“helicopterview”oftheextenttowhichnutritionaloutcomes,asmeasuredbyachild’sheight‐for‐ageZscore(HAZ)atanygivenpointintime,aswellasovertime,areassociatedwithinadequatefoodsecurity,inadequateenvironmentandhealth,andinadequatechildcarepractices.Inconsiderationofthecomplexityofthelinkagesbetweentheunderlyingcausesofmalnutritionandtheeconomicsituationofthefamily,theanalysisisalsocarriedoutseparatelyforresource‐rich(top60%)andresource‐poor(bottom40%)householdsbasedonanassetindexconstructedforthatpurpose.

Third,thereportprovidessomeofthefirstempiricalevidenceonthesynergiesatworkincombatingmalnutritioninthesetofcountriesusedinthisstudywhenthereissimultaneousaccesstoadequatelevelsintwoormoreofthethreepillarsoftheunderlyingcausesofmalnutrition.Whileintuitivelyappealing,thesynergiesamongthethreeclustersofmalnutritionhavereceivedlittleempiricalvalidation.Therecentemphasisonsector‐specificnutritionsensitiveinterventions(WorldBank,2014)rightlyemphasizesthesynergiesthatcanbeexploitedwithinspecificsectorssuchagriculture,waterandsanitation,orsocialprotection.Theanalysisinthereportunderscoresthepointthatthesuccessofthesesector‐specificnutrition‐sensitiveinitiativesmaybeconstrainedbytheslowprogressintakingadvantageofthesynergiesamongthethreebroadclustersoftheunderlyingdeterminantsofmalnutrition:foodsecurity,childcare,environmentandhealth.Giventhatthesynergiesamongthesethreepillarsarebeyondthescopeofanyspecificsectorsuchasagricultureorsocialprotection,thesimultaneousprogresstowardsinallofthethreepillarsiseithertakenforgrantedorunderemphasized.Asaconsequence,nutritionsensitiveinterventionsinspecificsectorsendupbeingasectoralpriorityincontextswherethechancesofsuccessmaybelimitedbecauseofnoorverylowaccesstoimprovedinfrastructurewaterandsanitationfacilities.

4

1.1 MethodologicalFramework

TheoriginalUNICEFconceptualframeworksummarizedinFigure1aviewsmalnutritionastheconsequenceofavarietyofinterlinkedandinterrelatedevents.Thecausesofmalnutritionareclassifiedintothreehierarchicalcategories:theimmediatecauses,theunderlyingcauses,andthebasiccausesofmalnutrition.Inanygivencontextidentificationoftheimmediatecausesofmalnutrition(diseaseorinadequatedietaryintake)isusefulforguidingpolicyactionsespeciallyinsituationsofcrises.However,diseaseorinadequatedietaryintakearetypicallyconsequencesofavarietyofunderlyingfactorsthatareinterrelated.Forconceptualsimplicitytheunderlyingcausesofmalnutritionarethemselvesgroupedintothethreeclusters:inadequatehouseholdfoodsecurity,inadequatecareandfeedingpractices,andunhealthyhouseholdenvironmentandinadequatehealthservices.Thebasiccausesofmalnutritionsummarizethesocial,cultural,economicandpoliticalcontextandtheprevailinginequalitiesinthedistributionofresourcesinthesociety.Incombinationthesecontextualorstructuralfactorsplayafundamentalroleintheextenttowhichthereareinequalitiesamonghouseholdsandtheirmembersinhavingadequatefoodsecurity,careandfeedingpractices,healthyenvironmentandadequatehealthservices(i.e.,theunderlyingcausesofmalnutrition).

Figure1a:DeterminantsofChildNutrition

Source:AnadaptationoftheUNICEF(1990)“StrategyforImprovedNutritionofChildrenandWomeninDevelopingCountries”

5

Sinceitsconceptionthisconceptualframeworkhasbeenrevisedandextendedinvariousdimensions.Variousinternationalorganizationshaveadoptedaswellasadaptedthisframework.Forexample,FAO(2011)discussesadaptationofthisframeworkforFAO’snutritionanalysis.USAID‐FANTA(FoodandNutritionTechnicalAssistance)alsoadaptedthisframework(Rielyetal.,1999).WorldFoodProgram(WFP)referstoitastheFoodandNutritionSecurityConceptualFrameworkinitsEmergencyFoodSecurityAssessmentHandbook(WFP,2009,pg.25).However,whatevertheadaptationsandtheextensionstotheoriginalframework,thefundamentalideasregardingthecriticalinteractions,interrelationsandsynergiesamongfoodsecurity,environmentandhealth,andcarehaveremainedatthecore.Thisisalsoverytransparentintheframeworkforactionstoachieveoptimumfetalandchildnutritionanddevelopmentextractedfromthe2013LancetMaternalandChildNutritionSeries(seeFigure1bbelow).

Figure1b:FrameworkforActionstoAchieveOptimumFetalandChildNutritionandDevelopment

Source:theExecutiveSummaryof“TheLancetMaternalandChildNutritionSeries2013.”

Theanalysisinthisreportfocusesontheunderlyingcausesofmalnutrition.Thisisbecauseactionsaimedataffectingtheunderlyingcausesofmalnutritionarelikelytobemorefeasibleinthemediumtermandthusmorelikelytohavealongtermeffectonmalnutrition.4Foodsecurity

4Oneshouldnotunderestimatethepotentialofreductionsininequalitiesinthedistributionofresourcesinthesociety,orchangesinthesocial,cultural,economicandpoliticalcontextinhavinglargeandlastingpositiveeffectonchildnutrition.However,changesinthebasiccausesofmalnutritionarelikelytobeinhibitedbymanymoreconstraintsincludingvestedinterestsandpoliticaleconomyconsiderations.

6

summarizestheavailability,access,andutilization(consumption)offood.5Environmentandhealthsummarizesthevarietyofcontextualfactorsthatcanimpactonchildren’ssusceptibilitytodiseasessuchaslackofimprovedsanitationandwaterfacilities,difficultyofaccesstohealthfacilities,orlowqualityofhealthcare.Finally,maternalandchildcaresummarizesthequalityofcareprovidedbythecaregiver,suchasfeedingandhygienepracticesadoptedandbytheavailabilityofthecaregiver.Furthermoreitmeasureshowwellthecaregiverissupportedinherchildrearingendeavors.

Althoughtheframeworkisaholisticwayofconceptualizingnutritionitisalsoimportanttoacknowledgethelimitationsoftheclassificationscheme.Prices,knowledge,education,andhouseholdincomeallinfluencecomponentsofthethreeclustersoftheframework,resultinginsomeoverlapinthemeasures.Themethodologyisinformativeinfindingtheoverallrelationships,fromwhichmorefocusedanddetailedanalysescanbecarriedouttodeterminemoreconcretelytheunderlyingcauses.Soforexample,moredetailedinformationwouldbeneededtodeterminewhetherfoodinadequacieswereduetothecostoffoodrelativetoincome,tolackofinformationontheimportanceofdiversifieddiet,orduetosomeotherfactor.Themodelsestimatedinthisreportarenotreducedformmodels(takingintoaccountbudgetconstraintsetc.)asdoneinBarrera(1990),butrathercorrelationsbetweennutritionaloutcomessuchaheightforagez‐scoresandhavingadequatelevelsofaccesstoeachofthevariablesgroupedintothethreeclusters.

5Itisimportanttobearinmindthatouruseoftheterm“utilization”representsanadaptationtoFAO’suseofthesametermintheirdefinitionoffoodsecurity.ForFAO,utilizationisthefollowing: “Utilizationoffoodthroughadequatediet,cleanwater,sanitationandhealthcaretoreachastateofnutritionalwell‐beingwhereallphysiologicalneedsaremet.”Thisbringsouttheimportanceofnon‐foodinputsinfoodsecurity.http://www.fao.org/forestry/13128‐0e6f36f27e0091055bec28ebe830f46b3.pdf

7

2 DataandMeasures

Measurementisveryimportantforaproperdiagnosisofthelongerrunconstraintstoreducingchronicmalnutritioninanygivencontext.Eachofthethreeclustersofunderlyingcausesofmalnutritionisinherentlymultidimensionalmakingmeasurementdifficultandcostly.AsaconsequencethemainconceptsunderpinningtheoriginalUNICEFframeworkregardingtheinterrelationshipbetweenandsynergiesamongfoodsecurity,environmentandhealth,andchildcareareusuallytakenforgrantedormistakenlyassumedtohavebeeninvestigatedbyotherearlierstudies.

Table1liststhecountriesandsourcesofdatausedinthisstudy.Twocountrieswerepurposefullychosenfromeachofthefourregionswheremalnutritionisaproblem,(i.e.SouthAsia(SAR),LatinAmericaandtheCaribbean(LAC),EastAsiaPacific(EAP),andSub‐SaharanAfrica(SSA).Additionalcriteriaappliedincluded:(i)thesurveycontainedreliableinformationofchildren’sheight(andweight)whicharethewidelyacceptedmeasuresofchronicandshort‐termmalnutrition.ThiscriterionlimitedtheanalysistotheDemographicandHealthSurveys(DHS)asthesearetheonlysurveysismostcountrieswithchildheight(andweight)measures.(ii)Datawereavailableforatleasttwoyearsinthelastdecade;(iii)malnutritionrateswerestableorslowlydecliningovertime;(iv)therewasparallelanalyticworkintheBankondifferentdimensionsofpovertyandnutritiontakingplaceinsomeofthesecountries(e.g.Ethiopia);and(v)thereweremorethanonesurveyavailableinthesamecountry(e.g.,Bangladesh).ThesamplesinmostcountriesaremainlyruralexceptforBolivia,Peru,Indonesia,andtheHelenKellersurveyinBangladeshwhereurbanhouseholdsmakeupbetweenone‐thirdandone‐halfofthesamples.TheIFPRIsurveyforBangladeshincludesonlyruralhouseholds.

Table1:CountriesandDataSources

Region Country‐datasource

SARBangladesh(HelenKeller2010,2011andIFPRI2011)Nepal(DHS2001and2011)

LACBolivia(DHS2003and2008)Peru(DHS2005and2012)

EAPCambodia(DHS2005and2010)Indonesia(Riskesdas2010)

SSAEthiopia(DHS2000and2011)Zimbabwe(DHS2005and2010)

Giventhesecriteriathesampleofchildrenusedisbetween0and24monthsinBangladesh(HelenKellerdata),Cambodia,andZimbabwe,between0and25monthsinBangladesh(IFPRIdata)andIndonesia,andbetween0and36monthsinBolivia,Ethiopia,Nepal,andPeru.6

6Inspiteofrecentfindingsthatcatch‐upgrowthoccurswithoutinterventions(Prentice,etal.,2013),orasaresultofinterventions,muchofstuntingoccursbeforetheageof24months(Victora,etal.,2010).Inadditionrecentresearchhasfoundthatcatchupgrowthinschoolagedchildrenisnotassociatedwithimprovementsincognitiveability(Sokolovic,Selvam,Srinivasan,Thankachan,Kurpadand

8

Thesetofcountriesusedintheanalysisisdiscussedandthenutritionliteratureissurveyedforthepurposeofidentifyingascompletelyaspossiblethesetoffactorsidentifiedinthenutritionliteratureastheidealvariablesormeasuresofunderlyingcausesofmalnutrition.

2.1 MeasuresofFoodSecurity

Thissectionservestohighlightthefactthatthenecessarydataformeasuringthedifferentdimensionsoftheconceptoffoodsecurityeitheratthehouseholdorattheindividuallevelarethedatamissingmostfrequentlyfromthestandardsurveysusedtoassessthelevelsanddeterminantsofmalnutrition.Incontrast,moredataseemtobeavailableforthemeasurementofsomeofthecomponentsofchildcareandenvironmentandhealthwiththelatterbeingmeasuredalmostasbestasonecouldhopefor.Thisisprobablyareflectionofthetradeoffsassociatedwithmeasurementandcost.Giventhelimitedsurveybudgetsallocatedtocollectinginformationonthedifferentdimensionsofnutrition,greateremphasismaybeplacedoncollectinginformationonenvironmentandhealthandchildcareasthesearecollectedatlowercostandperhapsevenmorereliablycomparedtothecostofadetailedsurveycollectinginformationonfoodavailability,accessandutilizationatthehouseholdorindividuallevel.

Asystematiceffortismadetoidentifytheextenttowhichanyorsomeoftheseidealvariables/measuresidentifiedintheliteratureasanimportantunderlyingcauseofmalnutritioncanbeapproximatedbyanyproxymeasurethatiscollectedintheexistingsurveysusedtoanalyzeormonitorchildmalnutritionatthenationalorevenattheregionallevelwithinanygivencountry.

2.1.1 “Ideal”IndicatorsofFoodSecurity

TheFoodandAgriculturalOrganization(FAO)definesfoodsecurityas“asituationthatexistswhenallpeople,atalltimes,havephysical,socialandeconomicaccesstosufficient,safeandnutritiousfoodthatmeetstheirdietaryneedsandfoodpreferencesforanactiveandhealthylife.”Thisdefinitionisasignificantdeparturefrompreviousconceptualizationsoffoodsecuritywhichfocusedinordinatelyontheavailabilityoffoodatthenationalorlocallevel.But,inbeingbroadandallencompassing,thisdefinitionisalsoadifficultonetooperationalize,asitemphasizestheimportanceofaccessandutilizationoffoodjustasmuchasavailability(Barrett,2009).

Whatconstitutesavailability,accessandutilization–thethreedimensionsofthecurrentthinkingonfoodsecurity?Availabilityisassociatedwiththesupplysideoffood,measuredmostoftenbytheextentofagriculturalproductionandfoodtradebalancerelativetothesizeofconsumptionforanygivencountry.Access,ontheotherhand,bringsinthedemandelementtotheequation:conditionalonwhatisavailableinthelocalmarketandthepriceatwhichitisavailable,whatistherangeoffoodchoicesthatareopentohouseholdsgiventheirincomes?Conceptually,itisthisdimensionoffoodsecuritythathasthestrongestresonancewithpovertyandvulnerabilitynotonlybecauseof

Thomas,2013)althoughcatch‐upgrowthinearlierchildhoodwasassociatedwithcognitiveabilitiessimilartothosewhohadneverbeenstunted(Crookston,Penny,Alder,Dickerson,Merrill,Stanford,PorucznikandDearden,2010).

9

itsdirectrelationshipwithincome,butalsobecauseofitslinkstobroaderissuesofsocialandpoliticalenfranchisement.Foodsecurityofindividualhouseholdmembers,forexamplehingesontheirsocialstandingwithinthehouseholdalmostasmuchasitdoesonthehousehold’soverallabilitytoprocureenoughfood(vulnerablegroupswithinthehouseholdmayincludechildren,daughters,daughters‐in‐law,ortheelderly).Finally,theutilizationdimensionbringstobearthequalitydimensionoftheaccessedfood.Dohousehold’smakegooduseofthefoodtheyareabletoaccess?Aredietsdiverseenoughtoprovideallthemicroandmacronutrientsnecessaryforhealthyphysiologicalandcognitivegrowth?Arecookingmethodssanitaryandhealthyenoughtopreservethenutritionalattributesoftheeatenfood?

FollowingtheFAO’sacceptedandapplieddefinitionoffoodsecurityaneffortismadetomapthemostcommonlyusedmeasuresintoatleastoneofthethreedimensionsoffoodsecurity.Specifically,themeasuresconsideredare:percapitaexpenditure;shareoffoodintotalexpenditure;percapitacaloricavailability;foodconsumptionscore(FCS);householddietarydiversityscore(HDDS);childdietarydiversityscore(CDDS);mother’sdietarydiversityscore(MDDS);householdfoodinsecurityaccessscale(HFIAS);starchystapleratio(SSR);andshareoffoodexpenditureonstarchystaples(SSEXR).

