亞開行-衡量蒙古乳制品價(jià)值鏈項(xiàng)目的影響：一項(xiàng)基線研究（英）

MEASURINGTHEIMPACTOFADAIRYVALUECHAINPROJECTINMONGOLIA

ABASELINESTUDY

EnkhtuvshinGunchinsurenandChristianAbeleda

NO.1

ADBPRIVATESECTOROPERATIONS

November2023

WORKINGPAPERSERIES

ASIANDEVELOPMENTBANK

ADBPrivateSectorOperationsWorkingPaperSeries

MeasuringtheImpactofaDairyValueChain

ProjectinMongolia:ABaselineStudy

EnkhtuvshinGunchinsurenand

ChristianAbeleda

No.1|November2023

TheADBPrivateSectorOperationsWorking

PaperSeriesfocusesonresearchfindingsfromongoingandcompletedprivatesectorprojects,highlightssignificantdevelopmentimpacts,andsharesinsightsonstrategiesandapproaches

promotingprivatesectordevelopment.

TheviewsexpressedarethoseoftheauthorsanddonotnecessarilyreflecttheviewsandpoliciesofADBoritsBoardofGovernorsorthegovernmentstheyrepresent.

EnkhtuvshinGunchinsuren,PhD,isaneconomistwho

consultsforgovernmentorganizationsandprivate

companiesinMongolia,andanumberofinternational

developmentorganizationsinthefieldsofeconomic

research,economicandfinancialanalysis,monitoring

andevaluation,andprojectevaluation.HefoundedandmanagesaconsultingfirmNewFrontiersLLCbased

inMongolia.

ChristianAbeledaworksasanassociateeconomics

officeratADB’sPrivateSectorTransactionDivision.

Hesupportsprivatesectorprojects,particularlyinthe

agribusinesssector,bypreparingeconomicevaluations

andenhancingdevelopmentimpactsperspectives.His

researchfocusisonmonitoringandevaluationofimpactsofagriculturaldevelopmentprojects.

ASIANDEVELOPMENTBANK

CreativeCommonsAttribution3.0IGOlicense(CCBY3.0IGO)

?2023AsianDevelopmentBank

6ADBAvenue,MandaluyongCity,1550MetroManila,Philippines

Tel+63286324444;Fax+63286362444

Somerightsreserved.Publishedin2023.

PublicationStockNo.WPS230516-2

DOI:

/10.22617/WPS230516-2

TheviewsexpressedinthispublicationarethoseoftheauthorsanddonotnecessarilyreflecttheviewsandpoliciesoftheAsianDevelopmentBank(ADB)oritsBoardofGovernorsorthegovernmentstheyrepresent.

ADBdoesnotguaranteetheaccuracyofthedataincludedinthispublicationandacceptsnoresponsibilityforany

consequenceoftheiruse.ThementionofspecificcompaniesorproductsofmanufacturersdoesnotimplythattheyareendorsedorrecommendedbyADBinpreferencetoothersofasimilarnaturethatarenotmentioned.

Bymakinganydesignationoforreferencetoaparticularterritoryorgeographicarea,orbyusingtheterm“country”inthispublication,ADBdoesnotintendtomakeanyjudgmentsastothelegalorotherstatusofanyterritoryorarea.

ThispublicationisavailableundertheCreativeCommonsAttribution3.0IGOlicense(CCBY3.0IGO)

/licenses/by/3.0/igo/.Byusingthecontentofthispublication

,youagreetobeboundbythetermsofthislicense.Forattribution,translations,adaptations,andpermissions,pleasereadtheprovisionsandtermsofuseat

/terms-use#openaccess

.

ThisCClicensedoesnotapplytonon-ADBcopyrightmaterialsinthispublication.Ifthematerialisattributed

toanothersource,pleasecontactthecopyrightownerorpublisherofthatsourceforpermissiontoreproduceit.ADBcannotbeheldliableforanyclaimsthatariseasaresultofyouruseofthematerial.

