EDGET project final evaluation report FINAL1 - SNV · 2018-08-31 · The project also includes two cross-cutting objectives: a) To promote women and youth entrepreneurship and b)

EVALUATIONREPORT

1

EDGETProjectFinalEvaluation

March2018

EVALUATIONREPORT

2

TableofContents

TableofContents.......................................................................................................................................2Abbreviations.............................................................................................................................................8ExecutiveSummary..................................................................................................................................10

Background..........................................................................................................................................10StrategicObjectives..............................................................................................................................11

Cross-cuttingStrategies....................................................................................................................21Conclusion............................................................................................................................................21

1.Introduction.........................................................................................................................................231.2.OverviewofEDGETproject...........................................................................................................231.2.Thisevaluation..............................................................................................................................24

2.Methodology........................................................................................................................................252.1.Overviewofmethodology.............................................................................................................25

Quantitativecomponent..................................................................................................................25Qualitativecomponent....................................................................................................................29Keylimitationstotheevaluationmethodology...............................................................................30

3.Context.................................................................................................................................................324.Evaluationfindings...............................................................................................................................34

4.1Overalldairyvaluechainsetup......................................................................................................34Socio-economicprofileoffarmersatbaselineandendline.............................................................35Gendereddivisionofrolesindairyactivities...................................................................................38

4.2 Strategic Objective 1: To enhance sustainable dairy production and productivity, input supplyandrelatedservices......................................................................................................................39

Theextensionsystem.......................................................................................................................39Forageandforageseedproduction.................................................................................................51AgroInputDealers...........................................................................................................................61Householdadoptionofinputsandpractices...................................................................................69

4.3StrategicObjective2:Toincreaseprocessingandmarketingofdairyproducts...........................87

EVALUATIONREPORT

3

Householdmilkprocessing,consumptionandsale.........................................................................87CooperativeswithDairyProcessingUnits........................................................................................95

4.4 Strategic Objective 3: To contribute to development of institutions and to dairy sector-wideinitiatives.....................................................................................................................................104

Institutionalsupporttoworedalivestockoffices...........................................................................105Engagementwithregional/nationalforageseedsuppliers...........................................................109

4.5StrategicObjective4:Todevelopaknowledgebaseondairyrelatedissues..............................1104.6StrategicObjective5:Toimprovenutritionalstatusofchildrenthroughdairyconsumption.....1124.7Cross-cuttingStrategies................................................................................................................117

Womenandyouthentrepreneurship............................................................................................118Climatechange...............................................................................................................................123

5.Inclusive,sustainabledairyvaluechaindevelopmentinEthiopia.....................................................1245.1KeyActorsDrivingBetterMilkProductionfromProducers.........................................................1255.2IntegrationandInterdependenceofcomponents.......................................................................1265.3AClearerUnderstandingofMilkMarketsandtheirdynamics....................................................1295.5PlanningandStrategy...................................................................................................................1305.6Metricsandaversatileandeffectivelearningsystem.................................................................131

EVALUATIONREPORT

4

ListofTablesTable1Evaluationfocusandevaluationobjectives................................................................................24Table2Stagesinsamplingprocedure......................................................................................................26Table3modulesandkeytopicscoveredforthehouseholdsurvey........................................................27Table4Overviewofquantitativeandqualitativedatacollected............................................................28Table 5 Socio-demographic background data for comparison and intervention group at base- and

endline.......................................................................................................................................36Table6Socio-economicdataforcomparisonandinterventiongroupatbase-andendline..................37Table7Cattleanddairycowownershipcomparisonandinterventiongroupatbase-andendline.......38Table8Responsibilitiesfordifferentdairyrelatedactivitiesinthehouseholds.....................................38Table9Extensionsystemactivities:achievementofoutputtargets.......................................................41Table10Receiptoftrainings,adviceandfollow-upsupport...................................................................43Table11Groupmemberships..................................................................................................................44Table12Percentofhouseholdsreceivingvarioustypesofinput............................................................45Table13Qualitativefindingsregardingtheextensionsystem................................................................46Table14Provisionofinputsonforagedevelopmentandforageseedmultiplication............................52Table 15 Land ownership and allocation for base- and endline data grouped by intervention and

comparisongroup......................................................................................................................53Table16Overviewofadoptionoffeedingpractices...............................................................................55Table17Qualitativefindingsontheforageproductionandseedsystem...............................................58Table18AgroInputDealers.....................................................................................................................62Table19QualitativefindingsonAgroInputDealers................................................................................65Table20Overviewofadoptionoffeedingpractices...............................................................................70Table21Overviewofequipmentusedbycomparisonandinterventiongroupsforstorage,milkingand

transport....................................................................................................................................72Table22Overviewofadoptionofhygienicmilkingpractices..................................................................74Table23Overviewofadoptionofanimalpractices.................................................................................77Table24Overviewofadoptionofcalfmanagementpractices................................................................78Table25Overviewofadoptionofhousingandmanuremanagementpractices....................................81Table26Overviewofadoptionofclimatesmartpractices.....................................................................83

EVALUATIONREPORT

5

Table27:Milkproduction........................................................................................................................85Table28%Householdsproducingdifferenttypesofdairyproducts......................................................88Table29Averagequantityofdifferenttypesofprocesseddairyproductsproducedbyhousehold......89Table30%ofhouseholdsinvolvedinsaleofrawmilkandprocesseddairyproducts............................90Table31Quantityofdairyproductssold.................................................................................................90Table32Incomeearnedfromsaleofmilk,dairyproductsanddairyrelatedactivities..........................91Table33:Incomeearnedfromsaleofmilk,dairyproductsanddairyrelatedactivities........................92Table34Costsofdairyproduction.Pleasenotefairlysmallnforsomeofthevariablespresented......92Table35NetincomeinBirrfromdairyrelatedactivities........................................................................94Table36AchievementofprojectoutputsforcooperativesandDPUs....................................................96Table37Qualitativefindingsoncoopsanddairyprocessingunits.........................................................99Table38Institutionalsupporttoworedalivestockoffices-achievementofoutputtargets................106Table39Qualitativefindingsontheinstitutionalstrengtheningcomponent(woredaservices)..........106Table40Achievementofprojectoutputsforcreatingaknowledgebaseondairyrelatedissues.......111Table41Achievementofprojectoutputsfornutrition.........................................................................113Table42Achievementofprojectoutputsforwomenandyouthentrepreneurship.............................118Table43Qualitativefindingsonwomen'sempowerment....................................................................121Table44Characteristicsandapproachesfordifferentworedacategories...........................................131

ListofFiguresFigure1DifferenceinDifferenceAnalysis...............................................................................................26Figure2OverviewofthedairyvaluechaininEthiopiaOverviewofthehouseholdsurvey.....................35Figure3TimelineofSO1implementation................................................................................................39Figure4Actormapoftheextensioncomponent....................................................................................41Figure5Actormapoftheforageproductionandforageseedsystem(requiresreview).......................52Figure6Bar-plotwith intervalestimateofpopulationproportions (CI95%)forproportionof farmers

allocationlandtoforageproductionnowandfouryearsback.................................................55Figure7Bar-plotwithintervalestimateofpopulationproportions(CI95%)foradoptionofforageseed

productionpracticesincomparisonandinterventiongroup....................................................56Figure8ActormapoftheAgroInputDealersystem...............................................................................62

EVALUATIONREPORT

6

Figure 9 Bar-plot with interval estimate of population proportions (CI 95%) for adoption of feedingpracticesincomparisonandinterventiongroup.......................................................................71

Figure 10 Histogram showing the count of years since adoption of practices for which we foundsignificantdifferencesbetweencomparisonand interventiongroup(interventiongrouponlydisplayedhere)..........................................................................................................................72

Figure 11 Bar-plot with interval estimate of population proportions (CI 95%) for usage of milkingequipmentincomparisonandinterventiongroup...................................................................73

Figure12Bar-plotwith intervalestimateofpopulationproportions (CI95%) foradoptionofhygienicpracticesincomparisonandinterventiongroup.......................................................................75

Figure 13 Histogram showing the count of years since adoption of practices for which we foundsignificantdifferencesbetweencomparisonandinterventiongroup......................................76

Figure 14 Bar-plotwith interval estimate of population proportions (CI 95%) for adoption of animalhealthpracticesincomparisonandinterventiongroup...........................................................77

Figure 15 Bar-plot with interval estimate of population proportions (CI 95%) for adoption of calfmanagementpracticesincomparisonandinterventiongroup................................................79


Figure17Bar-plotwith intervalestimateofpopulationproportions (CI95%) foradoptionofhousingandmanurepracticesincomparisonandinterventiongroup..................................................81


Figure19Bar-plotwith interval estimateof populationproportions (CI 95%) for adoptionof climatesmartpracticesincomparisonandinterventiongroup............................................................83