Percapitaexpenditureisawidelyusedmeasureofahousehold’swealthstatusandoverallwell‐being.Itisindicativeofresourcesthatareavailabletoahouseholdthatthehouseholdcantapintotosatisfytheirfoodrequirements.Itisthususedasoneofthemeasuresofaccesscomponentoffoodsecurity.Foodshareoftotalexpenditureisanindicatorofthehousehold’seconomicvulnerabilityandcanbeaproxymeasureofhousehold’sabilitytoaccessfood.Householdsthatspendalargerproportionoftheirtotalexpenditureonfooddonothavesufficientsafetynetofnon‐foodexpendituretorelyonandthusaremoresusceptibletofooddeprivation.Inaneventofanegativeincomeshockorincreaseinfoodprices,householdswithahighershareoffoodexpenditurewillhavetoadjustbyeitherreducingfoodquantityorbyloweringthequalityoffoodtheyeat.Percapitacaloricavailabilityisanindicatorofdietquantityandrelatestotheaccesscomponentoffoodsecurity.Itisoneofthemostwidelyusedquantitativeindicatorsoffoodsecurity.Itmeasureswhetherahouseholdhasacquiredsufficientcaloriestomeetthedailyenergyrequirementsofitsmembers.Ifahousehold’sestimatedpercapitadailyenergyavailabilityislowerthanthepercapitadailyrequirement,thenthehouseholdisconsideredenergydeficientandcanbeclassifiedasfoodinsecure.

Anothermeasureoffoodsecurityusedistheshareofcaloriesderivedfromstarchystaples–orstarchystaplesratio(SSR).Itismeasuredasthepercentageofcaloriesderivedfromstarchystaples.Starchystaplesareenergy‐densebutarelowinproteinandmicronutrientswhichmeansthathouseholdswithhighervalueofSSRwillhavealowerqualitydietandwillbemorevulnerabletoproteinandmicronutrientdeficiencies.Moreover,starchystaplesarenotonlycheapersourcesofenergybutalsofigureprominentlyasapartofhousehold’sstaplediet.JensenandMiller(2010)suggestthismeasuretobeapotentiallypromisingwaytocapturefoodsecuritywithinthehousehold.Itreliesonconsumptionbehavior,torevealthehouseholdfoodsecuritysituation,as

10

opposedtocaloricnorms.7Finally,thereisalsotheexpenditureanalogofthestarchystaplesratiowhichistheshareoffoodexpendituresthatisdevotedtothepurchaseofstarchystaples.Wecallthisthestarchystaplesexpenditureratio(SSEXR).

Dietarydiversityisameasureofdietqualityandreflectsthevariationinfoodtypicallyconsumedbyhouseholds.Ingeneral,itisdefinedasasumofthenumberoffooditemsorfoodgroupsconsumedoveragivenreferenceperiod.Althoughthereisnoageneralconsensusinconstructingameasureofdietarydiversity,studieshaveshownthatvariousmeasuresofdietarydiversityarepositivelycorrelatedwithothersmeasuresofhouseholdfoodsecuritysuchaspercapitaconsumption,calorieavailability,calorieintake,andintakeofessentialnutrients.8

Twoofthemostcommonlyusedindicatorsofdietarydiversityarehouseholddietarydiversityscore(HDDS)andindividualdietarydiversityscore(IDDS),developedbyUSAIDFoodandNutritionalTechnicalAssistance(FANTA).HDDSisdefinedasthenumberofdifferentfoodgroupsconsumedatthehouseholdlevelbyanaveragememberovera24‐hourrecallperiod.Whereas,IDDSisdefinedasthenumberofdifferentfoodgroupsconsumedbyanindividualovera24‐hourrecallperiod.FANTA/FAOusestwelve,eightandninefoodgroupclassificationtoconstructtheHDDS,CDDSandMDDS,respectively.ThevalueofHDDSrangesfrom0to12andthevaluesofCDDSandMDDSrangefrom0to8and0to9,respectively.

Foodconsumptionscore(FCS)isameasureofaccesscomponentoffoodsecuritydevelopedbytheWFP.WFPusesFCSasoneofthecoremeasuresoffoodconsumptionandfoodsecuritytomonitor,assess,andtrackchangesinfoodsecuritysituationandneedsofcountriesandregionsthatithasprogramsin.Itisacompositescorethatincorporatesdietarydiversity,foodfrequency,andrelativenutritionalimportanceofdifferentfoodgroupsconsumedbyahousehold.ForthecalculationoftheFCS,dataiscollectedonthe7‐dayrecalloffrequencyofconsumptionofdifferentfooditemsandfooditemsaregroupedinto8specificfoodgroupswitheachgroupgivenaweightrepresentingthenutrientdensityofthatfoodgroup.ThevalueofFCSrangesfrom0to112withahigherFCSrepresentingahigherdietarydiversityand/orfrequencyofconsumptionandhighernutritionalvalueofahousehold’sdietandviceversa.

Householdfoodinsecurityaccessscale(HFIAS)isameasuredevelopedbyFANTAtoassessfoodaccessproblemsfacedbyhouseholdduringarecallperiodof30days.Itaimstocapturethechangesinfoodconsumptionpatternsandreflecttheseverityoffoodinsecurityfacedbyhouseholdsduetolackoforlimitedresourcestoaccessfood.Itiscomposedofninequestionsandthesequestionsrelatetothreedifferentdomainsofaccesscomponentfoodinsecurity:anxietyanduncertaintyabouthouseholdfoodaccess,insufficientquality,andinsufficientfoodintake(Swindaleetal.2006).Eachquestionhasfourresponseoptions:never,rarely,sometimes,andoften,which

7Itisbasedontheideathatatlevelsbelowsubsistence,individualshavehighmarginalutilitiesforcaloriesandarelikelytochoosecheapsourcesofcaloriessuchasrice,wheat,cassavaetc.Astheypasssubsistence,theirmarginalutilityofcaloriesbeginstodeclineandtheybegintovalueothernon‐nutritionalattributesoffoodsuchastasteandstartdiversifyingtheirdiet.Whiletheactualsubsistencethresholdisunobserved,their“dietarytransition”isandthiscanbeusedtoidentifywhetherornottheyhavecrossedthefoodsecuritythreshold.Byrelyingdirectlyonconsumptionbehaviortoelicitinformationonhungerandfoodsecurity,thismethodobviatestheneedtoimposecaloricnormsandthresholds.8Ruel(2002),Weismannetal(2009)andHoddinottandYohannes(2002).

11

arecoded0,1,2and3inorderofincreasingfrequency.Responsestotheseninequestionsaresummedtoconstructafoodinsecurityscore,withamaximumscoreof27indicatingmostfoodinsecurehouseholds.

Table2belowpresentsasummarydescriptionofallthese“ideal”indicatorsofthedifferentdimensionsoffoodsecurity.

Table2:IdealFoodSecurityMeasures

DimensionsofFoodSecurity

FoodSecurityIndicator Description

Availability

Percapitadailycalorieavailability Indicateswhetherenoughfoodisavailabletomeetthedailyenergyrequirementsofanationalorlocalpopulation.Mostcommonlyusedmeasureoffoodsecurity.

Percapitahouseholdexpenditures Capturestheamountoffoodpurchasedoracquiredduringasurveyperiod.Associatedwithgeneralmeasuresofpoverty.

Shareoffoodintotalhouseholdexpenditures

Indicatorofthehousehold’seconomicvulnerability.Associatedwithgeneralmeasuresofpoverty.

Access&Utilization

FoodConsumptionScore(FCS) Compositescorethatincorporatesdietarydiversity,foodfrequency,andrelativenutritionalimportanceofdifferentfoodgroups.

Starchystaplesratio(SSR) Percentageofcaloriesthatahouseholdderivesfromstarchystaples.Alsoassociatedwithgeneralmeasuresofpoverty.

Starchystaplesexpenditureratio(SSEXR)

Shareoffoodexpendituresdevotedtothepurchaseofstarchystaples.Alsoassociatedwithgeneralmeasuresofpoverty.

DietaryDiversityScore(HDDS) Reflectsthevarietyoffoodsthatahouseholdtypicallyconsumes.Capturesinformationaboutdietaryquality.DDSdenotesthenumberofatotalofsevenfoodgroupsconsumedduringthepast24hours.Thesevenfoodgroupsconsideredare(1)grains,rootsandtubers;(2)legumesandnuts;(3)dairyproducts;(4)fleshfoodsincludingorganmeats;(5)eggs;(6)vitaminArichfruitsandvegetablesincludingorangeandyellowvegetables;(7)andotherfruits.Ideally,thisscoreshouldbedeterminedseparatelyforthechildandmother.InourstudyitwasonlyavailableforthechildandinthecaseofIndonesiaonlyavailableasanaveragedietarydiversityscoreforthehousehold.

MinimumAcceptableDiet– Forchildrenundertheageof6months,theonlyacceptabledietconsideredisbreastfeeding.Forchildren6‐36monthstheminimumacceptabledietconsistsofaDDSof4orgreater,currentlybreastfedorreceivingmilkfeedings(includingcowandothermilksinadditiontoformulafeeds)andageappropriateminimummealfrequency.Forbreastfedchildren6‐8monthsofage,thechildneedstohavebeenfedatleasttwiceinthepast24hours,forchildren9to36monthsatleastthreetimes.Fornon‐breastfedchildrenfrom6to23monthsofage,thechildneedstohavebeenfedfourtimesinthepast24hours

12

DimensionsofFoodSecurity

FoodSecurityIndicator Description

HouseholdFoodInsecurityAccessScale(HFIAS)

Subjectiveindicatorbasedonperceptionsandexperience.Themeasureisbasedonstatementssuchas:(1)Youwereworriedyouwouldrunoutoffoodbecauseoflackofmoneyorotherresources;(2)Youwereunabletoeathealthyandnutritiousfoodbecauseoflackofmoneyorotherresources;(3)Youateonlyafewkindsoffoodbecauseoflackofmoneyorotherresources;(4)Youhadtoskipamealbecausetherewasnotenoughmoneyorotherresourcestogetfood;(5)Youatelessthatyouthoughtyoushouldbecauseoflackofmoneyorotherresources;(6)Yourhouseholdranoutoffoodbecauseoflackofmoneyorotherresources;(7)Youwerehungrybutdidnoteatbecauseoflackofmoneyorotherresources;(8)Youwentwithouteatingforawholedaybecauseoflackofmoneyorotherresources. LatermodifiedintotheHouseholdHungerScale(HHS)

ChildDietaryDiversityScore(CDDS)Reflectsthevarietyoffoodconsumedbythemotherandchildandis

basedonindividuallevelintakeinformationMother’sDietaryDiversityScore(MDDS)

2.1.2 AvailableIndicatorsforFood

Table3liststhevariousindicatorsoffoodsecurityeachcapturingdifferentdimensionsoftheidealfoodsecuritymeasureandtheindicatorsthatcanbeconstructedfromthesurveys(mainlyDHS)inthecountriesstudied.

Table3:ComponentsofFoodSecurityMeasured

IdealIndicators AVAILABLE

Children'sDietaryDiversityScore(CDDS) Yes

Mom’sDietaryDiversityScore(MDDS) No

MinimumAcceptableDiet(forchildren6‐24months) Yes

FoodInsecurityAccessScale(HFIAS) No(OnlyinHelenKeller)

FoodConsumptionScore(FCS) No

Relativepricesofdifferentfoodgroups No

PROXIESIFIDEALINDICATORSARENOTAVAILABLE

HouseholdDietaryDiversityScore(forchild/mom) HelenKeller

StarchyStapleRatioortheFractionofhouseholdCaloriesDerivedfromStarchyStaples

No(sincesurveysinthisstudydonotcontaindetailedhouseholdconsumptionmodule)

13

Mostofthedatasetsusedinthisstudycontaininformationtoconstructadiversityscoreforthechildandtheminimumacceptabledietmeasure,butlackinformationrelevantfortheotherdimensionsoffoodsecurity.Thediversityscoresmeasurestheconsumptionfromdifferentfoodgroups.Ingeneraltherearesevengroups(grains,rootsandtubers;legumesandnuts;dairyproducts;fleshfoodsincludingorganmeats;eggs;vitaminArichfruitsandvegetablesincludingorangeandyellowvegetables;andotherfruits)ofwhichanindividualneedsdailyconsumefour.9Theminimumacceptabledietcombinesdietarydiversity,breastfeedingandmealfrequencies.10Thedietarydiversityscoredependsontheageofthechild.Forexample,thoseundersixmonthsofageshouldbeexclusivelybreastfedwhereasolderinfantsshouldreceivecomplementaryfeedingsaswellasmilk.11FortheHelenKellerdatafromBangladesh,weareabletoconstructahouseholdlevelfoodinsecuritymeasure,butnotameasureofminimumacceptablediet.Achildhastomeetboththedietarydiversityandtheminimumacceptablediet(orfoodsecurity)definitionsbeconsideredadequateinfood.

2.2 MeasurementofChildCare

2.2.1 IdealIndicatorsforChildCare

Adequatecaremeasuresthecapacityofthechild’scaregivertoprovideahealthyenvironmentforthechildtogrowupin.Ideally,themeasureisbasedoninformationon(1)thecaregiver’seducation,knowledge,andbeliefs;(2)thehealthandnutritionalstatusofthecaregiver;(3)thementalhealth,lackofstress,andself‐confidenceofthecaregiver;(4)thecaregiver'sautonomyandcontrolofresources;(5)theworkloadandtimeconstraintsofthecaregiver;and(6)thesocialsupportreceivedbythecaregiverfromfamilymembersandthecommunity.Belowweexpandonthespecificcomponentsofanidealadequatecaremeasure.

CaringBehavior:Breast‐feedingandComplementaryFeeding,HealthSeeking,Hygiene‐Relatedo Caregiverfeedingbehavior(observationofoneormoreeatingepisodeso Caregiverresponsivenessduringfeedingepisodeso Frequencyofbehaviorsuchasfeeding,numberofspoonfuls,numberoftoucheso Breastfeedingpractices(e.g.,exclusivebreastfeedingupto6months,earlyinitiationof

breastfeeding,breastfeedingat2years)o Introductionofsolid/semi‐solid/softfoods6‐8monthso Childfeedingindex(constructedfromDHSdatausingthefollowingyes/noquestions:

currentbreastfeeding;useofbottles;dietarydiversity;feeding/mealfrequency)o Takingachildtoahealthclinicfortreatmentofillnesso Maternalhand‐washingwithsoap

MaternalEducation,knowledgeandbeliefs

9ForIndonesiathedietarydiversitymeasuresisonlyavailableforthehousehold,notspecifictoeachchild.10MealfrequencyinformationisnotavailableintheHelenKellerBangladeshsurveyorintheIndonesiaRKDsurvey.11Formilkintakeweincludebreastmilk,formula,andcow’smilk.

14

o YearsofSchoolingo Literate/illiterate(Selfreport,simpletestorexistingdata)o Beliefsandknowledgeaboutinitiationofbreastfeedingo Beliefsaboutcomplementaryfeeding–timing,types,controlofintake

WorkloadandTimeAvailabilityofCaregivero Observedtimeinchildcare(observedinsampleoftimeorcontinuously)o Recalledtimeinchildcare(24hourrecall)o Qualityofcareduringworktime(characteristicsofalternatecaregivers(e.g.age,gender)

SocialSupportforCaregivero Availabilityofalternatecaregiverso Communitysupport(assessmentofcommunityinstitutionsforchildcare‐feeding

programs,childcareprograms) PsychosocialCare

o Caregiver/childinteractions–naturalisticobservationofcaregiverandchildforashortperiod(codevariablessuchasdelaytorespond,typeofresponse,levelofvocalizationbycaregiverandchild)

o Childappearance(ratingofappearanceeitherinapublicplaceoroveraperiodofvisits)o Caregiver’sunderstandingofmotormilestones

2.2.2 AvailableIndicatorsforchildcare

Oftheidealcomponentswehaveinformationonlyonafewofthecaringbehaviors,namely,someinformationonbreast‐feedingandcomplementaryfeeding.Table4liststheidealandtheavailablemeasures.Inourmeasureofadequatecare,initialbreastfeedingforimmediateskin‐to‐skincontacthastohaveoccurredwithinthefirsthourafterbirth.12Forchildrenundertheageof6months,adequatecareconsistsofexclusivebreastfeeding.Forchildren6to8monthsofagewerequirecomplementaryfeedings.Allchildrenunder24monthsarerequiredtobebreast‐fed.Althoughthesurveyshaveinformationontheeducationallevelofthemother,thepresumedcaregiver,thereisnoconsensusonhowtotranslatethatinformationtoameasureofthemother’scaregivingabilitiesandthuswedonotincludeitinourmeasure.