Pleasecontactpubsmarketing@ifyouhavequestionsorcommentswithrespecttocontent,orifyouwishtoobtaincopyrightpermissionforyourintendedusethatdoesnotfallwithintheseterms,orforpermissiontousetheADBlogo.

CorrigendatoADBpublicationsmaybefoundat

/publications/corrigenda

.

Note:

Inthispublication,“$”referstoUnitedStatesdollars.

TheADBPrivateSectorOperationsWorkingPaperSeriesservesasaplatformforsharingknowledge,insights,andresearchfindingsonprivatesectorresearchanddevelopment.ItaimstostrengthentheroleoftheAsianDevelopmentBank

(ADB)asaknowledgeproviderwhodeliversqualityanalysesthatarerelevanttodevelopingmembercountries,andsupportADB’sprivatesectorengagement.

TheseriesismaintainedbythePrivateSectorOperationsDepartment(PSOD)andismeanttoincreasethevisibilityofandsteercommunicationonthedevelopmentresultsofPSOD’sresearchwork.Sincepapersinthisseriesareintendedforquickandeasydissemination,thecontentmayormaynotbefullyeditedandmaylaterbemodifiedforfinalpublication.

CONTENTS

TABLESANDFIGURES

ACKNOWLEDGMENTS

ABBREVIATIONS

CURRENCYEQUIVALENTS

EXECUTIVESUMMARY

I.INTRODUCTION

II.THEPROJECT

III.METHODOLOGY

A.TheoreticalFramework

B.ProposedImpactEvaluationDesign

C.SelectionofRespondents

D.LimitationsoftheStudy

IV.BASELINESURVEYRESULTS

A.GeographicalDistributionoftheRespondents

B.DemographicCharacteristicsoftheRespondents

C.SocioeconomicConditionsoftheRespondents

D.DairyFarmingandProductivity

V.PRELIMINARYEVALUATIONOFTHEPROJECT

VI.RECOMMENDATIONSFORPOST-PROJECTEVALUATION

APPENDIX:BASELINESURVEYQUESTIONNAIRE

REFERENCES

iv

v

vi

vi

vii

1

1

3

3

4

5

6

6

6

7

9

12

16

17

18

26

TABLESANDFIGURES

TABLES

1MilkCollectionPoints2

2NumberofHouseholdsSurveyedintheTreatmentandComparisonGroups

6

3DemographicCharacteristicsoftheRespondents,2019

7

4EducationalAttainmentoftheHouseholdHead,2019

8

5SchoolAttendanceofChildren,2019

8

6PaymenttoSocialInsuranceContribution,2019

9

7SourceofIncomeoftheRespondents,2019

9

8NumberofLivestockOwnedbytheRespondents,2019

10

9NumberofLivestockOwnedbytheRespondents,inStandardizedLivestockUnit,2019

11

10NumberofMilkingCowsOwnedbytheRespondent,2019

12

11TypeofFarmingActivityAdoptedbytheRespondents,2019

13

12VolumeofMilkCollectedperHouseholdperday,2019

14

13Respondent’sMilkProductivity,2019

14

14AverageIncomeperCapitafromSalesofMilkandDairyProducts,2019

15

15AverageIncomefromAnimalHusbandryperStandardizedLivestockUnit,2019

15

16AverageMonthlyHouseholdIncomeforTreatmentandComparisonGroup,

16

2019and2020

FIGURES

1TheoryofChange

3

2DistributionofLivestockHoldings,forTreatmentandComparisonGroup,2019

10

3LivestockCompositionofTreatmentandComparisonGroups,2019

11

ACKNOWLEDGMENTS

EnkhtuvshinGunchinsuren(managingpartnerofNewFrontiersLLC)preparedthepublication,withinputsfromChristianAbeleda(associateeconomicsofficer,PrivateSectorTransactionSupportDivision,PrivateSectorOperationsDepartment).