Figure 20Median year since adoptionof practices for respondents for comparison and interventiongroup.Only showingpractices inwhichwe found significantdifferences. Themedian is thevalueseparatingthehigherhalfofadatasamplefromthelowerhalf....................................84

Figure 21 Differences between groups and over time for milk production per household. The 95%confidenceintervalsdepicttherangeofthemeanaveragemilkproductioninthepopulation...................................................................................................................................................86

Figure22TimelineforimplementationofSO2........................................................................................87Figure23Differencesbetweengroupsandover timeforannualnet incomeperhousehold.The95%

confidenceintervalsdepicttherangeofthemeanaveragenetincomeinthepopulation.....95Figure24ActormapoftheDPUcooperativeandMCCcomponent.......................................................96Figure25Actormapoftheinstitutionalstrengtheningcomponent(woredaservices)........................105

EVALUATIONREPORT

7

Figure26TimelineofimplementationforSO3......................................................................................106Figure27TimelineforimplementationofSO4......................................................................................110Figure28TimelineforimplementationofSO5......................................................................................112Figure29OverviewofdatafromtheTIPcomparinguptakeofrecommendedpracticesbymothersas

reportedduring follow-upvisits.Please see the limitationsofpossible conclusions from thequantitativepartofthestudy..................................................................................................115

Figure 30 Proportion of respondents by comparison and intervention group that mentioned dairyproductionrelatedchallenges(*p<0.1,**p<0.05).................................................................128

EVALUATIONREPORT

8

AbbreviationsAgID AgroInputDealer

AI Artificialinsemination

B2B BusinesstoBusiness

BOFED BureauofFinanceandEconomicDevelopment(atregionallevel)

CI Confidenceinterval

DA DevelopmentAgent

DFEG DairyFarmerExtensionGroup

DID DifferenceinDifference

DPU DairyProcessingUnit

EDGET EnhancingDairySectorGrowthinEthiopia

EM Effectivemicroorganism

FGD FocusGroupDiscussion

FTC FarmerTrainingCentre

GDP GrossDomesticProduct

HEW HealthExtensionWorker

HH Household

IYCF Indicatorsforassessinginfantandyoungchildfeedingpractices

KDDC KebeleDairyDevelopmentCommittee

kg kilogram

KII KeyInformantInterview

M&E MonitoringandEvaluation

MCC MilkcollectionCentre

MFI Micro-FinanceInstitution

MIYCN Maternal,infantandyoungchildrennutrition

MOLF MinistryofLivestockandFisheries

MTS MilkTransportationSystem

NSA NutritionSensitiveAgriculture

SARI SouthAgriculturalResearchInstitute

SBCC SocialandBehaviorChangeCommunication

SNV-DEP DairyExtensionPromoter(woredalevel)

SNV-ZDCM ZonalDairyCommunityMobilizer

EVALUATIONREPORT

9

SO Strategicobjective

TIP TrialofImprovedPractices

TOR TermsofReference

TOT TrainingofTrainers

W-MDDS MinimumDietaryDiversityforWomen

WCA WoredaCooperativePromotionAgencies

WLO WoredaLivestockandFisheriesResourceDevelopmentOffices

EVALUATIONREPORT

10

ExecutiveSummary

Background

BackgroundtotheEDGETproject

Agriculturecontributes35.8%toeconomicGDPinEthiopia,withinwhich,thedairysectorcontributes12-16%.TheEthiopiangovernment’sgoal is todoubledomesticmilkproductionbetween2015-2020toreducethedependencyondairyimports.In2015/6,11.33millionmilkingcowsinEthiopiaproducedatotalof3.06billionlitresofmilk.SNV’SEnhancingDairySectorGrowthinEthiopia(EDGET,2013-17)projectpromotes inclusivedevelopmentof thedairy sector.Working in close collaborationwith thenewlyformedMinistryofLivestockandFisheries(MOLF),theworkhasbeenfinancedbytheEmbassyof the Kingdomof theNetherlands. InOromiya, Amhara and SNNPR, the project aimed at doublingincomesfor65,000smallholderdairyfarmersin10zones,51woredasand353kebelesbytheendof2017. EDGET project interventions included extension services, input systems, dairy marketdevelopmentandinstitutionalstrengthening“toimprovehouseholdincomeandthenutritionalstatusofchildrenthroughincreaseddairyproductionandenhanceddairyprocessing&marketing”.

Specificobjectivesoftheprojectinclude:

1. Toenhancesustainabledairyproduction,productivity,inputsupplyandservices;2. Toincreaseprocessingandmarketingofdairyproducts;3. Tocontributetodevelopmentofinstitutionsandtodairysector-wideinitiatives;4. Todevelopaknowledgebaseondairyrelatedissuesand;5. Toimprovenutritionalstatusofchildrenthroughdairyconsumption.

The project also includes two cross-cutting objectives: a) To promote women and youthentrepreneurshipandb)topromoteclimatesmartpractices.

BackgroundtotheEvaluation

The EDGET project board commissioned ALINe to undertake a final and independent evaluationbetweenDec2017andMarch2018toassesstheperformanceandapproachofEDGET.Theevaluationcaptured the project outputs, outcomes and impacts and assessed strategies and approaches tostrengthenthedairyvaluechaininEthiopia.Theevaluationreflectstherelevance,effectiveness,andsustainabilityofinterventionsandtheiroutcomeswithinthedairyvaluechain,theirassociatedactorsandtheextenttowhichbenefitsresultedfromtheproject(e.g.extension,agroinputdealers,foragesystem, etc.). Insights from the evaluation provide recommendations for future inclusive dairydevelopmentinterventionsonalargerscale.

ALINeadoptedamixedmethodsevaluationcomprising:

• Aquantitativecomponentconsistingofanendlinehouseholdsurveycovering12interventionand5comparisonworedas,equivalentto432and218householdsrespectively.Inadditionto

EVALUATIONREPORT

11

the comparison of intervention and comparison groups at the endline, this was to becomparedtoabaselinedatasettobeprovidedbySNV,usingthesamecomparisonworedas.

• Qualitative case studies of key actors in the dairy value chain in 5 project woredas: Dangla(Amhara), Machakel (Amhara), Lemu Bilbilo (Oromia), Wuchale (Oromia) and Aleta Wondo(SNNPR).Woredaswereselectedbasedontheirperformance in termsofextensionservices,cooperatives with DPUs and Agro-Input Dealers (assessed by the EDGET team) to enablelearningacrossdifferentcontexts.ThequalitativeworkusedacombinationofKIIsandFGDs.

• Secondarydataanalysis, includingrecords forAgro-InputDealers,DairyProcessingUnitsandWoredaLivestockOffices,projectdocuments(includingstrategydocuments,progressreports,studiesandprojectM&Edata)andwiderliteratureonthedairysectorwerealsoreviewed.

• Analysis of the relevance, effectiveness and sustainability of the project components andsynthesisofevidenceandlearningwasexpectedtohelpinformanyfuturerefinements.

Keylimitationsoftheevaluationinclude:

• The quantitative and qualitative field work components of the evaluation were carried outundertimeconstraintsduringDecember2017,duetotheunavailabilityofprojectstaffbeyondthisperiod.Thisreducedthetimeavailablefortestingandrefinementoftoolsleadingtosomeshortcomings.

• The endline questionnaire differed from the baseline in a number of questions, limiting thecomparabilityofthedatasets.

• Baseline data was found to be very unsatisfactory. It had very significant credibility andusabilityissueswithsignificanterrorsforcingtheevaluationteamtointerrogateit,reorganiseit, clean it, recalculate it and address missing values. Comparisons with baseline shouldthereforebeinterpretedwithsomecaution.

• Due to the tight timeframe, some qualitative tools were only finalised after data collectionstarted. As a result, data collection in Yirga Chefe, Enemay and Kuyu utilised penultimateversionsofthetools.

StrategicObjectives

SO1 To enhance sustainable dairy production, productivity, input supply andservices;

StrategicObjective1 focuseson increasingmilkproduction throughstrengthened input systemsandextension, leading to theadoptionof improveddairymanagementpractices. The strategicobjectivewas subdivided into three main components: (1) the extension system; (2) the forage productionsystem; and (3) the agro-input dealer network. This section presents findings related to eachcomponent,followedbytheoverallchangeobservedintheadoptionofpractices.

Extensionsystem

EVALUATIONREPORT

12

EDGET’s support on strengthening the extension system comprises a) strengthening the capacity ofgovernmentextensionprovidersat theworeda(WoredaLivestockOffices)andkebele (DevelopmentAgent)levels,b)establishingandstrengtheningDairyFarmerExtensionGroups(DFEGs)asafarmer-to-farmerextensionmodelandc)developinganddistributingextensionmaterialsandotherkey inputs(calffeed,MTSandforageseed).