Table4:ComponentsofCare

IDEAL AVAILABLE

Workloadandtimeavailabilityofcaregiver No

Socialsupportforcaregiver No

Psychosocialcare No

CaringBehaviors:Breast‐feeding Yes

12ForBoliviaandPeruinitialbreastfeedingcanonlybeidentifiedinthefirst100minutes(insteadof60minutes)andthemeasureisnotavailableintheIFPRIBangladeshdata.

15

CaringBehaviors:Healthseeking No

CaringBehaviors:Complementaryfeeding Yes

CaringBehaviors:Hygiene No

CaringBehaviors:Childfeedingindex No

Notes:Anotherimportantindicatorofcareismaternaleducationthoughthereisnoconsensusonthethreshold(orlevelofeducation)foradequatenutritioncarebehaviors

2.3 MeasurementofEnvironmentandHealth

2.3.1 IdealIndicatorsforenvironmentandhealth

Theidealindicatorforadequateenvironmentandhealthcapturesaccesstoenvironmentalinfrastructureandhealthserviceutilization.TheconstructionoftheidealmeasureofadequateenvironmentandhealthisbasedonthedefinitionsadoptedbyWHOandUNICEF.Namely,

Accesstosafewater–AccordingtoWHOandUNICEF(2006),thewatersourceisconsideredimprovedifthedrinkingwaterispipedintothedwelling/yard/plot,comesfromapublictap/standpipe,comesfromatubewelloraborewell,comesfromaprotectedwellorspringorrainwaterisused.Furthermore,thedrinkingwaterisconsideredimprovedifitcomesfromanunimprovedsource(suchassurfacewater,unprotectedwellorspring)butitisdisinfectedbyeitherboiling,addingbleach/chlorineorbysolardisinfection.Unprotectedspringsordugwells,cartwithsmalltank/drum,tanker‐trucks,surfacewaterorbottledwaterareallconsideredunimproved.

Accesstoadequatesanitation–FollowingWHOandUNICEF(2006)householdsaredefinedashavingaccesstoimprovedsanitationifthehouseholdusesaflushorpourflushlatrinewithalatrinepit,septictankorpipedsewersystem,aventilatedimprovedpitlatrine,apitlatrinewithslaboracomposingtoilet.Unimprovedsanitationfacilitiesincludeflush/pourflushtounknownornon‐closedsystem,pitlatrinewithoutslab,bucket,hangingtoiletorlatrineornofacilities.Thesanitationisconsideredimprovedonlyifitisnotshared.Duetolackofinformationonsharedfacilities,theconditionisdroppedinallothercountries,butCambodia.

Communitylevelsanitation–Measuresthepercentageofhouseholdsinthechild’slocality(i.e.village)withaccesstoadequatesanitation.Athresholdof75%isused,exceptforPeruwheresanitationinformationwasnotcollected.

Useofprenatalservices–Numberofprenatalvisitsbymotherwhilepregnant.TheWHO(2007)recommendsatleast4prenatalvisitsandtheadequacymeasureuses4visitsasthethreshold.

Immunizationstatus–BasedonnationalorWHO(2013)recommendedimmunizationschedules.Dependingontheageofthechild,therequiredvaccinesdiffer.Ingeneral,forexample,thefirstDTP3(diphtheria‐tetanuspertussis)isrecommendedat6weeksofage.Givingaleewayofthreemonths,achildwhois4monthsoldisincomplianceonlyiftheyhavereceivedthefirstDTP3vaccine.

16

VitaminASupplementationstatus–BasedontheWHOguidelinesthatrecommendationchildren6through59monthsshouldreceiveavitaminAsupplementinareaswherevitaminAdeficiencyisaknownpublichealthproblem.

Presenceofunpennedanimals–Measuresthelikelihoodthatachildmaybeincontactwithnon‐humansourcedfecalmatter.

ORSusefortreatmentofdiarrhea–Measureofuseoforalrehydrationsolutionsintreatingdiarrheainyoungchildren.

Antibiotictreatmentforpneumonia–Anationalmeasurecapturingthepercentageofchildrenaged0–59monthswithsuspectedpneumoniareceivingantibiotics.

Thesurveystypicallycontainmoreinformationtoconstructacomprehensivemeasureofenvironmentandhealththancomprehensivemeasuresoffoodorcare.Inourmeasure,weconsideraccesstosafewater,improvedsanitation,andrequirethatmorethan75%ofchild’scommunityhaveaccesstoimprovedsanitation.Intermsofprenatalhealthservices,amothermusthavehadatleastfourprenatalvisits.Forpost‐natalhealthserviceswerequirethechildtohavetheirimmunizationsuptodateandthatthechildhasreceivedavitaminAsupplementation(asdropsortablets)sincebirth.AlthoughsomeofthesurveyscollectedinformationonORSuseorantibiotics,theinformationwasonlyavailableforthosechildrenwhohadrecentlyexperienceddiarrheaorabacterialinfectionandnotforallchildren.Table5liststheidealandtheavailablecomponentstobeconsideredforadequateenvironmentandhealth.

Table5:ComponentsofEnvironmentandHealth

IDEAL AVAILABLE

Accesstosafewater Yes

Accesstoimprovedsanitation Yes(exceptBolivia)

Communitylevelsanitation Yes(exceptBolivia,Ethiopia,Indonesia,andHK)*

Useofprenatalservices Yes(exceptIndonesia)

Ageappropriateimmunizationstatus Yes(exceptHK)VitaminAsupplementationstatus(typicallyforchildren6monthsandolder)

Yes(exceptIndonesia,Bolivia)

ORSusefortreatmentofdiarrhea No*CommunitylevelsanitationwasnotincludedforBoliviaasnosanitationinformationwascollected.CommunitylevelsanitationwasnotincludedintheadequacymeasuresforIndonesiaaswedidnothaveaccesstothevillageidentifiers.CommunitysanitationwasdroppedfromtheanalysesinEthiopiaandBangladeshHKastheresultingnumberofchildrenwithaccesstoadequateenvironmentbecomenegligible.

2.4 Otherdataissuesandrelatedconsiderations

17

Sinceeachsurveyisdesignedslightlydifferently,caremustbetakenwhencomparingacrossdifferentcountries.Table6summarizesthecomparabilityofthedifferentindicatorsacrossthecountriesandthesurveyyearforeachcountry.Inmostcases,thetwosurveyyearshavecomparabledataanditispossibletocomparetheevolutionoftheindicatorswithinacountry.ThenotableexceptionisBoliviawhereslightlydifferentdatawerecollectedinthetwosurveyyearsandthecomparisonofadequatefoodoradequatecaremeasuresthroughtimeisnotpossible.Comparisonsacrosscountriesaremoredifficult.Therearetwosetsofcountrieswhichhavesimilarinformation.ThefirstiscomprisedofCambodiaandZimbabweandthesecondofEthiopiaandNepal.Ifthefactthatthefirstsetcollectsinformationforunder24‐montholdsandthesecondforunder36‐montholdsisnotaconcern,thenallthefourcountriescanbecompared.

Currentlymanyhouseholdlevelsurveyshaveinformationonaspectsofenvironmentandhealthbuttheinformationregardingfoodandcareaspectsofnutritionislessrobust.Understandingthesynergiesfromtheadequaciesandnutritionwouldbenefitfrommoredetaileddatacollectionespeciallyoffoodandcarecomponents.Informationonfoodsecurityiscrucialincomplementingfooddiversityandminimumacceptabledietcomponentstoconstructaholisticadequatefoodmeasure.Informationontheknowledgeofthecaregiverregardingbestchildcarepracticesismissinginallthesurveysthatwereviewed.

Anotherimportantconsiderationisthatcurrentlymoststudiesarecross‐sectionalandwithfewquestionsregardingthetimingofvariousactionsorinformationonpastconditions.Theimplicationisthatwecanonlyassessthechild’scurrentsituationandnotthechild’scumulativeexperience.Thefactthatachildmeetstherecommendationsforhisorherageatthetimeofthesurvey,doesnotimplythatheorshehasalwaysmettheageappropriateadequacydefinitions.Soforexampleinthemajorityofourstudies,forachildthatisayearold,itisnotpossibletoknowwhetherthechildwasexclusivelybreastfeduntil6monthsandreceivedatleastonecomplementaryfoodat6to8months.SimilarlyitisnotusuallypossibletoknowifvaccinationsorvitaminAsupplementswerereceivedattherecommendedages,sincemanyfamiliesonlyreportwhetherthechildhasreceivedthevaccinationorsupplementbutnottheactualdate.Somecomponents,suchasaccesstoimprovedsanitation,mostlikelyhavenotchangedforthemajorityofchildrensincetheirbirth,butotherbehaviors,suchashandwashingmaybecomponentsthatmayhavechanged.Thelackofpastinformationpotentiallyleadstoinflatednumbersofchildrenidentifiedasadequateinaparticularcomponentwheninfacttheyhavebeeninadequateforsomespanoftimesincebirth.

Manyoftheunderlyingmeasuresaredichotomousleadingtobinaryadequacymeasures.Thechallengeandlimitationwithbinarymeasuresistheunderlyingassumptionsthatneedtobemaderegardingcut‐offvaluesorconditions.Continuousadequacymeasureswouldallowtomeasurethecorrelationsbetweenchangesintheadequacylevelsandnutritionaloutcomesandthusprovidingmoredetailedinformationontherelationshipsandsynergies.However,itisnoteasytoenvisiontheconstructionofmeaningfulcontinuousadequacymeasures.

18

Table 6: Comparability of indicators across years and countries  

  Country and data year 

   Food  Environment & Health  Care 

Maxim

um age (months) 

Dietary diversity 

Food security 

Meal frequen

cy 

Accep

table diet 

Second round comparable to 

first round 

Improved sanitation 

Community sanitation 

Access to safe water 

Vaccinations 

Prenatal checkups 

Vitam

in A supplemen

tation 

Second round comparable to 

first round 

Exclusive breastfeeding 

Immed

iate skin‐to‐skin contact 

Complemen

tary feeding (6‐8 

months) 

Breastfeeding for 24 m

onths 

Second round comparable to 

first round 

Cambodia (2005)  24  Y  N  Y  Y    Y  Y  Y  Y  Y  Y    Y  Y  Y  Y   

Cambodia (2010)  24  Y  N  Y  Y  Y  Y  Y  Y  Y  Y  Y  Y  Y  Y  Y  Y  Y 

Ethiopia (2000)  36  Y  N  Y  Y    Y  Y2  Y  Y  Y  Y    Y  Y  Y  Y   

Ethiopia (2011)  36  Y  N  Y  Y  Y  Y  Y2  Y  Y  Y  Y  Y  Y  Y  Y  Y  Y 

Nepal (2000)  36  Y  N  Y  Y    Y  Y2  Y  Y  Y  Y    Y  Y  Y  Y   

Nepal (2012)  36  Y  N  Y  Y  Y  Y  Y2  Y  Y  Y  Y  Y  Y  Y  Y  Y  Y 

Bolivia (2003)  36  Y  N  N  N    N  N  Y  Y  Y  N    Y  Y4  Y  Y   

Bolivia (2008)  36  Y  N  Y  Y  N  N  N  Y  Y  Y  Y  N  Y  Y4  Y  Y  Y 

Peru (2005)  36  Y  N  Y  Y    Y  Y  Y  Y  Y  Y    Y  Y4  Y  Y   

Peru (2012)  36  Y  N  Y  Y  Y  Y  Y  Y  Y  Y  Y  Y  Y  Y4  Y  Y  Y 

Zimbabwe (2005)  24  Y  N  Y  Y    Y  Y  Y  Y  Y  Y    Y  Y  Y  Y   

Zimbabwe (2010)  24  Y  N  Y  Y  Y  Y  Y  Y  Y  Y  Y  Y  Y  Y  Y  Y  Y 

Bangladesh (HK, 2010)  24  Y  Y  N  N    Y  Y2  Y  Y  Y  Y3    Y  Y  Y  Y   

Bangladesh (HK, 2011)  24  Y  Y  N  N  Y  Y  Y2  Y  N  Y  Y3  N  Y  Y  Y  Y  Y 

Bangladesh (IFPRI, 2011)  24  Y  N  Y  Y  n/a  Y  Y  Y  Y  Y  Y  n/a  Y  N  Y  Y  n/a 

Indonesia (RKD, 2010)  24  Y1  N  N  N  n/a  Y  N  Y  Y  N  N  n/a  Y  Y  Y  Y  n/a 

NOTES: Y = yes, N = no; 1At the household level instead of child‐specific; 2Community sanitation not included in the environment adequacy measure used in the regressions (only in summary statistics). 3Vitamin A supplementation information asked for only 6 to 24 month olds. 4Within first 100 minutes (not 60 minutes) as is used in the other countries.   

19

3 Prevalenceofstuntingandaccesstoadequatefoodsecurity,childcare,andhealthandenvironment

Inlinewiththecommonpracticeintheliterature,thereportadoptsheight‐for‐ageasthemeasureofchronicmalnutrition.Giventhatstunting,height‐for‐agemorethantwostandarddeviationsbelowthemedianheight‐for‐agefortheparticularageandgender,captureschronicmalnutrition,itischosenoverwasting,weight‐for‐heightmorethantwostandarddeviationsbelowmedianweight‐for‐height,sincethelattermostfrequentlyindicatesarecentepisodeofsevereweightlossassociatedwithstarvationand/orseverediseaseresultinginacutemalnutrition.

3.1 Prevalenceofstunting

Althoughstuntingremainshighinmostofthestudycountries,ithasdecreasedsignificantlyovertimeforbothundertwo‐yearoldsaswellasforunderfive‐yearolds.Figure2depictsstuntingintheundertwo‐yeargroupintheeightcountriesandFigure3depictsstuntingintheunderfive‐yeargroup.ItisimportanttobearinmindthatforBolivia,Cambodia,Ethiopia,Nepal,Peru,andZimbabwethestuntingratesarebasedonstuntingratesreportedbytheStatCompilerofDHSfortheindicatedyearsusingallthechildrenwithheightinformation.ForBangladeshtheinformationisderivedfromtheHKandIFPRIdatasetsandforIndonesiafromtheRKDdataset.Inbothagegroupsthesevencountriesforwhichtherearemultipleyearsofdatashownoincreasesinstunting.InEthiopiaandNepalthereductionshavebeenthegreatest.From2001to2011,Nepalexperienceda17percentagepointreductionintheprevalenceofstuntinginNepalintheunder5‐yearcohortand15pointreductionintheunder2‐yearcohort.InEthiopiafrom2000to2011,stuntingreducedby13percentagepointsinbothagecohorts.EthiopiaandNepalwerealsothetwocountrieswiththehighestbaselinestuntingrates.Inaddition,Peruachievedlargereductionsinstunting,bynearlyhalvingthestuntingprevalenceintheunder5‐yearoldcohortsothatin2012theprevalenceofstuntingwas18%.Intheothercountriesthereductionshavebeenintherightdirectionbutmodestreductionsoffivepercentagepointsorless.