Theauthorsaregratefultothefollowingindividualswhoextendedinvaluablesupportandprovidedassistanceintheconductofthestudyandtheproductionofthisreport:

FromMilkoLimitedLiabilityCompany

(i)MunkhzolbooBayaraa(headoftheFinanceDepartment)andGanbatBattulga(headoftheOperationsDepartment)ofMilkofortheirsupportandassistanceduringtheconductofthestudyinMongolia;and

(ii)ManagersofMilkocollectionpoints:NyamtsetsegBaasanjav(Bayangol,SelengeProvince),ErdenetsetsegSainkhuu(Zuunburen,SelengeProvince),EnkhtuvshinShagdarsuren(Khushaat,SelengeProvince),BolortsetsegEnkh-Amgalan(Darkhan,Darkhan-UulProvince),andShirmentogooGanchuluun(Mungunmorit,TuvProvince)fortheirassistanceincoordinatingtheparticipationofbeneficiariesduringtheconductofthestudy.

FromtheAsianDevelopmentBank

(i)ManfredKiefer,principaleconomist,PrivateSectorOperationsDepartment,forprovidingoverallguidanceinthestudyandthewritingofthereport;

(ii)TakashiYamano,principaleconomistfromtheEconomicResearchandDevelopmentImpactDepartment,forhispeerreviewofthereport;

(iii)JohnJuhyunJeong,currentsenioradvisortothePresident,whoservedastheseniorinvestmentspecialistfortheproject,forprovidingsupportduringtheconceptualizationofthestudy;

(iv)ArvinYana,communicationsconsultant,foreditingandcoordinatingthepublicationof

thereport;and

(v)JiaOrila,operationsassistant,PrivateSectorTransactionSupportDivision,PrivateSectorOperationsDepartment,foradministrativesupport.

ABBREVIATIONS

ADB

AsianDevelopmentBank

DiD

difference-in-differences

IE

impactevaluation

km

kilometer

MNT

Mongoliantogrog

SLU

standardizedlivestockunit

CURRENCYEQUIVALENTS

(asof15October2019)

Currencyunit

–

togrog(MNT)

MNT1.00

=

$0.000375

$1.00

=

MNT2,668.5

EXECUTIVESUMMARY

In2019,AsianDevelopmentBank(ADB)approvedaprojectprovidingMilkoLimitedLiabilityCompanywitha$7.5millionloanfacilitytosupporttheexpansionofthecompany’sdairyprocessing,andrawmilkandfruitprocurementcapacities.Thisworkingpaperpresentstheresultsofthebaselineevaluationsurveyconductedtodocumentcurrentconditionsofprojectbeneficiaries—primarilydairyfarmers—beforetheprojectwasintroduced.Theresultsofthebaselinestudy,togetherwithanex-postsurveytobeundertakenattheprojectcompletionin2023,willbeusedfortheimpactevaluation(IE)study.

TheIEstudywillemploythedifference-in-differences(DiD)analysistocomparethebeneficiarygroup(alsoknownasthetreatmentgroup)againstthecomparisongroup(alsocalledthecontrolgroup)beforeandaftertheprojectintervention.Thedataforthisbaselinesurvey,whichwilllaterbeanalyzedagainsttheresultsoftheex-postsurveytoassesstheproject’simpact,camefromahouseholdsurveycarriedoutthroughface-to-faceinterviewswithrespondentsusingastructuredquestionnaire.

Atotalof1,256householdsfromsixsoums(districts)weresurveyedforthestudy.Thetreatmentgrouphad556householdsandthecomparisongrouphad700households.Thesociodemographicandeconomiccharacteristicsofthetwogroupswerecompared.ThesurveyfoundthatherderhouseholdsinMongoliakeeparangeoflivestockasasourceofincome.Thisincludessheep,goats,cattle,andhorses,whichareutilizedformeat,milk,wool,andfarmlabor.Theirmainsourceofincomeisthesaleofmilkandmilk-basedproducts,aswellasliveanimals,includingmeatandcashmere.In2019,thetotalaveragemonthlyincomewasMNT983,000forthetreatmentgroupandMNT926,000forthecomparisongroup.