Developingcapacityofextensionproviders(WLO,DA)

By the end of 2017, the EDGET project had provided training to 1,476 public extensionproviders/officersatvariouslevels,butwithaparticularfocusonWLOsandDAs.Overall,thecapacitydevelopmentofpublicextensionproviderswasfoundtobehighlyvaluableandrelevant.Recipientsofthe training, includingDAsandWLOs, frequently reportedgains in knowledge related to theoveralldairydevelopmentapproachandonspecifictechnicaltopics.TheWLOsappreciatedtheadvantagesoftheprojectextensionapproach.

EDGET’sapproachtobuildingthecapacityofthegovernmentextensionsystemhasstrongpotentialtobesustainable.DAsarealong-term,paid,skilledworkforcethatcanplayacriticalroleinstrengtheningdairydevelopmentbeyondthelifeoftheproject.TheknowledgeacquiredbyDAsandtheiraccesstotrainingmaterials,increasestheirabilitytoprovideongoingextensionadvice.However,someriskstosustainabilitypersist,including;(1)highDAturnover,excessiveworkloadsandlowlevelsofmotivation;(2)SomeDAshavenotfullyembracedtheirdairydevelopmentroles;(3)inanumberofcases,WLOsandDAs depend significantly on SNVDEPs to be effective (i.e. DEPs play a direct role in training orfollowingupwithDFEGleadersandmembers,particularlywherethecapacityofDAsisweak).

EstablishingandstrengtheningDFEGs

DFEGs provide a viable mechanism for DAs to reach a larger number of farmers with improvedknowledge.68%ofinterventionfarmersreportedbeingamemberofaDFEG.However,theirrelevancedependsonthedemandfornewinformationbymembers,andthecapacityandmotivationofleaderstoplaytheirenvisionedroles.Overall, theDFEGmodelappearstohaveworkedwell.Theevaluatorsconsidertheapproachofdrawingonfarmers(notexistingmodelfarmers),asDFEGleadfarmerstobepositive,allowingmorefarmerstoplayaroleandovercomingentrenchedpowerrelationsassociatedwith the existing model farmers. While there are clearly examples of DFEGs functioning well, theoverall sustainability faces some significant risks: (1) DFEG leaders depend on project support formotivation;(2)WLOsandDAsfacebudgetaryconstraintsinprovidingcontinuedsupporttoDFEGs;(3)thereare limited incentives forDFEG leaders toplay their role (little formal recognitionof their roleamongstmembersandnoobservedfinancialreturns).

TrainingActivitiesandinputs

The project appears to have led to improved coverage in the provision of dairy extension/advisoryservices.68%interventionfarmersparticipatedinat leastonedairy-relatedtrainingorexposurevisitactivity, vs. 11% of comparison farmers (female andmale headed households). 47% of interventionfarmers received advice and follow-up support compared to 6% of comparison (both female/male

EVALUATIONREPORT

13

headed households) and 78% intervention farmers cited SNV/DEP as the key source for advice andpublicserviceprovidersin21%ofthecases.

Intermsofinputs,theEDGETprojectprovidedhouseholdswithforageseed,calffeedandMTS.64.4%of intervention farmers reported receiving MTS, compared to 0.6% comparison farmers. 33% ofintervention farmers reported receiving forage seed (13.9% comparison group) and 32% reportedreceivingcalffeed(8%comparisongroup).Similarinputprovisionactivitiesarecertainlycarriedoutinatleastsomecomparisonworedas,butatalowercoverage.

EDGET’s training materials were perceived to be useful and relevant especially illustrated andtranslated versions which increase their relevance to respondents - i.e. Amharic/Amhara and AfanOromo/Oromia.

Feedandforagesolutions

TheEDGETprojectsupplieddairyfarmerswithforageseedsandsupportedthemtoproducedifferenttypes of forage, in addition, to adopt improved feeding techniques, and use supplementaryconcentrate or other products to improve their feed e.g. urea or effective micro-organism (EM)treatment.TheEDGETprojectalsopromotedvariousmodelsofdecentralisedforageseedproductionthroughFTCsandfarmergroups.SNVreportsthatfarmersareabletosavemoneypreviouslyusedforbuyinghayandadditionalfeedsupplements;andgeneratenewincomebysellingforageseeds/splitstofarmers.

DuetoEDGET,thereisgrowingandrelativelywidespreadrecognitionthatfeedingappropriatetypesofforageincreasesthequantityandqualityofmilkproduced.Therearesuccessfulcasesofforageseeddevelopment through FTCs and farmers themselves. In the case of FTCs, institutional, financial andmanpowerconstraints limittheapproachtoscale -andraisestheriskthatdemandoutstripssupply.Where farmers are multiplying, exchanging and selling seed, either individually or through seedmultiplicationgroups,thisiscontributingtotheavailabilityofforageseed.However,forthesefarmersaccesstoqualityseedisessential forthemtoreplenishfreshforageseedstocksovermultipleyears.Thus,while significantprogresshasbeenmadeand themodelof farmer/groupbasedmultiplicationappears to be a successful model of addressing green forage supply a larger scale solution will berequiredtocreateasustainablesupplyofforageseed.

Agro-InputMarketingandSupplySystems

EDGET supported the establishment and development of 50 Agro Input Dealers through trainings,guidelines,BusinesstoBusiness(B2B)networking(AgIDs,nationalandregionaldairyinputsuppliers),andmicrofinanceinstitutions(MFIs).TheAgIDsprovideavenuesforgettingqualityinputs–particularlyvarious typesofconcentrateand improved feed– to farmers.Furthermore,byrouting thesupplyofkeyprojectinputs(e.g.calffeed)andequipment(e.g.theMTS)throughtheAgIDs,EDGEThashelpedAgIDstoestablishsustainablenetworksanddistributionchannelsforthebenefitoffarmers.

• MilkTransportationSystem(MTS):TheMTS(locallyreferredtoas‘Mazzican’)isahigh-qualityfoodgradeplasticcontainerwithlid,filterandmeasurementgauge-toimprovethehygieniccollection and transportation of milk for farmers. By 2017, a total of 95,000 MTS were

EVALUATIONREPORT

14

distributed(95%ofthenewtarget).FarmerssayMTSqualitycomeswithahigherprice (e.g.comparedtosimplebucketswhichcostsandweighsless).

• Calffeedsupplementstrategy:EDGETsoughttopromotethepracticeoffeedingcalveswithspecialised calf-feed to improve growth of the calf and reduce the time to fertility. EDGETprojectextensionstaff,DAsandAgIDsweretrainedonassessingconditionsofcalves,eartagapplications, and other topics to identify eligible calves for supplementary feed.Householdsreceived vouchers for feed supplementswhich reinforced the establishment of relationshipsbetweendairy farmersand theproject-supportedAgIDs in theworeda.By2017,EDGEThadsupportedthedistributionof14,176quintalsofcalffeedto14,683households(30%oforiginaltarget). The measurement of calf growth by DAs and DEPs illustrated the benefits ofsupplementary calf feed to farmers, changing their perceptions of the value of proper calffeeding andmanagement practices. EDGET also found that the age at which female calveswerereadyfortheir1stAIservicehadreducedfrom24-36monthsto14-18months (EDGETAnnualReport2016).

AgIDs rate highly the support from EDGET, particularly the B2B linkages providing access to betterdealsandfurtherbusinessdevelopment.Overall,theAgIDsfillgapsinthemarketbyprovidingbetterquality feed (than the traders) at amore affordable price (than high end feed businesses) and areincrementally adding more product lines (e.g. forage seed, milk collections). AgIDs are a promisingdistributionchannelfordairyrelatedinputswiththeirbusinessesgrowing(volumeofgoods,expandingcustomerbaseandprofitability).Theyexpectgrowingdemandfortheir inputs inthefuture.Quality,price, variety and availability of inputs provided do indicate some areas of weakness in the AgIDs’abilitytomeetthedemandforsufficientquantity,qualityanddiversityoffeed.

Householdadoptionofinputsandpractices

Theendlinestudygathereddataonatotalof34practicesrelatedtoimprovedforageandanimalfeed,milkingandmilktransportation,animalhealth,calfmanagement,housingandmanuremanagement,climatesmartpractices.Adoptionrateswerefoundtobesignificantlyhigherfor16ofthesepracticesintheinterventiongroupthaninthecomparisongroup,withrespondentsmostlikelytoreporthavingadoptedthesepracticeswithintheprojectperiod.

Keyresultsinclude:

• Atendline,47.2% interventiongroupfarmersare involved in forageproductioncomparedto40.0%ofcomparisongroupfarmers.Notably,however,interventionfarmersareconsiderablymorelikelytobegrowingmorethanonevariety(32.4%)thancomparisonfarmers(20.0%)atthe endline and compared to intervention farmers at the baseline (6.5%) and, when askedaboutperceivedchangesinforageproduction,moreinterventionfarmersreportedanincreaseinthelast4years(52.6%)vs.comparison(36.8%).