20

Figure2:Stuntingunder2‐yearolds

Figure3:Stuntingunder5‐yearolds

3230

37

25

20

2927

44

3130

42

26

22

17

30

26

010

2030

40P

erce

nta

ge

of c

hild

ren

with

HA

Z <

-2

SD

Bangladesh (HK)Bangladesh (IFPRI)

BoliviaCambodia

EthiopiaIndonesia (RKD)

NepalPeru

Zimbabwe

Source: Author estimates. Data for Bangladesh from 2010 (HK) and 2011 (HK, IFPRI); Bolivia 2003, 2008; Cambodia 2005, 2010; Ethiopia 2000, 2011; Indonesia 2010 (RKD); Nepal 2000, 2011; Peru 2005, 2012; Zimbabwe 2005, 2010

Based on official DHS estimates when available

Stunting in children 0 to 23 months

Stunting year 1 Stunting year 2

36

33

43

32

27

43

40

58

44

34

57

41

29

18

35

32

020

4060

Per

cen

tag

e o

f ch

ildre

n w

ith H

AZ

< -

2 S

D

Bangladesh (HK)Bangladesh (IFPRI)

BoliviaCambodia

EthiopiaIndonesia (RKD)

NepalPeru

Zimbabwe

Source: Author estimates. Data for Bangladesh from 2010 (HK) and 2011 (HK, IFPRI); Bolivia 2003, 2008; Cambodia 2005, 2010; Ethiopia 2000, 2011; Indonesia 2010 (RKD); Nepal 2000, 2011; Peru 2005, 2012; Zimbabwe 2005, 2010

Based on official DHS estimates when available

Stunting in children 0 to 59 months

Stunting year 1 Stunting year 2

21

Inalleightcountries,childrenlivinginresource‐poorhouseholdsaremorelikelytobestuntedthanchildrenlivinginresource‐richhouseholds.Ahouseholdisclassifiedasresourcerichifitisinthetop60%ofanassetindexandasresourcepoorthebottom40%oftheindexvalue.Annex1describestheconstructionoftheassetindex.ThedifferenceinstuntingratesbyaccesstoresourcesisespeciallymarkedinPeruwhereonly6%ofthechildreninresource‐richhouseholdsarestunted,but20%ofchildreninresource‐poorhouseholdsarestunted—aquadruplingoftheprevalence(Figure4).Similarly,theotherLACcountryofthestudy—Bolivia—showedlargedifferencesinstuntingratesbywealth.Childreninresource‐poorBolivianhouseholdsweremorethantwiceaslikelytobestuntedaschildreninresource‐richhouseholds.TheothercountrieswithalargeprevalencegapbyresourceaccesswereBangladeshbasedontheHelenKellerdataandNepal.InCambodia,Ethiopia,Indonesia,andZimbabwe,thestuntingratesforresource‐poorandresource‐richchildrenwerewithin4percentagepoints.

Figure4:Stuntingresource‐poorvsresource‐richhouseholds

3.2 AdequaciesinFoodSecurity,Environment,Health,andCarePractices

Giventheindicatorsavailableinthesurveys,thedeterminationofwhetherchildrenhaveaccesstoanadequateorinadequatelevelofeachspecificunderlyingdeterminantofnutritioniscarriedoutusingacceptedinternationalstandardsregardinginfantandchildfeedingpractices,foodsecurity,improvedwaterandsanitation,andpre‐aswellaspost‐natalcarepractices.Specifically,theWHO(2008)standardsareusedforassessinginfantandyoungchildfeedingpractices,theUSAID(2012)

42

26

44

32

35

16

3735

32

28 28

25

47

30

20

6

25

22

010

2030

4050

Per

cen

tag

e o

f ch

ildre

n w

ith H

AZ

< -

2 S

D

BGD, HKBGD, IFPRI

BOL*ETH*

IDNKHM

NPL*PER*

ZWE

Source: Author estimates. Data for Bangladesh from 2010 (HK) and 2011 (IFPRI); Bolivia 2008; Cambodia 2010; Ethiopia 2011; Indonesia 2010 (RKD); Nepal 2011; Peru 2012; Zimbabwe 2010. *Stunting for under 36 month olds. The rest under 24 month olds.

Stunting in children 0 to 23 or 35 months

Resource poor Resource rich

22

standardsonmaternaldietarydiversity,theWHO(2013)recommendationsonchildimmunizationsschedules,theWHOandUNICEF(2006)guidelinesondrinkingwaterandsanitation,aswellasthe1990UNICEFstrategyonimprovednutritionofchildrenandwomen.InBoliviaandPerucountryspecificimmunizationpracticesareused.TheseimmunizationschedulesfollowcloselytheWHOguidelines.

Theprevalenceofadequaciesvariesacrosscountriesandresourceaccess,butinmostcontextschildrenweremostlikelytobeadequateincareandleastlikelytobeadequateinenvironmentandhealth.AdequatecarewasthefacetwhichwasmostlikelytobemetinallcountrieswiththeexceptionofIndonesia,wheretheprevalenceofadequatefoodwasgreaterthantheprevalenceofadequatecare(Figure5a,6aand7a).Inallcountriesitisadequateenvironmentandhealththatismostlacking.WiththeexceptionofBolivia,inthemostrecentsurveyforeachcountryonly10%orlessofthechildrenlivedinahouseholdwherethehousinginfrastructureandhealthopportunitieswereadequate(Figure7a).However,forBolivianoinformationonsanitationwascollectedandthusthemeasureisnotcomparabletotheothercountrieswheresanitationmeasureswereincluded.

3.2.1 Adequatefoodsecurity

Inmostcountrieseitheradequatedietarydiversityoradequatemealfrequencyisthecomponentoffoodsecuritywhichisleastprevalentamongchildren.Infourofthesevendatasetwithcomparableinformationwefindmealfrequencytobetheleastprevalentcomponentandinthreeitisdietarydiversity(Figure5a).ThetwoSouthAmericancountrieshaveabove80%ofthechildrenconsumingfoodsfromatleastfourofthesevengroups.Indonesiaalsohasmorethan80%ofchildrenwithdiversediets,however,themeasureisnotcomparabletotheotherssinceitisbasedonhouseholdfoodconsumptionandnotchild‐specificfoodconsumption.ExceptforEthiopiaandZimbabwe,morethanhalfofthechildrenunder6monthsareexclusivelybreastfed.ThehighestprevalenceofexclusivelybreastfedchildrenareinCambodia(78%)andPeru(81%).

23

Figure5a:ComponentsofAdequateFoodSecurity

Applyingstricterstandardsbyrequiringsimultaneousaccesstoadequatelevelsofeachcomponentoffoodsecurity,lessthanone‐halfoftheyoungchildreninthestudycountriesappeartohaveaccesstoadequatefoodsecurity.Infactinsixoutofthe9surveyslessthanone‐thirdofthechildrenhadaccesstoadequatefood(Figure5b).InEthiopia,withthelowestaccesstoadequatefoodsecurity,only12percentofthechildrenmetthecriteriaforadequatefoodsecurity.Asfigure5aaboveclearlyindicates,thisisprimarilyduetotheverylowprevalenceofadequatedietarydiversity(13percent).BoliviaandIndonesiaweretheonlycountrieswheremorethanhalfofthechildrenhadaccesstoadequatefoodsecurity,at51and68percent,respectively.However,inIndonesiathedietarydiversitywasnotmeasuredforeachchild,andthemeasureisbasedexclusivelyonthedietarydiversityatthehouseholdlevelsoitisnotcomparablewiththeotherstudies.Dependingonhowdifferentthechild’sdietisfromthegeneraldietofthefamily,thetrueprevalenceofadequatefoodsecuritymaybemuchsmaller.

Inmostcountries,theimprovementsinaccesstoadequatefoodsecurity(asdefinedabove)havebeenmodest(Cambodia,PeruandZimbabwe)tonone(Ethiopia,Nepal)inthe5to10yearsbetweensurveys.ThedecreaseinBoliviainaccesstoadequatefoodsecurityisduetothechangeinthedefinition.In2003noinformationwascollectedonmealfrequenciesandthustheminimumacceptabledietvariablecouldnotbeconstructed.In2008theinformationwasavailableanditwasincludedaspartofthefoodadequacymeasure,thusmakingthe2008definitionamorecomprehensiveonethanthe2003measure.Furthermore,basedontheHelenKellerdata,inBangladeshtheaccesstoadequatefoodhalvedbetween2010and2011.Comparingthesecond

45

55

72

50 51

31

80

67

54

13

49

22

86

69

33

78

66

30

67

37

8681

66

40

3026

020

4060

8010

0P

erce

nta

ge

of c

hild

ren

wh

o m

eet

crit

eria

BGD (HK, '11)BGD (IFPRI, '11)

BOL ('08)ETH ('11)

IDN (RKD, '10)KHM ('10)

NPL ('11)PER ('12)

ZWE ('10)

Source: Author estimates.Note: (1) Dietary diversity: 4 out of 7 food groups. In Indonesia based on household (not child specific) dietary diversity;(2) Meal frequency depends on age

Components of Adequate Food

Dietary diversity > 5 months Exclusively breastfed < 6 months

Meal frequency Food security

24

yearoftheHelenKellersurveyandtheIFPRIsurveywhichwerebothcollectedin2011,weobservelargedifferencesinthenumberofchildrenclassifiedasadequateinfood.BasedontheIFPRIsurvey33%oftheruralchildreninBangladeshhadaccesstoadequatefoodbutonly14%ofbothruralandurbanchildrenwerefoundtobeadequateinfoodintheHelenKellersurvey.Besidessamplingadifferentpopulation,thetwosurveyslendthemselvestodifferentdefinitionsofadequatefood.InHelenKelleritisnotpossibletodeterminewhetherachildhasmettheminimumacceptabledietbasedonthecompositionandfrequencyofmeals,butitpossibletodeterminewhetherthehouseholdhasexperiencedfoodinsecurity.IntheIFPRIsampleitispossibletodeterminetheminimumacceptablediet,butnotwhetherthehouseholdexperiencedfoodinsecurity.

Figure5b:TheEvolutionofAccesstoAdequateFoodSecurity

Childreninresource‐poorhouseholdsarelesslikelytohaveaccesstoadequatefoodsecuritythanchildreninresource‐richhouseholds,butingeneralthedifferencesaresmall.Althoughinallthecountriesanalyzedchildreninresource‐poorhouseholdswerelesslikelytohaveaccesstoadequatefood,thereweredifferencesinthedegreeofdisparity(Figure5c).13Thelargestdisparity

13Weusethemostrecentyearofdataforthecomparisonsbetweenresource‐poorandresource‐richhouseholds.TheonlyexceptionistheHelenKellerdataforBangladeshwhereweuse2010andnot2011,thereasonbeingthatnovaccinationinformationwasavailablefor2011.

29

14

34

57

51

12 12

68

3134

29 28

4346

15

22

020

4060

8010

0P

erce

nta

ge

of c

hild

ren

wh

o m

eet

crit

eria

BGD, HKBGD, IFPRI

BOLETH

IDNKHM

NPLPER

ZWE

Source: Author estimates. Data for Bangladesh from 2010 (HK) and 2011 (HK, IFPRI); Bolivia 2003, 2008; Cambodia 2005, 2010;Ethiopia 2000, 2011; Indonesia 2010; Nepal 2000, 2011; Peru 2005, 2012; Zimbabwe 2005, 2010.Note: (1) Bangladesh (HK) includes food security but no information on meal frequencies.(2) Bolivia (2003) does not include information on dietary diversity.(3) Indonesia (RKD) only has information on household level dietery diversity, and no information on meal frequencies.

Prevalence of Adequacy in Food

Year 1 Year 2

25

wasobservedinBangladeshfortheHelenKellersample,whichincludebothurbanandruralchildren,wherethedisparitywas25percentagepoints.IntheIFPRIsamplewhichonlyincludesruralchildren,therewasnodifferencebetweenthetwogroups.InBolivia,Cambodia,NepalandZimbabwe,thedifferencebetweenthetwogroupswasbetweensevenandfourpercentagepoints.Thesefindingsareconsistentwiththefundamentalroleofnutrition‐relatedbehaviorsandcaringpracticesandthecriticalneedfornutritionknowledgeacrossallincomegroups.InEthiopia,IndonesiaandPerubothresource‐richandresource‐poorchildrenhadsimilaraccesstoadequatefoodasthedifferenceswereonlyonepercentagepoint.

Figure5c:AdequateFoodSecuritybyHouseholdWealth

3.2.2 AdequateCarePractices

Morethan80%ofthechildrenunderageof24monthswerebreastfedinthecountriesstudied,butaccesstotheotherthreecomponentsofadequatecarewasmixed.14InZimbabweonly30%ofthechildrenundertheageofsixmonthswereexclusivelybreast‐fedandonlyinCambodiaandPeruweremorethanthree‐fourthsofthechildrenundersixmonthsexclusivelybreast‐fed(Figure6a).Earlyinitiationofbreastfeedingrangedfrom45%ofchildreninNepalto78%ofchildreninBolivia.

14Sinceweusedchildrenunder24or36monthsofage,thisdoesnotimplythatmorethanfour‐fifthsofthechildrenreceivedmilkuntiltheywere24monthsold.

13

3834 34

47

53

12 13

67 68

32

37

2529

46 47

18

25

020

4060

8010

0P

erce

nta

ge

of c

hild

ren

wh

o m

eet

crit

eria

BGD (HK, '10)BGD (IFPRI, 11)

BOL ('08)ETH ('11)

IDN ('10)KHM ('10)

NPL ('11)PER ('12)

ZWE ('10)

Source: Author estimates.Note: (1) Bangladesh (HK) include food security but not information on minimum acceptable diet.(2) Indonesia (RKD) only has information on household level diatery diversity, nothing else.

By wealth

Prevalence of Adequacy of Food

Resource Poor Resource Rich

26

Complementaryfeedingsforsix‐toeight‐montholdswashighinBangladesh(intheIFPRIsample),Nepal,ZimbabweandBolivia,wheremorethan80%ofthechildrenoftheagerangereceivedcomplementaryfoods.InNepal,only44%ofthesix‐toeight‐montholdsreceivedcomplementaryfeedingsfromoneofthedietarydiversityfoodgroups.

Figure6a:ComponentsofAdequateCare

Applyingstricterstandardsbyrequiringsimultaneousaccesstoadequatelevelsofeachcomponentofchildcare,accesstoadequatecarehasincreasedsignificantlyovertimeinmostcountries.Adequatecareincreasedby24percentagepointsinCambodia,14percentagepointsinNepal,9percentagepointsinEthiopia,6percentagepointsinPeru,5percentagepointsinBangladesh,and3percentagepointsinZimbabwe(Figure6b).15Forthemostrecentyearofdata,mostcountrieshadaboutone‐thirdtotwo‐thirdsofthechildrenlivinginsituationswithadequatecare.However,adequatecarewasalsothemeasureforwhichmanyoftheidealcomponentsweremissing.Nonetheless,theresultssuggestanincreasedawarenessoftheimportanceofearlyinitiationofbreastfeedingandcontinuedbreastfeedinguntiltwoyearsofage,aswellasfollowingtherecommendationsonage‐appropriatecomplementaryfeedings.

15InBoliviatheprevalenceofadequatecarewasthesameforthetwosurveys.However,the2003definitiondoesnotincludevitaminAsupplementation(astheinformationwasnotcollected)andthe2008definitiondoes.

55

68 69

91

51

10096

67

7880 80

49 51 50

83

69

48

100

78

69

59

82

67

45 44

98

81

74

86

80

30

63

8482

020

4060

8010

0P

erce

nta

ge

of c

hild

ren

wh

o m

eet

crit

eria

BGD (HK, '11)BGD (IFPRI, '11)

BOL ('08)ETH ('11)

IDN (RKD, '10)KHM ('10)

NPL ('11)PER ('12)

ZWE ('10)

Source: Author estimates.Note: (1) Exclusive breastfeeding for first 6 months; (2) Complementary feedings for 6- to 8-month olds;(3) Breastfeeding within 1 hour of birth (100 minutes for Bolivia and Peru);(4) Breast-fed for 24 months or currently breastfeeding if less than 24 months.