In2019,bothgroupsmaintainedasimilarnumberofcattle(usedforbothmeatandmilk)withanaverageof41forthetreatmentgroupand42forthecomparisongroup.Intermsofthenumberofmilkingcows,thetreatmentgrouphadanaverageof13milkingcows,whereasthecomparisongrouphadanaverageof12.Householdswithupto19cowsmadeuproughly82%ofthesampleinbothdistributions.Mostherdersusedacombinationoftraditionalandintensivefarmingapproachesindairyproduction.Herderstypicallymilkedtheircowstwicedailyinthemorningandintheevening.Cows’milkproductionpeaksinJunewhenthepastureisgreen,andcontinuesuntilAugustandSeptember.Thetreatmentgroupcollected49.4litersofmilkperdayonaverage,whilethecomparisongroupcollected36.9liters.Thetreatmentgroup’smilkproductivity,definedasmilkvolumepercow,was4.2liters,whilethecomparisongroup’smilkproductivitywas2.9liters.

Anendlinesurveywillbeconductedaftertheprojectimplementation.Itwassuggestedthattheex-postsurveybeconductedayearafterprojectcompletion(i.e.,2024)andshouldadoptthequestionnaireusedduringthebaselinesurvey.

I.INTRODUCTION

Mongolia’slivestockpercapitaisamongthehighestintheworld,1withover67millionheadoflivestockcomparedtoahumanpopulationof3.3million.OverathirdofMongoliansliveinruralareas.In2020,about61%ofhouseholdsdependedonpastoralactivitiesandanimalhusbandryfortheirlivelihood.2

Despiteitsagriculturaloutputbeingmostlypastoralactivitiesandanimalhusbandry,domesticdairyproductionandprocessingcapacityinMongoliaremainslimited.In2018,892millionlitersofrawcowandsheepmilkwereproduced,ofwhichonly10%wasprocessedthroughdairyfactories(footnote2).Whiledairyprocessorsarewillingtobuyfromsmallholderproducers,forfarmers,gettingtheirmerchandisetoandfromcollectionpointsontimeisachallengeduetoimmensedistances,Mongolia’stopography,andtheperishabilityoftheirproducts.Herdersloseprospectiveincomeandlivelihoodopportunitiesonaccountofsupplychainchallengesandthelackofnearbyprocessingfacilities.

InNovember2019,theAsianDevelopmentBank(ADB)approvedaproject(henceforthreferredtoas“theMilkoproject”),extendinga$7.5millionloanfacilitytotheMilkoLimitedLiabilityCompany(Milko),oneoftheleadingdairycompaniesinMongolia,tosupporttheexpansionofitsdairyprocessingandrawmilkandfruitprocurementcapacities.MilkomanagesitsentirevaluechainatitsfactoryinUlaanbaatar,includingsourcingrawmilkandberriesdirectlyfromherdersandfarmers,andprocessingdairyproducts.

ADB’sinvestmentconcentratesontwoareas:(i)theconstructionofsevennewmilkcollectioncentersinruralareastoincreaseMilko’sdirectrawmilkprocurementfromherdersfrom1,500tonsperannum(tpa)to3,200tpa;and(ii)theinstallationoftwonewlow-temperaturedryingmachinestoincreaseMilko’spowderedmilkproductioncapacity.

Thisreportprovidesandanalyzesthebaselinedataofthebeneficiaries,whichwillbeusedasinputsfortheIEstudythatwillbecarriedoutattheprojectcompletion.

II.THEPROJECT

Milkohadfourmilkcollectionpointsin2019andbecauseoftheinvestment,sixadditionalmilkcollectionpointsinsixsoums(districts)locatedinfouraimags(provincialadministrativeunits)wereaddedin2020.Milkowasplanningtoestablishthreemorecollectioninthefuture(Table1).

1

2

FoodandAgricultureOrganization.2021.

FAOStatistics

.