• Intermsofforageseedproduction,19%interventionfarmersareengagedinseedproduction,comparedto3.2%comparison.9%interventionfarmersreportedpracticingfarmertofarmerseedexchange,whilenocomparisonfarmersdid.

EVALUATIONREPORT

15

• 27.9%interventionfarmersreportedusingsupplementarycalffeed,comparedto9.3%inthecomparisongroup.Farmersusingcalffeedreportedverypositivebenefits(67.9%acrossbothgroups). 78% intervention farmers reported planning a continuation of using supplementarycalffeed,comparedto66.7%inthecomparisongroup.

• Comparison farmers reported no income from the sale of forage seed, improved forage ornaturalgrass/pasture.Only11interventiongrouphouseholdsdid.Theaverageincomeearnedfromsellingforageseed(n=1),improvedforage(n=5)ornaturalgrass/pasture(n=5)forthese11households,is2951.5Birr.

• Intervention households were more likely than comparison households to report that theyprepare their own improved feed (26% vs 12%) and to vary feeding depending on lactation(35%vs20%)

• 67%ofinterventionhouseholdsreportedusingtheMTSformilkingand47%reportedusingitfortransportationofmilk

• Theadoptionofhygienicmilkingpractices improvedover theprojectperiod for interventionand comparison groups. At the endline, intervention householdsweremore likely to reportcleaninghandsbefore(63%vs42%)andafter(55%vs46%)milkingandtoreportcleaningthemilking area (77% vs 63%). Intervention group farmerswere also found to performhygienicmilkingpracticestoagreaterextentaftereverymilkingthancomparisongroupfarmers.

• Animal health practices showed very similar adoption rates for both intervention andcomparison households although intervention households were more likely to report usingantibioticsthancomparisongrouphouseholds.

• Intermsofcalfmanagement,interventionhouseholdsweresignificantlymorelikelytoreporthaving adopted improved practices than comparison households. For example, 53% ofhouseholdsintheinterventiongroupallowedthecalftosucklethemother,comparedto34%inthecomparisongroup.Interventionhouseholdswerealsomuchmorelikelytoapplyeartagsandconductregularheartgirthmeasurements(33%vs4%).

• Interventionhouseholdweremorelikelytoprovideadequateventilationandlightingforcows(53%vs39%)andadequatestorageofmanureforcropapplication(34%vs25%).

• The studydidnot detect statistically significant differences in the adoptionof climate smartpractices, suchas theuseofbiogas,enriching livestock feedwithagriculturalby-productsorusingmanuretofertilisethefield.

Milkproduction

Byendline, interventionandcomparisonworedaswerefoundtohaveanaveragemilkproductionof953and1068litresrespectively.Overall,milkproductionincreasedbetweenbaselineandendlineforbothgroups.While the increasewas larger for thecomparisongroup, thedifference in the increasewasnotfoundtobestatisticallysignificant.Moreover,issueswiththebaselinedata,limitthevalidityof thiscomparison.Perceptualdataonchanges inmilkproductionover the lastyears, revealed that38%ofcomparisonfarmersand47%ofinterventionfarmersreportedanincreaseinmilkproduction.Themainreasons for increaseswerebirthofcalves (88%vs.92%)andpurchaseofanimals (12.2vs.

EVALUATIONREPORT

16

4.4%).Keepinginmindthelimitationsofthedata,evidencefromtheevaluationdoessuggestthattheprojecthascontributedtoincreasedmilkproduction.

SO2ProcessingandmarketingofdairyProducts

Strategicobjective2focusesonincreasingtheprocessingandmarketingofdairyproductsbothatthehouseholdlevelandwithinthedairyvaluechain.Itaimstoenablehouseholdstoearnhigherincomeseitherbysellingtheirmilktoamoreremunerativemarketorproducingprocessed/value-addeddairyproductsthatcanbesoldatapremium.

Householdmilkprocessingandsale

Milkprocessing

Baselineandendlinesurveysgathereddataonhouseholdlevelmilkprocessing,focusinginparticularonbutter,cottagecheese,souredmilkandthequantitiesproducedandsoldofeach.Nostatisticallysignificant difference was found in the proportion of households involved in processing milk ininterventionandcomparisonvillages(82%and86%respectively).However,overall,theproportionofintervention group households producing all three products increased between the baseline andendlinethoughsubstantiallyonlyforbutterandcottagecheese.Intermsofperceivedchanges,51%oftheinterventiongroupfarmersreportedanincreaseinprocessingofmilkproducts,comparedto39%ofthecomparisongroup.Thecomparisongroupwere10%morelikelytoreportthatproductionhadstayedthesame.

Saleofrawmilkandmilkproducts

Accordingtotheendlinesurvey,moreinterventionhouseholds(32.4%)reportedhavingsoldmilkthancomparison households (21.5%). Individuals (>40%), followed by cooperatives (>21%) and traders(>20%)werethemainbuyerstowhomrawmilkwassoldbybothinterventionandcomparisongroups.Comparison householdsweremarginallymore likely to sell to a private company than interventionhouseholds (15% vs. 5%). A much smaller difference was found between the two groups for theproportion selling to cooperatives.However, of those selling to cooperatives in intervention groups,themajority(around60%)reportedsellingtheirmilktotheDPUcooperativessupportedbyEDGET.

Nosignificantdifferenceswerefoundintermsofthevolumeofrawmilksoldbytheinterventionandcomparisongroups(1524.4litresvs.1505.3litresonaverage)andtheaveragepricesobtainedduringboththefastingandnon-fastingseasonswerealsofoundtohavelittlevariation,generallybeingintherangeof 10.9 to 11.4 Birr per litre. Butterwas themost frequently sold product (comparison n=89,intervention n=154), followed by cottage cheese (comparison n=24, intervention n=30).Who theseproductsweresoldto,didnotvarysignificantlybetweencomparisonandinterventiongroups.Overall,butterwassoldatapriceof130Birrperkg,cottagecheeseat46Birrperkgandsouredmilkat13Birrperkg.

Incomefromthesaleofmilkanddairyderivedproducts

EVALUATIONREPORT

17

The total gross income earned from dairy related activities, has an average of 10,120Birr/household/year earned in comparison woredas and 9,553 Birr per/household/year earned ininterventionworedasatendline.Thisdifferenceisnotstatisticallysignificant.

Costsofproduction

Overall,thecostsofproductionincreasedforalmosthalfoftherespondentsinbothcomparisonandinterventiongroups.41.6%ofhouseholdsinthecomparisongroupreportedanincreaseascomparedto 50.0% in the intervention group. Where people reported increased costs, significant differenceswerenotfoundinthereasonsforthechangegiven.Themainreasonsforincreasesweremorecows(40%vs.36%)andbuyingbetterqualityfeeds(55%versus57.3%).

Netincomefromdairy

Average household net income from dairy at the endline was found to be in the region of 6,220Birr/year to 6,500 Birr/year across intervention and comparison groups (no statistically significantdifference). However, a significant difference was found in the net income between baseline andendline forbothgroups.Theaveragecosts incurred in thecomparisongroup increasedconsiderablyfrom 1,191 Birr per household to 3,595 Birr per household, but had not changed so much for theinterventiongroup(2,394to3,325Birr)overtime.

Average household net income in the intervention group increased from 792 to 6,221 Birr perhousehold, equivalent to a total increase of 7.8 times. For the comparison group, net income wasfound to have increased from254 to 6,525 Birr, equivalent to a total increase of 25.7 times.Whileproblemswiththebaselinedatalimittheaccuracyoftheresults,thefiguresdonotleadustoquestionthe notion that there have been significant increases in income from dairy related activities. It is,however,morechallengingtoattributethisveryclearlytotheEDGETproject.

CooperativeswithDairyProcessingUnits

EDGET’s approach to developing output markets for milk and milk products focuses on theestablishmentofDairy ProcessingUnits (DPUs) - at theworeda (mainly) and kebele (in some cases)levels. Dairy Processing Units are facilities that are attached to a cooperative and managed by adedicated management committee. Dairy farmers in the woreda can become members of thecooperative,whethertheyaremembersofaDFEGornot.EDGETprovidestrainingtothemanagementcommittees(onmanagement,bookkeeping,hygienicmilkproduction,milkqualitytesting,marketing,etc.)andequipmentformilkcollection,storage,testingandprocessing.EDGETalsoprovidestrainingstotheworedacooperativeagencyandtheworedalivestockofficetoorientthemontheDPU’sandgettheirsupportinkeytechnical,legalandoperationalmatters.

WherethecatchmentoftheDPUsarelarge,EDGEThaspromotedtheestablishmentofdecentralisedMilkCollectionCentres(MCCs)tofacilitatetheaggregationofmilkfromindividualdairyfarmerstothecooperative.CooperativeswithDPUsselleitherrawmilkorprocessedmilkproductstoprivatesectororinstitutionalbuyers,includingothermilkcooperatives/milkunionsandlargerscalemilkprocessors.EDGETprojectand theWoredaCooperativeAgenciesplaya role in facilitating linkagesbetween the

EVALUATIONREPORT

18

DPUsandtheseotheragencies.Farmersaretypicallypaidforthemilktheyprovideonamonthlyortwo-weeklybasisandinsomecasesalsoreceiveannualorbi-annualdividends.