Components of Adequate Care

Exclusive breastfeeding Early breastfeeding initiation

Complementary feedings Breast-fed up to 24 months

27

Figure6b:TheEvolutionofAccesstoAdequateCare

Childreninresource‐poorhouseholdsaremorelikelytohaveaccesstoadequatecarethanchildreninresource‐richhouseholds.Possiblycontrarytoexpectation,inallcountriesexceptforNepal,childreninresource‐poorhouseholdsweremorelikelytohaveaccesstoadequatecarethanchildreninresource‐richhouseholds(Figure6c).Thatis,childrenfromresource‐poorhouseholdsdidnothaveworseaccesstothecomponentsofadequatecarethatweremeasured.Althoughthedifferencesinmostcasesweremodest,inPeru72%ofthechildreninresource‐poorhouseholdswereadequateincarewhereasonly52%ofchildreninresource‐richhouseholdswereadequate,adifferenceoftwentypercentagepoints.InBangladesh(IFPRI)andBoliviathedifferenceinadequatecarewaseightpercentagepointsandfortherestthedifferencesweresixpercentagepointsorless.Acrosstheeightcountries,childreninresourcepoorhouseholdstendedtobebreastfedlonger,butwerelesslikelytoreceivecomplementaryfeedsduringtheagesofsixtoeightmonths(Annex2).Overall,thesefindingsfurtherunderscoretheimportantroleofnutritionknowledgeregardlessofsocioeconomicstatus.

5055

84

59 59

28

37 36

28

52

24

38

55

61

3639

020

4060

8010

0

BGD, HKBGD, IFPRI

BOLETH

IDNKHM

NPLPER

ZWE

Source: Author estimates. Data for Bangladesh from 2010 (HK) and 2011 (HK, IFPRI); Bolivia 2003, 2008; Cambodia 2005, 2010;Ethiopia 2000, 2011; Indonesia 2010; Nepal 2000, 2011; Peru 2005, 2012; Zimbabwe 2005, 2010.Note: (1) In Peru and Bolivia early initiation of breastfeeding is within 100 minutes, not within 60 minutes as in others.(2) Bolivia 2003 does not include information on Vitamin A supplemenation.(3) Bangladesh (HK, 2011) does not include information on vaccinations.(4) Bangladesh (IFPRI) does not include information on early initiation of breastfeeding

Prevalence of Adequacy in Care

Year 1 Year 2

28

Figure6c:AccesstoAdequateCarebyHouseholdWealth

3.2.3 AdequateEnvironmentandHealth

Ingeneralchildrenlackaccesstoimprovedwaterandsanitationinfrastructure.Figure7ahighlightsthelackofaccesstoimprovedsanitationatthehouseholdandcommunitylevel.Evenifachildhasaccesstoimprovedsanitationintheirhome,thegreatmajorityofthechildrenliveincommunitieswherelessthan75%ofthehouseholdsofthecommunityhaveaccesstoimprovedsanitation(Figure7a).OnlyintheIFPRIsampleofBangladeshichildrendomorethan50%ofthechildrenlivecommunitieswhere75%ofthehouseholdshaveaccesstoimprovedsanitation.InEthiopiaonly3%ofthechildrenliveinsuchcommunitiesandinCambodiaonly11%.However,exceptforEthiopia,morethanthree‐fourthsofchildreninthesurveyshaveaccesstoimproveddrinkingwater.Thatis,themajorityofthechildrendrinkwaterthathasbeentreatedtoreducedisease‐producingcontaminants.

52 50

89

81

64

56

3937

40

34

5451

3639

72

52

4038

020

4060

8010

0P

erce

nta

ge

of c

hild

ren

wh

o m

eet

crit

eria

BGD, HK '10BGD, IFPRI '11

BOL '08ETH '11

IDN '10KHM '10

NPL '11PER '12

ZWE '10

Source: Author estimates.Note: (1) In Peru and Bolivia early initiation of breastfeeding is within 100 minutes, not within 60 minutes as in others.(2) Bangladesh (IFPRI) does not include information on early initiation of breastfeeding(3) Indonesia does not include information on complementary feeds for 6 to 8 month olds.

By wealth

Prevalence of Adequacy of Care

Resource Poor Resource Rich

29

Figure7a:ComponentsofAdequateEnvironment

Incontrasttothecaseofbasicinfrastructure,childrenappeartodobetterinhavingaccesstoadequatepreventivehealthservicesandfacilities.Intermsofvaccinations,therearetwogroupsofcountries.ThefirstgroupconsistingofBangladesh,Bolivia,Cambodia,PeruandZimbabwe,wheremorethantwo‐thirdsofchildrenhavehadreceivedageappropriatevaccinations.ThesecondgroupconsistingofEthiopia(15%),Indonesia(17%)andNepal(30%)wherelessthanone‐thirdhavereceivedthem(Table7b).MorethanhalfofthechildreninBolivia,Cambodia,NepalandZimbabwehavehadfourormoreprenatalvisits,andlessthanathirddidsoinBangladeshandEthiopia.Morethan50%ofthechildreninthestudieshavereceivedvitaminAsupplementation,exceptforinEthiopia(47%)andPeru(10%).InEthiopianoneofthecomponentsreach50%prevalence.

47

18

93

75

65

7981

15

3

29

64

81

28

11

84

39

20

76

53

46

75

56

33

76

020

4060

8010

0P

erce

nta

ge

of c

hild

ren

wh

o m

eet

crit

eria

BGD (HK, '11)BGD (IFPRI, '11)

BOL ('08)ETH ('11)

IDN (RKD, '10)KHM ('10)

NPL ('11)PER ('12)

ZWE ('10)

Source: Author estimates.

Components of Adequate Environment

Improved sanitation Improved community sanitation

Improved drinking water

30

Figure7b:ComponentsofAdequateHealthServices

Applyingstricterstandardsbyrequiringsimultaneousaccesstoadequatelevelsofeachcomponentofenvironment/infrastructureandhealthservices,accesstoadequateinfrastructureappearstohaveincreasedovertimeinmostcountries.Forexample,accesstoadequateinfrastructurehasincreasedinBolivia,Nepal,Cambodia,andPeru(Figure7c).Ethiopiashowsverylittleprogressprimarilybecauseofthelowprevalenceofimprovedsanitationatthecommunitylevel(seealsoFigure7a).Alongsimilarlines,accesstoadequatehealthserviceshaveincreasedinBolivia,Cambodia,andZimbabwe,withlittleorveryminorprogressinEthiopia(Figure7d).16

16InPeruthedeclinewascausedbythedeclineinvitaminAsupplementaldosecoveragewhichfellfrom22%to10%.Alltheothercomponentsofthemeasureimprovedovertime.

23

53

80

26

72

80

72

67

15

28

47

17

80

64

58

30

55

7276

95

10

66

61 60

020

4060

8010

0P

erce

nta

ge

of c

hild

ren

wh

o m

eet

crit

eria

BGD (HK, '11)BGD (IFPRI, '11)

BOL ('08)ETH ('11)

IDN (RKD, '10)KHM ('10)

NPL ('11)PER ('12)

ZWE ('10)

Source: Author estimates.Note: For Bangladesh (HK) Vitamin A supplementation is only for children over 6 months.

Components of Adequate Health

Adequate vaccinations 4+ prenatal visits

Vitamin A supplementation

31

Figure7c:TheEvolutionofAccesstoAdequateEnvironment

Figure7d:TheEvolutionofAccesstoAdequateHealthServices

Childreninresource‐poorhouseholdsarelesslikelytohaveaccesstoadequateenvironmentandhealthservicesthanchildreninresource‐richhouseholds.Thus,incontrasttothesituationwith

1714

45

71

81

0 1

49

5

10

2

16

32

38 36

31

020

4060

8010

0P

erce

nta

ge

of c

hild

ren

wh

o m

eet

crit

eria

BGD, HKBGD, IFPRI

BOLETH

IDNKHM

NPLPER

ZWE

Source: Author estimates. Data for Bangladesh from 2010 (HK) and 2011 (HK, IFPRI); Bolivia 2003, 2008; Cambodia 2005, 2010;Ethiopia 2000, 2011; Indonesia 2010; Nepal 2000, 2011; Peru 2005, 2012; Zimbabwe 2005, 2010.Note: (1) Bolivia (2003 and 2008) do not include improved sanitation.(2) Ethiopia, Nepal, Bolivia, and Indonesia do not include community sanitation.

Prevalence of Adequacy in Environment

Year 1 Year 2

21

1618

42

78

4 5

18

4

35

0

13 13

8

22

27

020

4060

8010

0P

erce

nta

ge

of c

hild

ren

wh

o m

eet

crit

eria

BGD, HKBGD, IFPRI

BOLETH

IDNKHM

NPLPER

ZWE

Source: Author estimates. Data for Bangladesh from 2010 (HK) and 2011 (HK, IFPRI); Bolivia 2003, 2008; Cambodia 2005, 2010;Ethiopia 2000, 2011; Indonesia 2010; Nepal 2000, 2011; Peru 2005, 2012; Zimbabwe 2005, 2010.Note: (1) Bangladesh (IFPRI) has Vitamin A supplementation for 6 to 24 month olds only.(2) Indonesia does not have information on prenatal checkups or Vitamin A supplementation.

Prevalence of Adequacy in Health

Year 1 Year 2

32

childcarewheretheresourcepoorhouseholdsappearedtohavehigheraccesstoadequatecarethanchildreninresourcerichhouseholds(seefigure6c),childreninresource‐poorhouseholdshaveloweraccesstoadequateinfrastructureandadequatehealthservices(Figures7eand7f).).Thesedifferencesaresubstantial.ExceptforEthiopia,thedifferencesinaccesstoadequateenvironmentalinfrastructurearemorethantenpercentagepointsinthecountriesstudiedandmorethan25percentagepointsinBolivia,Peru,andZimbabwe.ExcludingPeru,thedifferencesinaccesstoadequatehealthservicesrangesfrom4percentagepointsto12percentagepointsinoursetofcountries,favoringchildreninresource‐richhouseholds.InPeruchildreninresourcepoorhouseholdsaremorelikelytohaveaccesstoadequatehealthservices.Theresultisdrivenbythehigherlikelihoodofachildfromaresource‐poorhouseholdinhavinghadavitaminAsupplement.

Figure7e:AccesstoAdequateEnvironmentbyHouseholdWealth

8

22

38

50

63

94

0 1

41

54

0

17

7

20

11

57

4

50

020

4060

8010

0P

erce

nta

ge

of c

hild

ren

wh

o m

eet

crit

eria

BGD, HK '10BGD, IFPRI '11

BOL '08ETH '11

IDN '10KHM '10

NPL '11PER '12

ZWE '10

Source: Author estimates.Note: (1) Bolivia does not include improved sanitation.(2) Ethiopia, Nepal, Bolivia, and Indonesia do not include community sanitation.

By wealth

Prevalence of Adequacy of Environment

Resource Poor Resource Rich

33

Figure7f:AccesstoAdequateHealthServicesbyHouseholdWealth

3.3 AdditionalConsiderations:WeakLinksinaChain

Theprecedingdiscussionsuggestsacorrelationbetweentheobservedreductionsinstuntingandtheincreasesinaccesstoadequatefoodsecurity,adequatecareandadequateenvironmentandhealthservicesoverthesameperiodinthecountriesstudied.Forthecountrieswithtwoyearsofsimilarsurveydata,theprevalenceofstuntingfell(Figures2and3).Inthesametimespanthereweremarkedimprovementsintheadequacyofcare(Figure6b),someimprovementinadequateenvironmentandhealth(Figures7cand7d)andamixedstoryinadequatefoodwithsomecountriesexperiencingadecreaseintheaccesstoadequatefood(Figure5b).Whetherthereisindeeda“causal”relationshipasopposedtoasimplecorrelationbetweenimprovedaccesstoadequatefood,careandenvironmentandhealthandtheobservedreducedstuntingandincreasedaverageheight‐for‐ageofchildren,isaquestionthatneedstobeaddressedwithmorecomplexmodelsusingdetailedpaneldataonchildrenthroughtime.

Asthefirststeptowardsexploringpotentialsynergiesinrelationtochildnutrition,theanalysisbelowinvestigatesinmoredetailtheextenttowhichchildrenhavesimultaneousaccesstoadequatelevelsinanyone,ortwo,orallthreeoftheclustersofunderlyingdeterminantsofnutrition.Sofartheaccesstoadequateenvironmentwastreatedseparatelyfromtheaccesstoadequatehealthservices.Intheremainderofthisreportenvironmentandhealtharecombinedintooneclustermainlyforthepurposeofsimplifyingthepresentation.Oneshouldbearinmind,thatcombininghealthandenvironmentintooneaggregateclustercreatesaverystrictsetofstandards

13

25

11

22

38

47

26

1719

31

38

9

14 14

4

22

30

020

4060

8010

0P

erce

nta

ge

of c

hild

ren

wh

o m

eet

crit

eria

BGD, HK '10BGD, IFPRI '11

BOL '08ETH '11

IDN '10KHM '10

NPL '11PER '12

ZWE '10

Source: Author estimates.Note: (1) Bangladesh (IFPRI) has Vitamin A supplementation for 6 to 24 month olds only.(2) Indonesia does not have information on prenatal checkups or Vitamin A supplementation.

By wealth

Prevalence of Adequacy of Health

Resource Poor Resource Rich

34

foradequacyinhealthandenvironmentrequiringsimultaneousaccesstoimprovedlevelsforeachcomponentofinfrastructureaswellasaccesstoadequatelevelforeachcomponentofhealthservices.Inthiscasethe“chainisonlyasstrongasitsweakestlink”inthesensethataslongasthereisverylowaccesstoanadequatelevelinanyoneofthecomponentsofthecluster(e.g.sanitationatthecommunitylevel)thenoverallaccesstoadequatehealthandenvironmentisforcedtobeverylow,inspiteofconsiderablyhigheraccesstoadequatelevelsintheothercomponentsofhealthandenvironment.Theprecedingargumentisalsovalidforthedeterminationofwhetherachildhasaccesstoadequatecareoradequatefoodsecurity.Aslongaseachoneoftheclustersconsistsofnumerousindicatorsmeasuringthedifferentdimensionsofchildcareorfoodsecurity,adequatecareoradequatefoodsecurityoverallwillbedeterminedbythelevelofadequacyinoneofthesub‐componentsorthe“weakestlink.”Inotherwords,theanalysisfollowingtakesitasagiventhatthethreeclusters—food,care,environmentandhealth—andthevarioussub‐componentsofeachoftheseclustersworkintandemtoproducenutritionaloutcomes.

UsingZimbabweasanexample,althoughthereare22%ofchildrenadequateinfood,39%adequateincare,and10%adequateinenvironmentandhealthcombined(Figures5a,6aand7a),manychildren(47%)arenotadequateinanycomponentandonlyafew(17%)areadequateinmorethanonecomponent(Figure8a).Asexpected,fromtheearlieranalysis,childreninresource‐richhouseholdsaremorelikelytobeadequateintwoofthethreefacets(17%vis‐à‐vis11%)andinallthreecomponents(4%vis‐à‐vis0%)thanchildreninresource‐poorhouseholds(Figure8b).ThesamegeneralpatternsappeartoholdforNepal(seeFigures8cand8d).

Figure8a Figure8b

47

8

23

6

15

2

020

4060

8010

0P

erc

ent

age

of

child

ren

who

me

et c

rite

ria

Source: Author estimates.

Adequacy status in Zimbabwe (2010)

None Food only

Care only Environmental health only

Two of three adequacies All three adequacies

53

7

29

0

11

0

44

8

18

9

17

4

020

4060

8010

0P

erc

ent

age

of

child

ren

who

me

et c

rite

ria

Resource-poor Resource-rich

Source: Author estimates.