MongoliaNationalStatisticalOffice.2021.

https://1212.mn/en

.

MeasuringtheImpactofaDairyValueChainProjectinMongolia3

III.METHODOLOGY

A.TheoreticalFramework

Impactevaluation(IE)isaquantitativeassessmentofthecausaleffectsofagivenintervention.Itaimstoquantifytheeffectthattheprojecthasonitsbeneficiariesthroughtheprocessofreachingtheprojectoutcome.IEiscarriedoutbycomparingtheendlineoutcomesagainstbaselineconditionscontrollingforeffectsunrelatedtotheproject.Figure1presentsthetheoryofchangeoftheproject.

Figure1:TheoryofChange

Implementation

Inputs

ADBInvestment

MilkoResources

Outputs

Outcomes

Qualityoflife

Impact

Increasedhousehold

incomeandimproved

well-beingamong

dairyherders

IntermediateOutputs

Immediate

Outputs

ImmediateOutcome

IntermediateOutcome

Constructionofnewcollection

centers

Increased

productivityof

herdersandberry

collectors

Increasedcapacity

toprocureand

storemilk

Regularsourceofincomeforherders

andcollectors

Newmachines

installedinthe

processing

facility

Increasedcapacity

toprocessmilkand

milk-based

products

Highernutritionalintakeby

consumers

Reliablefood

supply;amore

competitive

production/

supplychain

Improvedquality

andhighervalue

addedatthe

production.

milkcollectorsand

Morejobscreatedin

thesupplychain(for

factoryworkers)

Increased

productionofmilk

products

Source:NewFrontiersLLC.

Inputs.TheinputsoftheMilkoprojectincludealoanfromADB($7.5million)andthecompany’sinternallygeneratedcash,amountingto$2.1million.

Outputs.TheMilkoprojecthastwomainoutputs:(i)expansionofdairyandfruitsupplychaininruralMongolia;and(ii)expansionofthecompany’sdairyprocessingcapacityinUlaanbaatar.Thefirstoutputpertainstotheconstructionofsevennewmilkcollectioncentersintendedtoincreasethecompany’sprocurementcapacity,whilethesecondoutputreferstotheinstallationofnewmachinestorampupproductionofmilk-basedproducts.

Outcomes.TheMilkoprojectaimstoimprovethelivelihoodsofsmallholdermilksuppliersandberrycollectors.AmongtheimmediateresultsexpectedaretheincreasedproductivityofherdersandberrycollectorsduetotheincreaseddemandfromMilko.TheMilkoprojectalsohopestocreatemoreopportunitiesforpeoplethatmaybehiredforMilko’svaluechain(i.e.,factoryworkersandmilkcollectors).This,inturn,willimproveincomes,employment,andgeneralwell-beingofthecommunitiessurroundingthecollectioncenters,andraiseproductivityforMilkoitself.TheMilkoproject,byprovidingnutritious,fresh,local,andhigh-qualityrawmaterialstothefactory,alsohasthepotentialtoimprovethehealthandwell-beingofMilko’sconsumers.

4ADBPrivateSectorOperationsWorkingPaperSeriesNo.1

Impact.SmallholdermilksupplierhouseholdswillfeelthemainimpactoftheMilkoproject,astheywilldirectlybenefitfromthenewcollectionpointsestablishedwithprojectsupport.Ultimately,theprojectwouldimprovelivelihoodconditions,increasingmilkproductivityandhouseholdincome.

B.ProposedImpactEvaluationDesign

ThestudycomprisesabaselinesurveywhichwillcollectinformationtoimplementtheselectedIEmethod.Experimentalevaluationdesigns,suchastherandomizedcontroltrial,arenotappropriatefortheprojectasbeneficiaryherderscannotbechosenrandomly.Othertechniqueswouldemploystatisticalmethodstoestablishacomparisongrouptoallowavalidanddefensibleevaluationofprojectimpacts.Thedifference-in-differences(DiD)methodwasconsideredmostappropriate.DiDaimstocomparethetreatmentwithacomparisongroup,beforeandaftertheprojectintervention.3

TheIEmustbedesignedcarefullytoattributeonlytheeffectsofinterventionandtoexcludeeffectsofotherfactors.Theattributioniscarriedoutbyestablishingacounterfactualgroupthatisindependentoftheintervention.Theinterventioncanexplainthedifferencesbetweenthechangesintheimpactindicatorsforthetreatmentgroupandcomparisongroup,providedthatthecontrolandtreatmentgroupsaresufficientlysimilarandselectionbiasisaddressed.