Overall, the number of establishedMCCs andDPUswas 86 (96%of revised target)with small-scaletechnologysupport for70of them(76%ofrevisedtarget).44DPUs(83%ofrevisedtarget) receivedbusiness linkagessupport.Bytheprojectend,atotalof3,198dairyhouseholds(20%oftarget)werelinkedwiththeformalmilkmarketthroughtheDPUsandMCCs.EDGETconductedmarketstudies inWoredas in which DPUs and MCCs were to be supported with capacity development, storage andprocessing technology and market linkage support. A number of delays were faced in theestablishment of DPUs, including the slow process of membership mobilization, registration as acooperative, and fulfilling the necessary prerequisite for a dairy processing unit (e.g. securing thebuilding,equipment,etc.).Atotalof22DPUshadreceivedprocessingequipmentandrelevanttrainingby2016.SixDPUsreceivedsupportonbusiness linkages. In2017,34MilkCollectionCentres (MCCs)andfurther21DPUswereestablished.

The evaluation found that DPUs address – or have the potential to address – key gaps in existingoutput markets for milk and milk products for quality standard buyers who can purchase fromsmallholder dairy farmers. Overall SNV’s support in establishing and developing DPUs has beeneffective,butmarketdevelopment isnot linearandcontextual factorsareachallenge.Delays in theprovisionofequipmentlimitedtheextenttowhichthedairycooperativessupportedwithDPUshavebeenabletofunctionasintendedduringtheprojectperiod.Indeed,DPUsarenotallfunctioningwell.Somewereevenfoundtohavestoppedcollectionatthetimeoftheevaluation.DPUsfacedifficultyinestablishingmarketlinkageslimitingtheirabilitytopurchaseandsell-onthemilkproducedbyfarmersand;managementissuesthatunderminetheirfunctionalityandoperations.DespitetheseDPUshaveincreasedinterest indairyfarmingandfacilitatedrecognitionoftheimportanceofmilkquality.Theircontributiontoincreasedincomesfordairyfarmersisstillunclearandthesustainabilityuncertain.

SO3DevelopmentofDairyInstitutionsandDairySectorWideInitiatives

StrategicObjective3engageswithselecteddairyvaluechainactorsattheregionalandnationallevelsaswell aswithworeda livestockoffices.Preliminarydiscussionswith relevant institutions (e.g. SARCandzonalagriculturalbureausinSNNPRandAmhara)wereheldin2014toexploreopportunitiesandneedsforcapacitystrengthening.Aneedsassessmentconductedin2014ledtotheprioritisationof(1)institutional support to woreda livestock offices; and (2) engagement with regional/national forageseedproducersandmultipliers.

Institutionalsupporttoworedalivestockoffices(WLO)

EDGET'sinstitutionalsupporttoWLOsfocusonaddressingkeyconstraintstotheprovisionofArtificialInseminationservices.EDGETprovidedmotorcycles,largeandsmallliquidnitrogenflasksforstorageattheworedalivestockofficeandfortransportationbymotorbikeaswellasotherAIrelatedequipmentTechnical trainings and capacity developmentwere alsoprovided toAI technicians to enhance theirability to provide services. Animal health services, also a key responsibility of the woreda livestockofficewerenotidentifiedasapriorityareaforsupportbytheworedalivestockoffice.

EVALUATIONREPORT

19

More than 55 government offices benefitted fromAI equipment, even though delivery of procuredgoods and services by contracted suppliers was sometimes late. A total of 183 AI techniciansweretrained.Sixregionalandfederalleveldairysectorinstitutionsweresupported(100%ofrevisedtarget).Challenges were experienced in engaging more institutions due to budget constraints and limitedengagementbysomeregionalpartners.

ThesupportprovidedbyEDGETtotheWLOaddressedclearneedsoftheWLOintheprovisionofAIservices.Equipmentprovidedincludedtheprovisionofnitrogenflasksandmotorcycles,criticalforAItechnicianstomaintainthequalityofsemen,increasetheirresponsetimeandexpandtheircoveragearea.ThesupporttoAItechniciansincreasedtheircoverageandshowedaperceivedimprovementinthedeliveryofAIservices.ThesuccessoftheAIsupporthasbeenfurtherbolsteredwheretherewasanoverlapwithAGP(e.g.AletaWondo),whichsupportedtherecruitmentofadditionalAItechnicians.Overall, the capacity training and equipment support played a constructive role in enabling theongoing provision of AI services in theworeda. The technical trainings and equipment provided arelikelytomakeasustainabledifference(goodquality,etc.).However,themotivationofAItechniciansappears tobe variable, budget constraintson theWLO side for logistics/transportation/fuel and theirregularandinsufficientsupplyofqualitysemenandliquidnitrogenremainchallenges.

Finally, a number of actors thought animal health (i.e. access to vets/medicines) was important,despitethisbeingbeyondthescopetheproject.

Engagementwithregional/nationalforageseedsuppliers

In2016,engagementwithgovernment regional forage seedmultiplication centresbegan toaddressthe shortage of forage seed and planting materials. Selected institutions had to develop projectproposals for the futuresupportbyEDGETproject.OnlySARI’s (southagricultural research institute)proposalonforageseedmultiplicationledto130quintalsofforageseedmultiplied.

Thesupplyofimproved/qualityforageseedconstitutesamajorconstraintinthedairyvaluechain.Assuch,workingwithcredibleandreliable-supplyinstitutionstodevelopseed/plantingmaterialsishighlyrelevant. While the work with SARI has been positive, it is clear that the quantity of forage seedproducedfallsshortoftherequirements.Assuchtheoveralleffectivenessofthiscomponentrevealssomesignificantgaps.ThemechanismofrequestingproposalsfrominstitutionsseekingtocollaboratewithEDGETprovedchallengingduetolimitedcapacitywithinthesector.Alternativeapproachesmaybe required to address this gap. Finally, it is unclear if farmers’ demand or forage seed is wellunderstood. In light of this, a clear assessment of thequantity of seed required and a clear plan tomeetthedemandshouldbeundertaken.

SO4Developmentofaknowledgebaseondairyrelatedissues

EDGETdevelopedanddisseminatedextensionrelatedandgoodpracticematerialsindairyproduction,processing, marketing and development and developed its Learning and Knowledge ManagementStrategy in 2014. EDGET is an activemember of the Livestock Broader Platform and Livestock TaskForce.Theprojectorganisedavarietyofknowledgerelatedactivitiesatvariouslevels(woreda,zone),facilitateddiscussionamongstlivestockexpertsandextensionpersonnel,conductedreviewsessionsat

EVALUATIONREPORT

20

thecentralandregional levelsandpublishedvariouspracticebriefsandextensiontrainingmaterials.Regional managers attended various technical working groups and multi-stakeholder meetingscontributing to cross-organisational learning, collaboration and knowledge sharing. Overall theseactivities are seen to constitutean important setof contributions to knowledgedevelopment in thedairysector.

SO5Toimprovethenutritionalstatusofchildrenthroughdairyconsumption

Dairydevelopmentandnutritionhaveanumberofimportantlinkagesandcanincreaseconsumptionofmilkandprocesseddairyproductswithinthehouseholdcreatingnutritionaloutcomes.ThethrustofEDGETproject’sworkonnutritionwasthetestingandpilotingofapproachestoSBCC.Thiscameaboutaftertheinitialnutritionstrategywasfoundtobeunviable.

AwarenessraisingCampaignsandNutritionPilot

TheEDGETnutritionstrategy(2015)includedawareness-raisingqualitynutritionandmilkproductsindiversifieddiets;andmilk fortificationasa solution tomicronutrientdeficiencies inchildrenunder2andpregnantandlactatingwomen.EDGETfocusedontheimplementationofanutritionalbehaviour-change communications strategy and a pilot directed at behavior-change at the household level.Awareness raising campaigns in 2016 and2017built onpast experiences including ‘Worldmilk day’withtheLivestockResourceDevelopmentandPromotionAgency.

TheNutritionPilot

EDGET commissioned quantitative and qualitative research in 2017 on key nutritional indicators forwomen, infants and young children, identifyingbarriers to improvedoutcomesandopportunities todesignaneffectivenutritionSocialandBehavioralChangeCommunication(SBCC)intervention.Resultsshowedonly39.4%ofchildreninthesamplemettherequireddietarydiversity,thatminimumdietarydiversity for women showed only 4.3% meeting a minimum of five food groups out of 10 forconsumption, 67.9% received information on infant and young child feeding practices and 58%mentioned health education by health workers. To test potential messages for use in the SBCCintervention, EDGET project commission EUREKA Health Services to conduct a Trial of ImprovedPractices (TIPs) in 2017. This pilot tested the compatibility of SBCC message-materials-channelsstrategy, i.e. sixmajor infant and young child feeding recommendations identified at the formativeresearchstage.