By wealthAdequacy status in Zimbabwe (2010)

None Food only

Care only Environmental health only

Two of three adequacies All three adequacies

35

Figure8c Figure8d

Overall,thisdescriptiveanalysisrevealsthat(i)thereareonlyafewchildrenadequateinallthreeclustersand(ii)manychildrenareadequateinnoneorinonlyoneoftheunderlyingclusters.WiththeexceptionofBoliviaandBangladesh(IFPRI),morethanonequarterofthechildrendonothaveadequateaccesstoanyofthethreepillars(Figure9).Bolivia,withthecaveatthattheadequateenvironmentdoesnotmeasuresanitation,hasthehighestpercentage(13percent)ofyoungchildrenwithaccesstoadequatelevelsinallthreeclusters.TheothercountriesincludedinthestudyhavelessthanfivepercentofthechildrenmeetingthethreeadequacieswithCambodiaandEthiopiahavingthefewest,withlessthanhalfapercentofyoungchildrenhavingaccesstoallthreeadequacies.Between6%(Ethiopia)and39%(Bolivia)ofchildrenareadequateintwoofthethreeadequacies.Whenonlyoneadequacyismet,inmostofourstudycountriesitiscare.TheonlyexceptionisIndonesiawhereachildwhoonlymeetsoneoftheadequacycategoriesismorelikelytobeadequateinfoodthanincareorinenvironmentandhealth.However,inIndonesiathefoodadequacymeasureisbasedonhouseholdlevelinformationonfoodconsumptionandnotchildspecificinformationonwhatkindsoffoodsthechildconsumed.

44

14

23

2

14

2

020

4060

8010

0P

erc

ent

age

of

child

ren

who

me

et c

rite

ria

Source: Author estimates.

Adequacy status in Nepal (2011)

None Food only

Care only Environmental health only

Two of three adequacies All three adequacies

49

14

24

1

12

1

42

14

23

3

15

3

020

4060

8010

0P

erc

ent

age

of c

hild

ren

who

me

et c

rite

ria

Resource-poor Resource-rich

Source: Author estimates.

By wealthAdequacy status in Nepal (2011)

None Food only

Care only Environmental health only

Two of three adequacies All three adequacies

36

Figure9:PrevalenceofAdequacies,recentyear

Althoughchildreninresource‐richhouseholdsareingeneralmorelikelytohavemetallthreeadequacies,insomecountriestheyareaslikelyaschildrenfromresource‐poorhouseholdstomeetnoneoftheadequacies.Between6%(Ethiopia)and48%(Bolivia)ofthechildrenfromthepoorest40%ofhouseholdsareadequateinatleasttwoofthethreemeasures(Figure10a).However,childrenfromthewealthiest60%ofthehouseholds,donotfaremuchbetter,sinceonlybetween7%(Ethiopia)and55%(Bolivia)ofthechildrenareadequateinatleasttwoofthethreemeasures(Figure10b).17Thatis,inthestudycountries,childreninthewealthiesthouseholdsarestillmorelikelytobewithoutaccesstoadequatelevelsinanyofthethreeclustersthantohaveaccesstotwoormoreadequacies.

17TheprevalenceofadequaciesinBolivia,Ethiopia,NepalandIndonesiaareinflatedincomparisontotheothersasthereisnocommunitysanitationinformation.

19%3%

36%

3%

39%

13%24%

6%

33%

3% 14%2%

34%

1%

15%2%

Bangladesh (HK, 2010) Bangladesh (IFPRI, 2011) Bolivia (2008) Cambodia (2010)

Ethiopia (2011) Indonesia (RKD, 2010) Nepal (2011) Peru (2012)

Zimbabwe (2010)

None

Food only

Care only

Environmental health only

Two of three adequacies

All three adequacies

Source: Author estimates. See Table 5 for the specific components used in each country for the adequacy measures.

Recent year

Adequacy status

37

Figure10a:PrevalenceofAdequacies,Resource‐poorhouseholds

Figure10b:PrevalenceofAdequacies,Resource‐richhouseholds

42%

3%41%

3%11% 11%

52%

34%

3%19%

9%

20%5%

38%

9%

36%

10%32%

22%

55%

6%

33%

6%25%

31%6%1%

33%

3%

49%

14%

24%

1%12%1%

21%

7%

32%

40%

53%

7%

29%

11%

Bangladesh (HK, 2010) Bangladesh (IFPRI, 2011) Bolivia (2008) Cambodia (2010)

Ethiopia (2011) Indonesia (RKD, 2010) Nepal (2011) Peru (2012)

Zimbabwe (2010)

None

Food only

Care only

Environmental health only

Two of three adequacies

All three adequacies

Source: Author estimates. See Table 5 for the specific components used in each country for the adequacy measures.

Recent year

Adequacy status - Resource-poor

28%

12%

25%

6%

24%

5% 16%

41%2%

37%

3% 15%

9%

10%

11%39%

15%32%

12%25%

4%

26%

56%

6%

30%

1%7%28%

32%1%3%

33%

4%

42%

14%

23%

3%

15%3%

30%

17%22%

1%

30%

1%

44%

8%18%

9%

17%4%

Bangladesh (HK, 2010) Bangladesh (IFPRI, 2011) Bolivia (2008) Cambodia (2010)

Ethiopia (2011) Indonesia (RKD, 2010) Nepal (2011) Peru (2012)

Zimbabwe (2010)

None

Food only

Care only

Environmental health only

Two of three adequacies

All three adequacies

Source: Author estimates. See Table 5 for the specific components used in each country for the adequacy measures.

Recent year

Adequacy status - Resource-rich

38

4 Synergies

Inthissection,aneffortismadetoderivesomequantitativeestimatesoftheroleofsynergiesassociatedwithhavingsimultaneousaccesstoadequatelevelsinoneormoreoftheclustersoffoodsecurity,childcare,andhealthandenvironment,onchildnutrition.Theanalysisispurelydescriptive,inthatitisaimedatquantifyingthecorrelationbetweenimprovedorhigherheightforageZ‐scoresandsimultaneousaccesstoadequatelevelsinmorethanoneoftheclusters.Toexplorethepotentialsynergiesamongthethreeclustersofunderlyingdeterminantsandnutritionaloutcomes,asimpleregressionmodelisusedtosummarizeinaparsimoniouswaythedifferencesinthemeanheight‐for‐ageamongchildrenwithaccesstoonlyormoreofthethreeofthethreeclustersoftheunderlyingdeterminantsofnutrition.Purposefully,noadditionalcontrolsareusedintheseregressionssinceincludingsuchcontrolsislikelytocreatetheimpressionthataneffortismadetominimizetheinfluenceofconfoundingfactorsintherelationshipbetweenthedependentandindependentvariablesintheregression,apracticecommontoallstudiesaimedatestimatingcausalregressionswithinaneconometricframework

4.1 Model

Thefollowingeconometricspecificationisestimated

+

∗ ∗ ∗ +

∗ ∗ (A1)

whereHAZiistheHeight‐for‐AgeZ‐scoresforthechildi,andAidenotesaccesstothethreeadequacies,foreachchildi.Namely,A1is1whenthehouseholdisadequateinfood(F)andis0otherwise;A2is1whenthehouseholdisadequateinenvironmentandhealth(H)andis0otherwise;and,A3is1whenthehouseholdisadequateincare(C)andis0otherwise.Thesebinaryvariablesareconstructedwithoutanyconsiderationofwhetherthechildhasaccesstoadequatelevelsintheothertwoclusters.ItisalsoimportanttokeepinmindthattherearenoadditionalcontrolvariablesusedintheregressionbecausetheobjectivehereissimplytocomparemeanvaluesinHAZamongchildreninthesedifferentsub‐groupsofchildrendefinedbytheextenttowhichtheyhaveaccesstooneormoreofthepillars.

Inthisspecificationtheconstantterm providesanestimateofthemeanvalueofHAZscoresforchildrenwithoutaccesstoadequatefoodsecurity( =0),adequateenvironmentandhealth( =0),andadequatecare( =0).Thatis,with | 1 0 denotingtheexpected(ormean)valueofheight‐for‐age(outcome),conditionalonhavingadequateaccess(A=1)orinadequateaccess(A=0)toclusterA,theexpectedheight‐for‐ageforwhenthechilddoesnothaveadequateaccesstoanyofthethreeclustersis:18

18Itisalsoassumedthat | , , 0.

39

| 0, 0, 0

Thecoefficients yieldestimatesoftheincreaseinthemeanHAZscoreofchildrenwhenachildhasaccesstoadequatelevelsinoneoftheclustersonly(andnottheothers).Thatis:

| 1, 0, 0

| 0, 1, 0

| 0, 0, 1 ,respectively.

Specifically,thecoefficient yieldsanestimateoftheincreaseinthemeanHAZscoreofchildren(comparedtothemeanHAZscoreofreferencegroupsummarizedbytheconstantterm, )thathaveaccesstoadequatefoodsecurityonly( =1)butdonothaveaccesstoadequateenvironmentandhealth( =0),andnoaccesstoadequatecare( =0).Thecoefficients and haveananalogousinterpretationsforenvironmentandhealthandcare,respectively.

Moreover,thecoefficients yieldestimatesofthesynergiesorcomplementaritiesassociatedwith

havingaccesstoadequatelevelsinmorethanoneoftheclusterofunderlyingdeterminantsofnutrition.Specifically,themeanHAZscoreofchildrenhavingaccesstoadequatefoodsecurity( =1)andadequateenvironmentandhealth( =1)issummarizedbytheexpression

| 1, 1, 0 .

TheexpressionforthemeanvalueofHAZscoresofchildreninhouseholdswithaccesstoadequatefoodsecurityandadequateenvironmentandhealthcanbeconsideredasconsistingofthesumofthreecomponents:thefirstcomponentistheincreaseinHAZscoresassociatedwithchildreninhouseholdswithadequatefoodsecurityonly(i.e., );thesecondcomponent(i.e., )istheincreaseinHAZscoresassociatedwithchildreninhouseholdswithadequateenvironmentandhealthonly,andthethirdcomponent(i.e., )istheincreaseinHAZscoresassociatedwithchildreninhouseholdsthathaveaccesstobothadequatefoodsecurityandadequateenvironmentandhealth.Thusthecoefficient yieldsinformationonwhetherthereareadditional(extra)gains(orlosses)inHAZscoresderivedfromaccesstoadequatefoodonlyoraccesstoadequateenvironmentandhealthonly.Asignificantandpositivevalueofthecoefficient impliessynergiesfromthesimultaneousaccesstoadequatefoodsecurityandadequateenvironmentandhealthintheproductionofchildnutrition.

Alongsimilarlines,themeanHAZofchildrenfromhavingaccesstoadequatefoodsecurity( =1)andadequatecare( =1)oradequateenvironmentandhealth( =1)andcare( =1)aresummarizedbytheexpressions

| 1, 0, 1 .

| 0, 1, 1

40

withthecoefficient summarizingpotentialsynergiesfromsimultaneousaccesstoadequatefoodsecurityandadequatecareandthecoefficient summarizingpotentialsynergiesfromsimultaneousaccesstoadequateenvironmentandhealthandadequatecare.

ThemeanHAZofchildrenfromhavingaccesstoallthreecomponents(i.e.adequatefoodsecurity( =1)andadequateenvironmentandhealth( =1)andadequatecare( =1))isgivenbytheexpression

| 1, 1, 1

withthecoefficient summarizingthepotentialsynergiesfromsimultaneousaccesstoallthreecomponents.

ThemodelemployedabovedoesnotallowforcausalinferencesontheeffectsofhavingaccesstoadequatelevelsinthevariousclustersadequacycomponentsonnutritionnorprovideaformaltestoftheUNICEFframework.Amorerigorouscausalanalysiswouldrequiretheuseofmethodsaimedataddressingtheendogeneitybiasassociatedwiththefactthatmanyoftheclustersthemselvesaretoalargeextentchoicevariables(e.g.suchaschildcarevariables,vaccinations,andvisitsforprenatalcare)aswellastheinclusionofadditionalcontrolvariablesaimedatreducingoreliminatingtheimpactofothercontextualvariableomittedfromtheregression(omittedvariablebias).

Nevertheless,theestimatesfromsuchamodelserveasausefulbenchmarkforpolicyintermsofhighlightingthepotentialgainsthatcouldbeaccomplishedwithhavingsimultaneousaccesstoadequatelevelsoftheotherclusters.Thisspecificationallowsfortheexplorationofthepatternsofcorrelationbetweenthevariousadequacymeasuresandnutritionaloutcomesasmeasuredbyheight‐for‐age.Thatis,themodelestimatesthecorrelationbetweenadequaciesandheight‐for‐ageforeachchildbasedoninformationinonetimeperiod.

Animportantcaveattothemodelisthatinmanyofthecountriesanalyzedonlyasmallfractionofchildrenareadequateinenvironmentandhealth.Inmostcases,thesamplesizesaretoosmalltoyieldreliableestimatesofthesynergieswithenvironmentandhealth.Infact,inEthiopiaandNepalthecommunitylevelsanitationconditionwasdroppedgiventherestrictionitplacedonthenumberofchildrenadequateinenvironmentandhealth.

4.2 Results

Therearesignificantandsizablesynergiesassociatedwithsimultaneousaccesstoadequatelevelsin2ormoreclusters.Table7andFigure11presentasummaryoftheresultsfromestimatingEquationA1.Mostofthesynergycoefficients, , ,and ,associatedwithsimultaneousaccesstoadequatefoodsecurityandadequatecare,oradequateenvironmentandhealthandadequatecare,oradequatefoodssecurityandenvironmentandhealtharepositive.Therefore,thereare

41

additional(extra)gainsinHAZscoresoverandabovethegainsderivedfromhavingaccesstoadequatefoodonlyoraccesstoadequateenvironmentandhealthonlyoraccesstoadequatecareonly.Focusingonthefivecountriesforwhichwehavethemostcompleteinformation—Cambodia,Ethiopia,Nepal,Peru,andZimbabwe—wefindthatthesynergycoefficientsbetweentwoclustersareingeneralpositiveandthatsixoutofthefifteencoefficientsarestatisticallysignificantlygreaterthanzero.Thatis,therearepositivesynergiesfromhavingaccesstoadequatelevelsinmorethanonecluster.Giventhesmallnumberofchildrenwhoareadequateinmorethanonecomponentinmostofthecountries,itmaybethatwithadditionalobservationsthecurrentlystatisticallynon‐significantcoefficientestimateswouldbecomestatisticallysignificantandsuggestpositivesynergiesinadditionalcountries.

Accesstoadequatecareonlyisinmanycasesnegativelyassociatedwithheight‐for‐age.Allthecoefficientestimatesforcarearenon‐positivewithsixoftheeightcoefficientestimatesbeingstatisticallysignificantlynegative(Table7).Thisnegativeassociationisthelikelyconsequenceoftargetinghealthychildcarecampaignstopopulationsatrisk.Evenafteradoptingthepromotedbehaviors(suchasexclusivebreastfeedingandbreastfeedinguntilthechild’ssecondbirthday)thepopulationatriskmaystillnotbeabletoovercometheotherobstaclesinprovidingadequatenutritiontothechildren.Ofthethreemeasures,thecomponentsofadequatecarearemoreeasilymodifiedbythemotherandthefamilyastheyarebehavioral.Ultimatelyadequatefooddependsontheavailabilityofdifferentfoodtypesinthehousehold’scommunityandthefinancialabilityofthehouseholdtopurchasenutritiousfood.Adequateenvironmentandhealthpartlydependsonlargerinvestmentsininfrastructurethatahouseholdmaynothavethefinancialabilitytodo(eveniftheyhavetheknowledgeandwillingness).