Intheprojectcase,theprojectimpactwillbereflectedinthechangesinimpactindicatorsforherderswhosupplymilktoMilko’scollectionpointsandherderswhodonotsupplymilkbetweenthebaselineandendlineyears.Forinstance,takingtheaverageincomefromsalesofmilkanddairyproducts(denotedasDIPC)astheimpactindicatortobemeasured,theDiDimpactestimatewouldbe:

where,

DIPC

DiD

DIPCe

t

DIPCe

t

DIPCe

c

DIPCb

c

DIPCDiD=(DIPCte–DIPCtb)–(DIPCce–DIPCcb)

DiDprojectimpactestimateforincomefromsaleofmilkanddairyproductsTheaverageincomefromthesaleofmilkanddairyproductsforherderhouseholdsinthetreatmentgroupattheendlineyear

Theaverageincomefromthesaleofmilkanddairyproductsforherderhouseholdsinthetreatmentgroupatthebaselineyear

Theaverageincomefromthesaleofmilkanddairyproductsforherderhouseholdsinthecomparisongroupattheendlineyear

Theaverageincomefromthesaleofmilkanddairyproductsforherderhouseholdsinthecomparisongroupatthebaselineyear

Thechangesintheaverageincomefromthesaleofmilkanddairyproductsinthecomparisongroupwillbenotedtodescribewhatwouldhavesimilalryhappenedtothetreatmentgroupiftheprojectdidnotexist.TheDiDapproacheliminatestheimpactoffactorsnotrelatedtotheproject,suchasdrought,harshwinters,andchangesinpricesofanimalby-productsthataffectbothtreatmentandcomparisongroups.

Theseexternalfactorsneedtobecommontobothcontrolandcomparisongroups.Researchersusedthisassumptionwhentheychosecomparisongroupsfromwithinthesoumswithcollectionpoints,orfromneighboringsoums.TheDiDmethodalsoremovesanydifferenceintheindicatorsbetweentreatmentandcomparisongroupsthatwerepresentatthebaseline,suchasdifferencesinmilkoutputspercowduetodifferentcowbreeds.

3White,H.andD.Raitzer.2018.ImpactEvaluationofDevelopmentInterventions:APracticalGuide.ADB.Manila.

MeasuringtheImpactofaDairyValueChainProjectinMongolia5

C.SelectionofRespondents

Researchersinterviewedtwosetsofrespondents:onetreatmentandonecomparisongroup.ThetreatmentgroupconsistedofherderhouseholdswhoaresupplyingmilktoMilko’snewly-establishedcollectionpoints.Atotalof556householdswereincludedunderthisgroup.

Thecomparisongroup’sidealcompositionshouldconsistofhouseholdswithidenticalcharacteristicsasthetreatmentgroup,exceptthattheydonotsupplytoMilko.Householdsforthecomparisongroupwereselectedfromsoumswithcollectionpoints,andneighboringsoumswithsimilargeographicalconditions,livestockcomposition,andmarketaccess.Researcherssetthefollowingrulesinchoosinghouseholdsforthecomparisongroup:

(i)InthesoumswithonlyMilkocollectionpoints.ResearcherschosehouseholdsthatdidnotsupplymilktoMilkoandotherdairyprocessors,andareunlikelytosupplyinthefutureduetolocation(forexample,beingseparatedbyariverfromthecollectionpointsorbeingremotelylocated).Ifthesoumdoesnothaveasufficientnumberofherderhouseholds,researcherslookedinneighboringsoums.