Whilethestudywasnotwithoutlimitations,itwasusedtoinformanutritionSBCCStrategyandScaleupPlanforthepromotionofappropriatenutritionfocusedonthe‘first1000days’.TheSBCCStrategybuilds on insights generated from thenutritionpilot and the TIPs report and is also integratedwithproject interventionandoutcomes.Mostof theobjectivesand indicators referenced in theplanarenow outdated (time-bound to December 2017) (BehavioralM&E process objectives and indicators).The proportion of SNV-EDGET supported households with less than two children is 20%. It willtherefore be challenging to reach 65,000 households with a women-child (<2 years) with MIYVNmessages that have concurrently been supported on dairy production, processing and marketing -buildingsynergieswiththedairyandnutritioncomponent.Furthermore,anindependentexternaland

EVALUATIONREPORT

21

rigorousevaluationofoutcomesandcampaigneffectivenessaresensiblestepsbeforelargescaleroll-outofactivities.

Cross-cuttingStrategies

Cross-cutting strategies include the promotion of women and youth entrepreneurship and climatechange.Thissectionfocusesspecificallyonthewomenandyouthentrepreneurshipcomponentsincethe climate change component is addressed in relation to SO1 and the adoption of climate smartpracticesbydairyfarmers.

Womenandyouthentrepreneurship

EDGET project has sought to promotewomen and youth entrepreneurship in the dairy value chainthroughpromotinglocalbusinessinitiativesinclusiveofunemployedwomenandyouthininputsupply,seedmultiplicationanddairyprocessingactivities.AstudytoassessgenderintegrationintheprojectledtothedevelopmentofanewgenderandyouthstrategyforEDGET.In2015,13DPUshadatleastone female board member plus one women out of two technicians hired for milk processing. ForAgIDs,womenapplicantsweregivenpriority.Butduetolimitednumberofapplicationsfromwomen,only 6 out of 51 AgIDswere actuallywomen-led in 2015. In 2016, only threewomen groups and 6youthgroupswereestablishedforforageseedmultiplicationandmarketing,shortofthetargeted36women/youthdairygroups.Byendof2016,anassessmentonhowtoengagewomenandyouthforthe project was completed, showing for instance that women struggle to acquire improved breeds(EDGETProgrammeGenderandYouthMainstreamingStrategyreportinSNNPregionalstate,2016).Apilotforgenderandyouthranin2017.Bytheprojectend,43womenandyouthenterpriseshadbeenestablished(86%ofthetarget).

Whilethepromotionofwomenandyouth-ledenterprisesandinitiatives,clearlyhasanimportantroleto play, evidence from the endline survey and the qualitative survey suggest that household levelgenderdynamicsandnormsareinneedofelucidation.Mostfindingsshowthatwomencarryoutthemajor share of dairy related activities (looking after the cows, milking them and producing milkproducts) at the household level, adding significantly to their existingwork load.Womenwere alsofound to participate less in trainings thanmen andhave less of a role in economic decision-makingsuch as the purchase of inputs and the marketing of milk products. Despite this, an in-depthassessment of the effectiveness or sustainability of the women and youth enterprise developmentcomponentwasnotcarriedoutaspartofthisevaluation.Analysisiscompoundedbytheabsenceofagenderstrategyduringtheinitialstagesoftheinterventionanddelaysinstrategicimplementation.

ConclusionThe evaluation found that the EDGET project has made significant and valuable contributions tostrengthening inclusive dairy value chains in Ethiopia. The project has tested and demonstratedapproaches foranumberofkeysubsystemsof thedairyvaluechain,encompassingextension, inputsupply,forageproduction,institutionalservicesandmilkaggregationandmarketing.

EVALUATIONREPORT

22

Overallthequantitativestudyattheendlinedemonstratedpositiveresults,particularlytheadoptionofimproved practices. Milk production and net income frommilk were also found to be significantlyhigher than at baseline.Unfortunately, however, issueswith thebaselinedata limited theextent towhichstrongquantitativeconclusionsregardingtherelativegainininterventionworedascomparedtocomparison woredas could be made for a range of indicators, including milk production and netincome.However,theendlineresultsareencouraging.Thequalitativepartoftheevaluationrevealedthatmostoftheeffortstostrengthenkeysubsystemsofthedairyvaluechainwerehighlyrelevantandeffective. Some risks to sustainability have been identified and require attention when consideringfutureinclusivedairyvaluechaininterventions.

More broadly, the evaluation concludes that future interventions would benefit from an increasedfocusatbothstrategicandmeasurementlevelsonunderstandingthekeyinterdependenciesbetweendifferentvaluechaincomponents;carefulconsiderationofthekeyfactorsthatdrivesustainabilityandperformance for each of the actors in the value chain; and a stratified/segmented approach thatdevises differentiated approaches for woredas with different levels of dairy potential.

EVALUATIONREPORT

23

1.Introduction

1.2.OverviewofEDGETprojectThe five-yearproject (2013-2017) is fundedby theEmbassyof theKingdomof theNetherlands andworksonthedifferentcomponentsof theruralmilkvaluechain in threeregions -Oromiya,Amharaand SNNPR. It covers 10 zones, 51 districts/woredas with dairy potential and around 353 kebeles,targeting 65,000 smallholder households. The EDGET project is implemented by SNV NetherlandsDevelopmentOrganizationinclosecollaborationwiththeMinistryofLivestockandFisheries(MOLF-formerlyitwasMinistryofAgriculture)anditslinebureausinthreeregions.

Theprojectaimsatunlockingthepotentialofdairydevelopment inEthiopia“to improvehouseholdincomeandthenutritionalstatusofchildrenthroughincreaseddairyproductionandenhanceddairyprocessing&marketing”.Specificobjectivesoftheprojectinclude:

SO1: To enhance sustainable dairy production and productivity, input supply and relatedservices;

SO2: Toincreaseprocessingandmarketingofdairyproducts;SO3: Tocontributetodevelopmentofinstitutionsandtodairysector-wideinitiatives;SO4: TodevelopaknowledgebaseondairyrelatedissuesandSO5: Toimprovenutritionalstatusofchildrenthroughdairyconsumption. Cross-cutting:a)Topromotewomenandyouthentrepreneurshipandb)climatechange

Theseobjectivesareachievedthroughfiveprincipalstrategiesthataddress

(i)inputandproductionsystems;

(ii)milkcollectionprocessingandmarketingarrangements;

(iii)womenandyouthentrepreneurshipwithindairyvaluechains;

(iv)dairy-relatedinstitutionaldevelopment;and

(v)widerknowledgedevelopmentforthedairysectorinEthiopia.

Input and production systems: EDEGT promotes increased forage production, increased access tosupplementalfeeds,strengtheningagro-inputmarketingsystems,andpromotestheuseof improvedmilk transportation equipment to achieve hygienic collection and safe transportation of milk toprocessing units and ultimately output markets. This is achieved through strengthening the publicextensionssystem,establishingDFEGsfor farmer-to-farmerextension,supportingAgro InputDealersandpromotinghouseholdlevelforageproduction.

Milk collection processing and marketing arrangements: The project creates or establishes DairyProcessingUnits(DPU)inareaswherea)thereisademandforproductsb)thereisapotentialforlocalsourcingof rawmilk, toestablishsustainablebusinesseswhichemployadaptedequipmentandbestpractices.TheprojectalsoengagesthewidersystemofdairyvaluechainactorsassociatedwithDPUs,

EVALUATIONREPORT

24

suchasdairyfarmerorganisations,smallandmediumenterprisesandcooperatives,asappropriatetoeachcontext.

Womenandyouthentrepreneurshipwithindairyvaluechains isencouragedacrossthevalue-chainbypromotingtheinvolvementofthesegroupsinsuchorganisations,withaviewtostrengtheningtheirrole in theprovisionofdairyextension services, inputmarketing,milk collectionandprocessingandthemarketingofmilkandmilkproducts.Theprojectalsoaimstoensurethatithasapositiveimpactonchildren’snutrition throughadedicatedawarenesscampaign, though thishasonly recentlybeenrolledout(2017).

Atthesectoral level,dairyinstitutionaldevelopment isfosteredbyworkingcloselywithdairysectorinstitutionssuchasregionalartificialinseminationcentresandagriculturalresearchcenters.

Finally,theprojectputsemphasisonknowledgedevelopmentthatincludesfarmer-to-farmerlearning,documentation and development of good practices through write shops and the design anddisseminationofknowledgeproducts.