Ingeneral,thesynergycoefficientfromhavingaccesstoadequatelevelsinallthreeclustersisnegative.Havinganegativesynergycoefficientinallthreedoesnotimplythatchildrenwhoareadequateinallthreeareshorter,butthatthesynergiesfromtheadequacypairsoverestimatethedifference.Table8andFigure12presentthetotalcoefficientestimateforchildrenineachspecificadequacycategory.ExceptforPeruandBangladesh(IFPRI)thechildrenwhoareadequateinallthreearealsotaller.Infourofthesevencasestheyarestatisticallysignificantlytaller.19Again,giventhesmallnumberofchildrenwhoareadequateinallthreethesecoefficientestimatesarebasedononlyafewobservations.

Synergycoefficientsdonotappeartobesystematicallydifferentforchildreninresource‐poorhouseholdsthanforchildreninresource‐richhouseholds.Thereissomepreliminaryevidencethatinsomecountries(namelyPeru,Bolivia,Nepal,BangladeshandZimbabwe)beingadequateinmorethanoneoftheadequacymeasuresisassociatedwithlargersynergiesinresource‐poorhouseholdsvis‐à‐visresource‐richhouseholds(Annex3).However,inothercases(Ethiopia,Indonesia,andCambodia)thesynergiesarelargerinresource‐richhouseholdsthaninresource‐poorhouseholds.

19InPeru,ifvitaminAsupplementsareremovedfromtheadequateenvironmentalhealthconditionthenthetotalcoefficientestimateforbeingadequateinallthreebecomes0.270anditisstatisticallysignificantatthe99%levelofconfidence.

42

Table 7: Marginal increases (Synergies) in mean HAZ of children under 37 or 24 months 

Adequacy  Coefficien

t 

  Cam

bodia (2010) 

Ethiopia (2011) 

Nep

al (2011) 

Bolivia (2008) 

Peru (2012) 

Zimbabwe (2010) 

Bangladesh (HK, 

2010) 

Bangladesh 

(IFPRI, 2011) 

Indonesia (RKD, 

2010) 

Food only    0.220  1.000***  0.498***  0.140  0.004  ‐0.031  0.055  ‐1.094*  ‐0.187** 

Environment and health only    0.214  0.863***  ‐0.212  0.059  ‐0.210  ‐0.085  0.570***  ‐0.824*  0.720*** 

Care only    0.092  ‐0.255***  ‐0.074  ‐0.570***  ‐0.133**  ‐0.158*  ‐0.121*  ‐0.602***  -0.219

Environment x Food     0.208  ‐0.479  0.671*  0.166  ‐0.478  ‐0.028  ‐0.261  2.107***  ‐0.563** 

Environment x Care    0.525  0.685*  0.951***  0.329**  0.476*  0.418  ‐0.371  1.104**  ‐0.136 

Care x Food    0.172  0.708***  0.331  0.473***  ‐0.090  0.332*  0.241  1.478**  0.240 

Food x Care x Environment    ‐0.557  ‐1.175**  ‐1.701***  ‐0.573**  0.012  ‐0.387  0.415  ‐2.559***  0.511 

Observations    1,354  5,605  1,397  4,311  5,248  2,022  4,994  1,056  6,671 Source: Author calculations. Notes: *** p<0.01, ** p<0.05, * p<0.1 

Table 8: Simultaneous Adequacy in 2 or more clusters and the total effect on mean HAZ of children under 37 or 24 months 

Adeq

uacy 

Coefficient 

  Cam

bodia (2010) 

Ethiopia (2011) 

Nep

al (2011) 

Bolivia (2008) 

Peru (2012) 

Zimbabwe (2010) 

Bangladesh (HK, 

2010) 

Bangladesh (IFPRI, 

2011) 

Indonesia (RKD, 

2010) 

E x F    0.643  1.384***  0.957***  0.365**  ‐0.684**  ‐0.144  0.364*  0.188  ‐0.030 

E x C    0.832  1.293***  0.665**  ‐0.183*  0.133  0.175  0.078  ‐0.323  0.365 

C x F    0.484***  1.453***  0.756***  0.043  ‐0.219***  0.143  0.175*  ‐0.218  ‐0.166** 

F x C x E  

0.875***  1.346***  0.465  0.023  ‐0.419***  0.062  0.528***  ‐0.391  0.365** 

Source: Author calculations. Note: *** p<0.01, ** p<0.05, * p<0.1 

43

Figure11:Synergyeffects

Figure12:Totaleffects

-.5

0.5

11.

52

Coe

ffici

ent

est

ima

te fr

om

Mo

del A

1

BGD, HKBGD, IFPRI

BOLETH

IDNKHM

NPLPER

ZWE

Source: Author estimates. Data for Bangladesh (HK) 2010; Bangladesh (IPRI) 2011; Bolivia 2008; Cambodia 2010; Ethiopia 2011; Nepal 2011; Peru 2012; Zimbabwe 2010

Synergies among adequacies

Environment & Food Environment & Care

Food & Care

-.5

0.5

11.

5C

oeffi

cie

nt e

stim

ate

fro

m M

ode

l A1

BGD, HKBGD, IFPRI

BOLETH

IDNKHM

NPLPER

ZWE

Source: Author estimates. Data for Bangladesh (HK) 2010; Bangladesh (IPRI) 2011; Bolivia 2008; Cambodia 2010; Ethiopia 2011; Nepal 2011; Peru 2012; Zimbabwe 2010

Total effects among adequacies

Environment & Food Environment & Care

Food & Care

44

5 ConcludingRemarksandPolicyConsiderations

ThisstudyprovidedoneofthefirstcomprehensiveinvestigationsofthedataavailabilityanddataconstraintsassociatedwithamoresystematicapplicationoftheUNICEFconceptualframeworkemphasizingtheinterrelationshipamongaccesstoadequatefoodsecurity,environmentandhealth,andchildcarepracticesintheprevalenceofmalnutritionratesamongchildrenindevelopingcountries.UsingdetaileddemographicandhealthsurveydatafromBangladesh,Bolivia,Cambodia,Ethiopia,Indonesia,NepalandZimbabwe,acomparisoniscarriedoutoftheidealvariablessummarizingthevariousdimensionsorcomponentsoffoodsecurity,environmentandhealth,andchildcareagainstthemeasuresavailablefromcurrenthouseholdsurveys.Thesecomparisonsservetohighlightthelimitationsofmostdatasetsandthepotentialgainsassociatedwiththecollectionandavailabilityofadditionalinformation.

Bearinginmindthelimitationsimposedbythedataavailable,thereportalsoprovidedapracticaldiagnosticframeworkofthemaincorrelatesofchildmalnutritionthatcouldbeappliedtoidentifypotential“bindingconstraints”towardstheefforttoreducechildmalnutrition.Specifically,theUNICEFconceptualframeworkis“operationalized”byservingasaguideforinvestigatingtherelationshipbetweentheprevalenceofmalnutritioninthecountryandinadequatelevelsandaccesstothethreepillarssummarizingtheunderlyingcausesofmalnutrition.Next,foreachindicatoravailablefromthesurveyusedincountry,adefinitionof“adequacy”isconstructedusingthresholdsbasedonacceptedinternationalstandards.Thereportaimstoprovidea“helicopterview”oftheextenttowhichnutritionaloutcomes,asmeasuredbyachild’sheight‐for‐ageZscore(HAZ)atanygivenpointintime,aswellasovertime,areassociatedwithinadequatefoodsecurity,inadequateenvironmentandhealth,andinadequatechildcarepractices.Inconsiderationofthecomplexityofthelinkagesbetweentheunderlyingcausesofmalnutritionandtheeconomicsituationofthefamily,theanalysisisalsocarriedoutseparatelyforresource‐rich(top60%)andresource‐poor(bottom40%)householdsbasedonanassetindexconstructedforthatpurpose.

Thereportalsoprovidedsomeofthefirstempiricalevidenceonthesynergiesatworkincombatingmalnutritioninthesetofcountriesusedinthisstudywhenthereissimultaneousaccesstoadequatelevelsintwoormoreofthethreepillarsoftheunderlyingcausesofmalnutrition.Whileintuitivelyappealing,thesynergiesamongthethreeclustersofmalnutritionhavereceivedlittleempiricalvalidation.Therecentemphasisonsector‐specificnutritionsensitiveinterventions(WorldBank,2013)rightlyemphasizesthesynergiesthatcanbeexploitedwithinspecificsectorssuchagriculture,waterandsanitation,orsocialprotection.Theanalysisinthereportunderscoresthepointthatthesuccessofthesesector‐specificnutrition‐sensitiveinitiativesmaybeconstrainedbytheslowprogressintakingadvantageofthesynergiesamongthethreebroadclustersoftheunderlyingdeterminantsofmalnutrition:foodsecurity,childcare,environmentandhealth.Giventhatthesynergiesamongthesethreepillarsarebeyondthescopeofanyspecificsectorsuchasagricultureorsocialprotection,thesimultaneousprogresstowardsinallofthethreepillarsiseithertakenforgrantedorunderemphasized.Asaconsequence,nutritionsensitiveinterventionsinspecificsectorsruntheriskofbeingasectoralpriorityincontextswherethechancesofsignificantsuccessmaybelimitedbecauseofnoorverylowaccesstoimprovedinfrastructurewaterandsanitationfacilities.

45

Overall,theanalysishighlightedthecriticalneedforadditionalinformationonnationalhealthsurveys.Asystematicreviewofthedataavailableinthecurrentnationalhealthsurveysinrelationtothespecificcomponentsofthethreeclustersofunderlyingcausesofmalnutritionconsideredasimportantbythenutritionalscientificcommunityrevealsthatthesesurveyscontaininformationononlyasmallfractionoftheidealmeasuresoffoodsecurity,childcareandenvironmentandhealth.Thecollectionofinformationontheseadditionalvariablescurrentlymissingfromthesesurveyswillenableamorerobustunderstandingoftheextenttowhichchildrenhaveaccesstoadequatecareoradequatefoodsecurityorenvironmentandhealth.Paneldataofyoungchildrenbetweentheagesof6and18monthswouldalsofurtherourunderstandingoftherelativecontributionofthevariouscomponentsofthethreepillars,andhowaccesstodifferentcomponentsofthethreepillarsatdifferentagesofachildaffectsnutritionaloutcomes.Inthesampleofcountriescoveredinthisreportitisfoundthatthemostcompleteinformationiscollectedonthecomponentsofenvironmentandhealth,whilethereismuchlessrobustinformationcollectedonthecomponentsofchildcareandfoodsecurity.AspointedoutbyBarrett(2010)inhisarticleonmeasuringfoodsecurity,recognitionthat“measurementdrivesdiagnosisandresponse”iscriticalforsettingarenewedandhigherstandardintheefforttoreducechildmalnutritionworldwide.

Usingalltheinformationavailable,albeitincomplete,theanalysisrevealedthatprogresstowardsimprovedaccesstoadequatefoodsecurityandtoadequateenvironmentandhealthhasbeenquitelimited.Itappearsthataccesstoadequateenvironmentandhealthremainsthegreatestchallengerelativetotheothertwopillarsoffoodsecurityandcare.Accesstoimprovedsanitationbothinthehouseholdandinthecommunityasawholewerethelowestofalltheothercomponentsofenvironmentandhealth.Inthecountriescoveredinthisreportaccesstoadequateenvironmentandhealthingeneralisnotsurpassing10%ofthepopulationofchildren.Accesstoadequatecarehasimprovedinsomecountriesbymorethan10percentagepointssuggestingthesuccessofcampaignsontheimportanceofbreastfeedingandcomplementaryfeedingwhichcomprisethemeasure.However,thishastobeviewedinperspectiveasthisimprovementisbasedonlyontheoneortwochildcarevariablesthatmaybefoundinthesurveyscollected.

Inspiteofthelimitationsinherentintheavailabledata,thereportpresentedevidenceofimportantsynergiesamongadequatefood,childcareandenvironmentandhealth.Theeconometricanalysiscarriedoutisaimedatsummarizingsimplecorrelationsratherthancausalrelations,nordoesitprovideaformaltestofthevalidityoftheUNICEFconceptualframework.However,itdoesserveasusefulevidencethataccesstoadequatelevelofserviceinmorethanonepillaratthesametimeisassociatedwithbetternutritionaloutcomescomparedtothe“sumofitsparts”.Inotherwords,intheaggregate,thereductionsinchildmalnutritionthatcanbeaccomplishedaregreaterwhensomechildrenareprovidedsimultaneouslywithaccesstoadequatefoodsecurity,care,andenvironmentandhealthcomparedtothereductioninmalnutritionthatcanbeaccomplishedbyprovidingsomeofthechildrenwithadequatehealthenvironmentandanothergroupofchildrenwithadequatefoodsecurityandanothergroupofchildrenwithadequatecare.ThedegreeandtheimportanceofthesynergiesvariesfromcountrytocountrybutingeneralsimultaneousaccesstothethreeclustersisassociatedwithhigherHAZscores.

46

Thereportalsohighlightedtheneedforincreasedawarenessattheoperationalandprojectlevelofthelimitsofexclusivelyfocusingontacklingoneparticulardimensionofinadequacywithoutconsideringthegreatercontext.Forexample,thebestdesignedprogramforimprovedaccesstonutritiousfood,mayyieldminimalreturnsiftheprogrambeneficiariesdonothaveaccesstosanitationorcleanwater.Similarly,healthcampaignstoencouragebreastfeedingorvaccinationsmaychangethesebehaviorstobeinlinewithbestpractices,butwithoutaddressingtheneedsforaccesstoadequatefoodsecurityoradequateenvironmentandhealth,mayresultinverylimitedreductionsinmalnutrition.Moreover,suchcoordinationacrosssectorsmaybenecessarynotonlyatthebroadlevelbutalsowithinsector‐specificprojects.Forexample,emphasisonimprovednutrition‐sensitiveagriculturalinterventions,suchashomegrownvegetablegardensmaybemootifmostdwellingsaresurroundedbyunpennedanimalsroamingfreelyandincreasingthepotentialexposureofachildtoanimalfecalmatter.

47

References

Alderman,H.,Hentschel,J.,andSabates,R.(2003):“WiththeHelpofOne’sNeighbors:ExternalitiesintheProductionofNutritioninPeru”SocialScience&Medicine56(2003)2019–2031

Alderman,H.,Hoogeveen,H.,&Rossi,M.(2005):“ReducingchildmalnutritioninTanzania–combinedeffectsofincomegrowthandprograminterventions,”WorldBankPolicyResearchWorkingPaper3567.TheWorldBank,Washington,DC.

Alderman,H.,Hoddinott,J.,andB.Kinsey(2006):“LongtermConsequencesofEarlyChildhoodMalnutrtion”,OxfordEconomicPapers,58(3):450‐474.

Alderman,H.,M.Garcia(1994):“FoodSecurityandHealthSecurity:ExplainingtheLevelsofNutritionalStatusinPakistan,”EconomicDevelopmentandCulturalChange,42(3):485‐507.

Arimond,M.,MT.Ruel(2004):“DietaryDiversityisAssociatedwithChildNutritionalStatus:Evidencefrom11DemographicandHealthSurveys,”JournalofNutrition,34(10):2579‐85.

Barrera,A.(1990):“TheRoleofMaternalSchoolingandItsInteractionwithPublicHealthProgramsinChildHealthProduction,”JournalofDevelopmentEconomics32:69–91.

Barrett,C.B.(2009):“MeasuringFoodInsecurity”,Science,327,825(2010)

Beck,N.,andJ.N.Katz.(1995):“Whattodo(andnottodo)withtime‐seriescross‐sectiondata.”AmericanPoliticalScienceReview,89(3):634‐47.

_________(2001).“Throwingoutthebabywiththebathwater:AcommentonGreen,Kim,andYoon.”InternationalOrganization,55(2):487‐95.

Behrman,J.R.andWolfeB.L.(1987):“HowDoesMother’sSchoolingAffectFamilyHealth,Nutrition,Medicalcareusage,andHouseholdSanitation?”JournalofEconometrics,36:185‐204.

Behrman,J.R.,andWolfe,B.L.(1984):“Moreevidenceonnutritiondemand:Incomeseemsoverratedandwomen’sschoolingunderemphasized.”JournalofDevelopmentEconomics,14(1),105–128.