(ii)Intheneighboringsoums.Researchersfirstselectedaneighboringsoumwithnodairycollectionpoints,eitherbyMilkooranyotherdairyprocessors.ThesesoumsshouldbelocatedascloselyaspossibletosoumswithMilkocollectionpointsandshouldhaveasimilargeographicalconditionandlivestockcomposition.4Researchersthenchosehouseholdsthatowncowsandarelocatedclosetomainroadsthatconnecttothesoumcenterandbhags.Insum,researchersselectedhouseholdsthatareverylikelytobemilksuppliershadtherebeenacollectionpointintheirsoum.

Theneighboringsoumswereincludedbecauseofthefollowingreasons:

(i)InsoumswithMilkocollectionpoints,householdsthatarenotcurrentlysupplyingtoMilkomaybeconvincedtostartsupplyinginthefuture.Inotherwords,theriskofcontaminationforthetreatmentandcomparisonsampleishighiftoomanyhouseholdsareselectedforthecomparisongroupfromwithinthesoumswithcollectionpoints.

(ii)OtherdairyprocessorsalsocollectmilkfromalmostallsoumswithMilko’snewcollectionpoints.Therefore,thesetofhouseholdsthatdonotsupplymilktoanydairyprocessorisverysmall.

(iii)Mosthouseholdsthatarenotsupplyingmilktoanydairyprocessor(whileatleasttwodairyprocessorsarecollectingfromtheirsoums)arelikelytohavefundamentallydifferentcharacteristicsfromthemilksupplierhouseholds,suchasnothavingenoughcowsthatcanproducemilkintheamountbeyondtheirownhouseholdconsumption.Includingthesehouseholdsinthecomparisongroupwillresultinselectionbias.

Basedontheaboveconditions,researcherssurveyed1,198householdsforthecomparisongroup.Ofthese,around700householdswereselectedaspartofthefinalcomparisongroupbasedontheirsimilaritieswithhouseholdsinthetreatmentgroup.Thesamplesizeforthecomparisongroupissetatalargerlevelthanthetreatmentsampleduetopotentialattrition,toensurethattheendlinesurveywillhaveasufficientnumberofhouseholdsinthecomparisongroup.

4Livestockcompositionforsoumsdifferdependingongeographicconditions.Forexample,insoumswithforestandalpinepastures,herderstendtohavemorecattleandlesssheepandgoatthanherdersinsoumswithsteppeandGobigeography.

6ADBPrivateSectorOperationsWorkingPaperSeriesNo.1

D.LimitationsoftheStudy

Researchersencounteredthefollowingissuesduringthebaselinestudy:

(i)Theprojectidentifiedtwomaingroupsofbeneficiaries:(a)herderssupplyingmilktothenewcollectionpoints;and(b)berrycollectors.Thestudyonlycoversherderhouseholdsduetothedifficultyinidentifyingtheberrycollectorbeneficiaries.Themarketforberriesisnearlyperfectlycompetitive,andanyonecanpickwildberries.Apersonwhocollectedwildblueberriesinanygivenyearmaynotcollecttheminthesucceedingyearduetoawiderangeoffactors.

(ii)Theprojecteffectivelystartedin2020,andbaselinesshouldhavebeenestablishedfor2019.However,asthebaselinesurveywasconductedinspring2021(April),householdswereaskedtoanswerquestionsabouttheirconditions2yearspriortothesurvey.Thiscouldhavesomenegativeimpactontheaccuracyoftheanswersinvolvingspecificnumberssuchasincome,livestockholdings,andvolumeofmilk.Toimproveaccuracy,householdswereaskedabouttheir2020conditions,andresearcherscomparedthetwodatasetsforeachyear.Correctionsweremadewhennecessary.

(iii)Milkohasnoformalcontractswithanyherderhouseholds.Thelistofbeneficiarieswascompiledbytheconsultantbasedonthepartiallistprovidedbythecompany,andalistprovidedb