1.2.ThisevaluationTheEDGETprojectboardcommissionedALINe toconducta finaland independentevaluationof theEDGETproject fromDecember2017-March2018.The focusof theevaluationwasonassessing theachievements of project in terms of results (output, outcome and impact), assessing its overallcontributiontostrengtheningdairyvaluechainsinEthiopia,andcapturinglessonlearntforupcomingsimilar dairy programs/interventions. As part of this, the evaluation also assesses the relevance,effectiveness, and sustainability of interventions and their outcomes for key dairy value chain sub-systems(e.g.extension,agroinputdealers,foragesystem,etc.).Morespecifically,theevaluationseekstoaddressthefollowingkeyaspects:

Table1Evaluationfocusandevaluationobjectives

ThreeAreasofFocus EvaluationObjectivesasnumberedintheTORS

Evaluationofperformanceandapproach

(1)Toassesswhetherandtowhatextenttheprojectachieveditsoutput,outcomeandimpactresults;

(5)Toassessappropriatenessandeffectivenessofstrategiesandapproachesusedintheprojecttorealisetheintendedresults;

Measurementofchange/impactandbeneficiaryvoice

(2)Toassessimpactonkeydairyvaluechainactors/stakeholderswhohavebenefitedfromtheprojectinterventions;

(4)Toassesstheprojectcontributiontoeconomic&socialempowermentofwomenandclimatechangeadaptation&mitigation;

Learning,widerimpactandreplication/scaling

(3)Toassesswiderrelevance&contributionoftheprojecttodairydevelopmentinEthiopia;

(6)Toidentifyanddocumentlessonslearntforthedesignandimplementationofafutureprojectforsmallholderdairyfarmers,and,developmentofthedairysectorinEthiopia.

EVALUATIONREPORT

25

2.MethodologyDetailed evaluationmethodology is reported in the Inception Report and field level data collectionimplementationisintheFieldReport.Inthefollowingsection,summariesofbothreportsareincludedalongwiththelimitationstoquantitativeandqualitativedataanalysisandinterpretation.

2.1.OverviewofmethodologyIn response to the specifications set out in the TORs and the evaluation and learning questions, amixedmethodsevaluationapproachwasproposedcomprisingthefollowingcomponents:

● Aquantitative component focusedon the incomecomponentof the results chain throughahouseholdsurveycoveringtreatmentandcomparisonworedas;

● Aqualitativesurveyofallthekeyactorsinthedairyvaluechainacrossthethreeregions,fromthe regional level through to the kebele level covering government officials involved in theextensiondeliverysystem,DairyProcessingUnits,variouscategoriesofinputdealers/suppliersandotherkeyinstitutionsandprojectpartners;

● A review of secondary data, including project documents and M&E data as well as widerdocumentationrelatedtothedairysectorinEthiopia;

● Ananalysisof the relevance,effectiveness,efficiencyandsustainabilityofall thekeyprojectcomponents (seetablebelowforasummaryofhowthesewillbeconsidered)aswellas theoverallapproachadoptedbyEDGETinstrengtheningthedairyvaluechain;

● A synthesis of the evidence, insights and lessons learned that will inform scaling-up and orfurtherdevelopmentoftheapproachinasecondphaseoftheprojectinEthiopia.

Quantitativecomponent

The quantitative component included an endline household survey targeting both beneficiary andcomparison group farmers.Data available from thebeginningof theproject (baselinedata) allowedsomecomparisonofchangesinindicators,suchasincomeormilkproductionperhousehold,withinaquasi-experimentaldesign.Analysis includeddescriptivestatistics,testsforstatisticalsignificanceandanapproachusingaDifference-in-Differences(DID)analysis.Resultswillbedisaggregatedformaleandfemale-headedhouseholds.

DIDreliesonmakingacomparisoninkeyindicatorsbetweenthebaselineandendlineforatreatmentandcomparisongroup. Itentails comparing thechange in income in thecomparisongroup (withoutexposuretotheEDGETproject)withthechangeinincomeintheinterventiongroup(withexposuretotheEDGETproject),asshowninFigure1,below.

EVALUATIONREPORT

26

Figure1DifferenceinDifferenceAnalysis

Samplesizeestimation

The minimum required sample size for the endline was estimated based on the impact indicator“averageincrementofannualdairyincomeoftargeteddairyhouseholds”.Assumingwewanttoseea100% increase innet income fromthebaselineandassuming that thecomparisongroupwill seeanincreaseofnetincomeof10%between2013and2017,wewouldliketodetectadifferenceofatleast2,454Birrbetweenthetwogroups.Inordertodetectadifferenceofthismagnitudethatissignificantwith90%confidenceandapowerof80%,therequiredsamplesizeforeachgroupis95.Withadesigneffectof2, therequiredsamplesizeforeachgroup is190.This issmallerthantheproposedsamplesize of 432 for treatment and 218 for comparison, indicating that the proposed sample size issufficient.1

Samplingprocedure

Athree-stagesampleselectionprocedurewasusedtoselectthehouseholdstobeinterviewedforthisevaluation.TheprocessisexplainedinmoredetailintheTablebelow.ThefinalelectionofcomparisonandinterventionworedasandkebelescanberequestedforfromSNVEDGETProject.

Table2Stagesinsamplingprocedure

Interventiongroup Comparisongroup

1stStage Atotalof12projectworedasweredrawnfromthelistofthe51 project woredas. For this, woredas in each region werecategorised into three strata based on their project basedperformance (low, medium, high). Woredas were thenrandomly selected using a probability-proportional-to-size(PPS) sampling procedure, with the number of woredasselectedfromeachregionandeachstratabeingproportionaltothenumberofworedasintheregionandstrata.

Itwasagreedtousethesame5woredasasperthebaselinestudybasedondairypotential.

1https://select-statistics.co.uk/calculators/sample-size-calculator-two-means/

EVALUATIONREPORT

27

Interventiongroup Comparisongroup

2ndStage 3 project kebeles were selected from each woreda, atrandom.

Hence,weselected4kebelesatrandomfromthe35kebelesofeachofthe5woredas.

3rdStage 12 project households were randomly selected in eachkebele for interviewing. This accounts for a total of 36project households per woreda. ALINe used the lists offarmersprovidedthroughtheEDGETprojectM&Especialisttoidentifyindividualhouseholds.

11householdswillberandomlyselectedfromeachkebele,amountingtoatotalof220comparisonhouseholds.Theteamofenumeratorswillcloselyworkwithworedaorkebelelevelofficialstopreparelistofhouseholdswhohavedairycows.Theselectedhouseholdsincomparisonworedasweregivenincentiveforinterviewintermsofconsumableitems(soap,sugar,etc.)tocompensatefortheirtime.

Householdsurveytool

The quantitative component of the study used a household questionnaire. The householdquestionnairewasadministeredtothehouseholdheadalongwithoneotherpersonwhoisprimarilyinvolvedindairyrelatedactivities,typicallyoftheoppositesex.

ThehouseholdsurveywasstructuredassetoutinTablebelow.ThefinalhouseholdquestionnairecanberequestedforfromSNVEDGETProject.

Table3modulesandkeytopicscoveredforthehouseholdsurvey

Module Keytopics

Backgroundinformation • Infoonthesurveyrespondents(sex,age,etc.)

HHprofile/socio-demographics

• HHtype(maleorfemaleheaded),sizeandchildren6-23months• Primarysourceofincome• Landownershipandcultivation(includingforage)andchangeoverlast4years• Groupmembership• Dairyassets

Livestockownership • Currentownership• Changeinownershipoverlast4years• Whomanagesdairycattle

Participationinprojectactivities

• Participationintraining(bytopic)• Receiptofcoaching/follow-upsupport(bytopic)

EVALUATIONREPORT

28

Module Keytopics

Adoptionofpractices • Breeding• Health• Forageseed• Feeding• Calfmanagement• Housingandmanuremanagement• Climatechange• Milkingpractices• Otherinputs

Dairyproduction,processingandmarketing

• Annualmilkproduction• Marketingofrawmilk• Processingofmilkandmarketingprocessedmilkproducts

Non-dairyproduce/income • Saleofcattle• Saleoffeed/seeds• Saleofmanure• Saleofbreedingservices

Householdconsumptionofdairyproducts

• Purchaseofmilk• Milkanddairyproductconsumption(byagegroup)• Milkgivenaway,milkwasted

Keyconstraintstodairyexpansion

• Listingofconstraintsandranking

Impactsonwomen'slabour

• Changesinwomen'stimeallocationtodairyactivities

Secondarydata

Secondary data was collected on a limited number of indicators from Agro-Input Dealers, DairyProcessing Units andWoreda Livestock Offices. This was analysed with simple descriptive statistics(includingmeanand%increases).Issuesrelatedtotheavailabilityofthedatainasuitableformatforcapturelimitedtheextenttowhichthisdatacouldbeused.Resultswereintegratedintocasestudiesandtheoverallqualitativeanalysis.