Behrman,J.,Deolalikar,A.,(1987):"Willdevelopingcountrynutritionimprovewithincome?AcasestudyforruralsouthIndia,"JournalofPoliticalEconomy,95,492–507.

Behrman,J.andDeolalikar,A.,(1988):"HealthandNutrition”,inHollisCheneryandT.N.Srinivasan(eds),HandbookofDevelopmentEconomics,NorthHolland,Amsterdam.

BellA.,andJones,K.,(2012)“ExplainingFixedEffects:RandomEffectsmodellingofTime‐SeriesCross‐SectionalandPanelData”Unpublishedmanuscript

Borooah,V.K.(2005)“TheHeight‐for‐AgeofIndianChildren,”EconomicsandHumanBiology,3(1),45–65.

Chong,A.,Hentschel,J.,andSaavedra,J.(2003)“BundlingofServicesandHouseholdWelfareinDevelopingCountries:TheCaseofPerú”WorldBankPolicyResearchWorkingPaper3310.TheWorldBank,Washington,DC.

Christiaensen,L.,andAlderman,H.(2004).ChildmalnutritioninEthiopia:Canmaternalknowledgeaugmenttheroleofincome?EconomicDevelopmentandCulturalChange,52(2),287–312.

Coates,J.,A.Swindale,andP.Bilinsky(2007),“HouseholdFoodInsecurityAccessScale(HFIAS)forMeasurementofHouseholdFoodAccess:IndicatorGuide(v.3).”Washington,D.C.:FoodandNutritionTechnicalAssistanceProject,AcademyforEducationalDevelopment.

48

CrookstonBT,PennyME,AlderSC,DickersonTT,MerrillRM,StanfordJB,PorucznikCA,DeardenKA(2010):Childrenwhorecoverfromearlystuntingandchildrenwhoarenotstunteddemonstratesimilarlevelsofcognition.TheJournalofNutrition,140(1):1996–2001.

Deolalikar,A.(2008):Malnutritionandhunger.Copenhagenconsensus2008perspectivepaper.Copenhagen:CopenhagenConsensusCentre.

ElfindriandGourangaLalDasvarma(1996):‘ChildMalnutritioninIndonesia’,BulletinofIndonesianEconomicStudies31(1):97–111.

FAO,(2011):DevelopingAResponseAnalysisFrameworkForFoodSecurityEmergencies,DiscussionPaperFAO.http://www.fao.org/fileadmin/templates/tc/tce/pdf/Response_Analysis_Framework_Discussion_Papers.pdf,accessedonJune6,2013.

Haddad,L.,Alderman,H.,Appleton,S.,Song,L.,&Yohannes,Y.(2003):“ReducingChildMalnutrition:HowFarDoesIncomeGrowthTakeUs?”WorldBankEconomicReview,17(1),107–131.

Hoddinott,J.,Rosegrant,M.,&Torero,M.(2012):“Investmentstoreducehungerandundernutrition.”CopenhagenConsensusChallengePaper,preparedfor2012GlobalCopenhagenConsensusChallenge,CopenhagenConsensusCentre,Copenhagen.

Hoddinott,J.andYohannesY.,(2002):“DietaryDiversityasaHouseholdFoodSecurityIndicator,”FoodandNutritionTechnicalAssistanceProject,AcademyforEducationalDevelopment,Washington,D.C.

HortonS.,ShekarM.,McDonaldC.,MahalA.;KrysteneB.J.(2010):ScalingUpNutrition:WhatWillitCost?.Washington,DC:WorldBank.http://hdl.handle.net/10986/2685accessedonJune6,2013.

Prentice,A.M.,Ward,KA,Folgerg,GR,Jarjou,LM,Moore,SE,Fulford,AJ,andPrentice,A(2013):“Criticalwindowsfornutritionalinterventionsagainststunting,”TheAmericanJournalofClinicalNutrition,97(5),911‐918.

Rajaram,S.,Zottarelli,L.K.,&Sunil,T.S.(2007):”Individual,household,programmeandcommunity:EffectsonchildhoodmalnutritioninruralIndia.”MaternalandChildNutrition,2007(3),129–140.

Riely,F.,Mock,N.,Cogill,B.,Bailey,L.&Kenefick,E.(1999):“FoodSecurityIndicatorsandFrameworkforUseintheMonitoringandEvaluationofFoodAidPrograms”,FoodSecurityandNutritionMonitoring(IMPACT)ProjectfortheU.S.AgencyforInternationalDevelopment,Arlington,VA.http://www.fantaproject.org/downloads/pdfs/fsindctr.PDF,accessedonJune6,2013.

Ruel,M.(2002):“IsDietaryDiversityanIndicatorofFoodSecurityorDietQuality?AReviewofMeasurementIssuesandResearchNeeds,”FCNDDiscussionPaperNo.140,IFPRI,Washington,D.C.

SavetheChildren(2012):“NutritionintheFirst1,000Days:StateoftheWorld’sMothers2012”http://www.savethechildren.org/atf/cf/%7B9def2ebe‐10ae‐432c‐9bd0‐df91d2eba74a%7D/STATEOFTHEWORLDSMOTHERSREPORT2012.PDF

Sahn,D.andAldermanH.(1997):“OntheDeterminationofNutritioninMozambique:TheImportanceofAge‐SpecificEffects,”WorldDevelopment,25(4):577‐588.

49

Schnepf,R.D.(1992):“NutritionalStatusofHouseholds:SurveyEvidenceontheRoleofHouseholdConsumptionBehavior,”WorkingPaperno.23.CornellFoodandNutritionPolicyProgram.

Skoufias,E.(1999):“ParentalEducationandChildNutritioninIndonesia.”BulletinofIndonesianEconomicStudies,Vol.35,No.1,(April),pp.99‐119.

Sokolovic,N,Selvam,S,Srinivasan,K,Thankachan,P,KurpadAV,andThomas,T(2013):“Catch‐upgrowthdoesnotassociatewithcognitivedevelopmentinIndianschool‐agechildren,”EuropeanJournalofClinicalNutrition,68,14‐18.

Strauss,John,andThomasD.,(1995):“EmpiricalModelingofHouseholdandFamilyDecisions”,inJ.BehrmanandT.N.Srinivasan(eds),HandbookofDevelopmentEconomics,3A,NorthHolland,Amsterdam.

Strauss,J.,andThomas,D.(1998):“Health,nutrition,andeconomicdevelopment”.JournalofEconomicLiterature,36(2),766–817.

Swindindale,A.andBilinskyP.,(2006):“HouseholdDietaryDiversityScore(HDDS)forMeasurementofHouseholdFoodAccess:IndicatorGuide(v2),”Washington,D.C.:FoodandNutritionTechnicalAssistanceProject,AcademyforEducationalDevelopment.

Tiwari,S.,SkoufiasE.,SherpaM.,(2013):“Shorter,Cheaper,Quicker,Better:LinkingMeasuresofFoodSecuritytoNutritionalOutcomesinBangladesh,Nepal,Pakistan,Uganda,andTanzania”.PolicyResearchWorkingPaperNo.6584,August2013,TheWorldBank,WashingtonD.C.http://go.worldbank.org/BWBRP91A50

UNICEF(1990):“StrategyforImprovedNutritionofChildrenandWomeninDevelopingCountries”

USAID(2012):“MaternalDietarydiversityandtheimplicationsforchildren’sdietsinthecontextoffoodsecurity.”USAID’sinfantandyoungchildnutritionprojectlink;www.iycn.org

Victora,C.G.,deOnis,M.,Hallal,P.C.,Blössner,M.,&Shrimpton,R.(2010):“Worldwidetimingofgrowthfaltering:revisitingimplicationsforinterventions.”Pediatrics,125(3),473‐480.

Webb,P.,andBlock,S.A.(2004):Nutritioninformationandformalschoolingasinputstochildnutrition.EconomicDevelopmentandCulturalChange,52(4),801–820.

Weismann,D.,Hoddinott,J.Aberman,N.,andRuel,M.,(2006):“ReviewandValidationofDietaryDiversity,FoodFrequency,andOtherProxyIndicatorsofHouseholdFoodInsecurity,”

Wiesmann,D.,Bassett,L.Benson,T.,andHoddinott,J.,(2009):“ValidationoftheWorldFoodProgramme’sFoodConsumptionScoreandAlternativeIndicatorsofhouseholdFoodSecurity,”IFPRIDiscussionPaper00870

WHO(2008)“Indicatorsforassessinginfantandyoungchildfeedingpractices–partI:definition”link

WHO(2011).Guideline:VitaminAsupplementationforinfantsandchildren6‐59monthsofage.Geneva,WorldHealthOrganization.

WHO(2013).Table2SummaryofWHOPositionpapers‐recommendedroutineimmunizationsforchildrenlink

WHOandUNICEF(2006)“Corequestionsondrinking‐waterandsanitationforhouseholdsurveys,”link

WorldBank(2013)."ImprovingNutritionThroughMultisectoralApproaches."Washington,DC,TheWorldBank.

50

WorldFoodProgramme,(2009)EmergencyFoodSecurityAssessmentHandbookWFP,Rome,http://documents.wfp.org/stellent/groups/public/documents/manual_guide_proced/wfp203244.pdf,accessedonJune24,2013.

WorldFoodProgramme(WFP).(1996)RomeDeclarationonWorldFoodSecurity

WorldFoodProgramme(WFP).(2008).“FoodConsumptionAnalysis:CalculationandUseoftheFoodConsumptionScoreinfoodSecurityAnalysis,”TechnicalGuidanceSheet.Rome.

Annex1

51

Annex1:Assetindex

Table A3: Components of asset index used to determine resource‐rich and  ‐poor households 

Asset   Cam

bodia 

Ethiopia 

Nep

al 

Bolivia 

Peru 

Zimbabwe*

 

Bangladesh 

(HK)** 

Bangladesh 

(IFPRI)*** 

Indonesia**** 

Electricity    yes  yes  yes  yes  yes  yes     

Radio  yes  yes  yes      yes  yes1  yes   

TV  yes  yes  yes      yes  yes1  yes   

Computer        yes  yes         

Refrigerator  yes          yes  yes     

Garbage disposal        yes  yes         

Mobile/telephone    yes  yes  yes  yes  yes  yes  yes   

Telephone    yes  yes      yes    yes   

Floor type  yes  yes  yes      yes    yes   

Cooking fuel type  yes  yes  yes  yes  yes  yes    yes   

Roof type        yes  yes  yes    yes   

Wall type        yes  yes  yes    yes   

Kitchen        yes  yes         

Bicycle  yes  yes  yes      yes  yes  yes   

Motorcycle  yes  yes  yes  yes  yes  yes  yes  yes   

Car/truck  yes  yes  yes  yes  yes  yes  yes     

Boat with motor  yes          yes  yes  yes   

Oxcart/horse  yes          yes  yes  yes   Notes: 1 TV and radio are combined 

* Also included source of water supply, sanitation facility, ownership of a watch, land, livestock, cattle, goats, sheep, pig, poultry, having a bank account, domestic servant 

** Also includes solar panel, khat/chawki, almirah, table/chair, watch/clock, rickshaw/van, power tiller, shallow machine, and fishing net. *** Also includes ownership of suitcase, buckets, stove, metal cooking pots, bed/khat/chowki, armoire/cabinet/alna, table/chair, hukka, electric fan, electric iron, audio cassette/cd player, wall clock/watch, jewelry, sewing machine, rickshaw, van, boat, dheki, jata, randa, saw, hammer, patkoa, fishing net, spade, axe, shovel, shabol, daa, khacchar, ass, solar energy panel, electricity generator, IPS, other assets, kaste, nirani, ladder, rake, plough, reaper/sickle, manual sprayer, wheelbarrow, bullock cart, push cart, light machinery, tractor, power tiller, trolley, thresher, fodder cutting machine, swing basket, don, hand tube well, treadle pump, rower pump, low lift pump, shallow tube well, deep tubewell, electric  motor pump, diesel pump, spraying machines, harvester, other heavy machinery, mason equipment, potters chaka, blacksmith's hapor, charka, a servant in household, more rooms than median; the household has access to health center/hospital, bus stop, main road, railway station, local shop/shops, bazaar, nearest town, college, agricultural extension office, post office, bank, brac, Grameen bank, asa.  **** Used per capita expenditures to divide the sample to resource rich (top 60%) and resource poor (bottom 40%). 

Annex2

52

Annex2:Componentsofadequacymeasuresbywealth

2625

3227

4434

6764

7687

8981

3735

6964

3128

6665

7679

3629

6965

7591

802223

5245

1412

5453

6373

8375

2935

4857

5346

3175

5952

5745

0 20 40 60 80 100Percentage of children who meet criteria

ZWE ('10)

PER ('12)

NPL ('11)

KHM ('10)

IDN ('10)

ETH ('11)

BOL ('08)

BGD (IFPRI, 11)

BGD (HK, '10)

Source: Author estimates.Note: In Indonesia dietary diversity based on household (not child specific) information

By wealth

Components of Adequate Food

Resource-poor DD Exclusive Meal frequency Food security

Resource-rich DD Exclusive Meal frequency Food security

53

808586

816363

322776

869177

668576

8797

994540

4741 69

6480

866650

706776

7910010047

4965 75

838549

5151525245

7685

847576

806373

9697

10010048

579496

777959

6459

52

0 20 40 60 80 100Percentage of children who meet criteria

ZWE ('10)

PER ('12)

NPL ('11)

KHM ('10)

IDN ('10)

ETH ('11)

BOL ('08)

BGD (IFPRI, 11)

BGD (HK, '10)

Source: Author estimates.Note: Exlusive breastfeeding for first 6 months; Complementary feedings from 6 months;Early initiation of breastfeeding within 60 mins (100 mins BOL and PER) of birth; Continued breastfeeding until 24 months

By wealth

Components of Adequate Care

Resource-poor Exclusive Early initiation Complementary Breastfeeding

Resource-rich Exclusive Early initiation Complementary Breastfeeding

54

8661

526

7728

8566

6415

7521

7967

259

4524

9077

201

4212

7491

7646

3517

42

1711

9364

8079

6663

7871

9897

2411

5525

0 20 40 60 80 100Percentage of children who meet criteria

ZWE ('10)

PER ('12)

NPL ('11)

KHM ('10)

IDN ('10)

ETH ('11)

BOL ('08)

BGD (IFPRI, 11)

BGD (HK, '10)

Source: Author estimates.

By wealth

Components of Adequate Environment

Resource-poor Sanitation Community Water

Resource-rich Sanitation Community Water

55

6159

6456

7160

518

9792

7776

7076

6141

3225

5958

7057

8376

1815

4941

3024

189

6669

8258

8079

7172

3218

8278

7070

3418

9283

0 20 40 60 80 100Percentage of children who meet criteria

ZWE ('10)

PER ('12)

NPL ('11)

KHM ('10)

IDN ('10)

ETH ('11)

BOL ('08)

BGD (IFPRI, 11)

BGD (HK, '10)

Source: Author estimates.Note: In Helen Keller (HK) Vitamin A supplementation only for 6 months and older

By wealth

Components of Adequate Health

Resource-poor Vaccinations Prenatal Vitamin A

Resource-rich Vaccinations Prenatal Vitamin A

Annex3

56

Annex3:Synergiesbywealthgroup

-4 -2 0 2 4Coefficient estimate from Model A1

ZWE

PER

NPL

KHM

IDN

ETH

BOL

BGD, IFPRI

BGD, HK

Source: Author estimates. Data for Bangladesh (HK) 2010; Bangladesh (IPRI) 2011; Bolivia 2008;Cambodia 2010; Ethiopia 2011; Nepal 2011; Peru 2012; Zimbabwe 2010

By wealth

Synergies among adequacies

Resource-poor Environment & Food Environment & Care Food & Care

Resource-rich Environment & Food Environment & Care Food & Care