Table4Overviewofquantitativeandqualitativedatacollected

Quantitative Qualitative

Region Type #Households #Organisationsforsecondarydata

KeyInformantInterviews

FocusGroupDiscussions

Amhara Intervention 144 18 19 8

Comparison 88

Oromia Intervention 180 14 22 8

EVALUATIONREPORT

29

Quantitative Qualitative

Region Type #Households #Organisationsforsecondarydata

KeyInformantInterviews

FocusGroupDiscussions

Comparison 88

SNNP Intervention 108 14 11 4

Comparison 44

Total Intervention 432 46 52 20

Comparison 220

Qualitativecomponent

Fivequalitativecasestudiesat theworeda levelwerecompiled,each,coveredsix foci including; theextension system, the forageproduction system,agro-inputdealers, dairy collection,processingandmarketing,andinstitutionaldevelopmentsupport.

Dangla (Amhara), Machakel (Amhara), Lemu Bilbilo (Oromia), Wuchake (Oromia) and AletaWondo(SNNPR)wereselectedforthecasestudies.Theworedaswereselectedonthebasisoftheirdiversitytoenablelearningacrossthespectrum.Theydifferintheirperformanceandthesupportreceivedfordairyextensionservices,cooperativeswithDairyProcessingUnits(DPUs),andAgro-InputDealers.TheassessmentofprojectworedasusedtoinformtheselectionofcasestudyworedaswasconductedbyEDGETprojectofficers.

Case studies were assembled through mapping of key dairy value chain actors, conducting KeyInformant Interviews (KIIs)andFocusGroupDiscussions (FGDs)with them,andbygathering recordsfromcooperativesandDAsthatdescribedthenatureandscaleoftheiroperations.Atotalof52KeyInformantInterviewswereheldwithDevelopmentAgents,DairyProcessingUnitcommittees,aswellas Agro Input Dealers. In addition, 20 Focus Group Discussionswere heldwith DFEGmembers andDFEGleadfarmerrespondents.

Inaddition,17interviewswereconducteddirectlywithprojectstafftogetmoreinformationabouttheprojectperformance,contextand lessons learned.Thesestakeholders includedMOLFStateMinister,DutchEmbassyrepresentative,projectandregionalmanagersofEDGETproject,businesspartners,andstaffoftheEthiopianBureauofFinance&EconomicDevelopmentandregionalofficersattheBureauofLivestockandFisheries.

A seriesof interviewguides for all KIIs and FGDsweredeveloped toensure the rangeanddepthofinformation sought. Data was summarized by field staff and shared in prepared categories foradditionalanalysis.

EVALUATIONREPORT

30

Keylimitationstotheevaluationmethodology

Quantitativedata

• It was apparent that there were inconsistencies and gaps in time-series secondary dataobtained from AgIDs, WLOs, DPUs on key metrics (e.g. volume of inputs produced, milkcollected, payments made to coop members, etc.). AgIDs and DPUs often had incompleterecordsinplaceorwerehesitanttosharesecondarydatawithfieldinterviewers.Furthermore,thedata fromtheAgIDandDPUrecordswasgenerallynotavailable ina formthatcouldbereadily inserted in the forms for secondary data capture. In many cases this made theextractionofsecondarydataimpractical,particularlygiventhetimeconstraints.

• There was poor cooperation with government staff to provide adequate and detailedsecondarydataontime.SecondarydatarequestedfromkebelesinKuyuandWuchaleworedaarestilloutstandingbutwedonotexpecttoreceivethem.

• The sampling decisions for the quantitative household survey component of this evaluationweremadeonthebasisof (1)theprioritycomparisonsthatwererequired(interventionandcomparison,baselineandendline);(2)estimatesofthevarianceinpriorityvariablesbasedonthe baseline data; (3) the assumed suitability of the comparison woredas selected at thebaselinestage;and(4)thefeasibilityofalternativesforcreatingacounterfactualgroup.Thesechoices were made under considerable time pressure and with limited opportunity to fullyinterrogatepoints(2)and(3).Ithassinceemergedthatthebaselineresultsthattheevaluationteamhadaccesstocontainederrorsandthatthecomparisonworedashadacombinationofhistoric andongoing livestock/dairy interventions and, as such, alsounderwent considerablechangeduringtheprojectperiod.Thissignificantly limitstheextenttowhichthecomparisongroupcanbeusedtomeasuretheeffectoftheEDGETproject.

• TheendlinesurveywascarriedoutinDecember2017.Itisworthnotingthatimplementationofprojectactivitieswascarriedoutmostintensivelyandatthegreatestscaleduring2016and2017,withagoodproportionofworedasnothavingreceivedallplannedinterventions(suchassupplyofequipmenttocooperativeswithDPUs)untilwellinto2017.Asaresult,itwouldbereasonable to expect that the full extent of benefits of participation in the project for dairyfarmerswouldnotyethavematerialised.Thiscould limittheextenttowhichproject-relatedchangestovariablessuchasmilkproductionandincomemightbeobserved.

• The endline questionnaire differed from the baseline for a number of questions.While thisallowed us to have clearer questions formulated in the manner required at the endline, itlimitedcomparabilityofthebaselineandendlinedatainsomecases.

StatusoftheBaselineHouseholdData

• The baseline dataset provided to ALINewas found to have a set of issues that significantlyimpactedonitscredibilityandusability.Keyissuesinclude:

o Poorcleaningofthedatawithmanyextremevaluesnotremoved;o Complexformatofthedata;o Significanterrorsinautomaticallycalculatedfields(e.g.totals);

EVALUATIONREPORT

31

o Large number of cases with missing values without clarity on how they should betreated;

o Somevariableshadnozero-valuesbutmissingvaluesinstead.• Asaresultofthis, theevaluationteamhadtospendan inordinateamountoftimeengaging

with the baseline data, reorganising it, cleaning it, addressing missing values, etc. Thisconsumedaconsiderableamountof time for theevaluation teamanddetracted fromothertypesofanalysis.

• Ultimately - and where possible - the evaluation team transformed (recalculated from rawfigures)thebaselinedatainordertomakeitcomparable.Thesetransformationsarebasedonbest practice (e.g.when substitutingmissing values) but also on best judgement (e.g.whendeciding ifablankvalueshouldbeconsidered’missing’ortakenaszero).This leftsignificantroomforerrorgivenour limitedunderstandingandopportunity to interrogatehowbaselinedata was actually collected, its quality and an assessment that it was neither cleaned nordocumentedusingprinciplesofbestpractice.2Itisimportantthattheinterpretationoffindingsthatreflectbase-andendlinecomparisonsshouldbemadewithcaution.

Qualitativedata

• Thequalitativetoolsweredelayed intheir finalisationduetosignificantrevisionupuntil theday before fieldwork commenced. Additional tools required more extensive discussion andtookmore time than has been planned. Qualitative data collection in three woredas (YirgaChefe,EnemayandKuyu)wasdoneusing tools thatwerepenultimateversionsof tools thatwerelaterupdated.

2PleaseseeEmailfromALINetoEDGETfrom30.01.2018forfurtherdetails.

EVALUATIONREPORT

32

3.ContextEthiopiaisafast-growingeconomyranking9thinpopulationgrowthinAfricaand14thintermsofGDPpercapitagrowth.Agriculturecontributes35.8%toeconomicGDPwiththedairysectorcontributing12-16%withinthat.3TheEthiopiangovernmentaspirestodoubledomesticmilkproductionbetween2015-2020toreduceitscurrentdependencyonimportsofdairyproducts.

Nationalstrategy

The national Growth and Transformation Plan II of 2015-2020 prioritizes transformation of theagricultural sector including a LivestockMaster Plan. In relation to the dairy sector, the Cow DairyDevelopment Roadmap (2015/16 - 2019/20) specifies ‘raising total cattle milk production to 7967millionlitresby2020throughgenetics,feedandhealthinterventionstoimprovetraditionalfamilycowdairyproductionandexpandandimprovespecialiseddairyproductionunits’(cf.page17,Roadmapforgrowthandtransformation).ForsmallholderdairyfarminginterventionstheRoadmapproposescross-breedingeffortswithexoticdairybreeds throughAI and synchronization, improved feedanduseofveterinaryservices.4

Milkproduction

11.33millionmilkingcowsinEthiopiaproducedatotalof3.06billionlitresofmilkannually(2015/16)5withstarkvariationsacrossdifferentregions.TheEthiopian‘milk-sheds’existinAdama-Asella,GreaterAddis,Mekele,Ambo-Woliso,Humera,BahirDar,Hawassa,DireDawaandJimma.Thevastmajorityofmilk(97%)isproducedbysmallholderfarmers(95%ownlessthan5cattle)whoaremostlypastoral,agro-pastoralfarmersorinmixed-croplivestocksystems,i.e.traditionalhighlandmixedfarming.Milkproduction is predominantly from indigenous breeds (97%) rather than the more productivecrossbreedsorpuregradeexoticcattle(3%).6

RuralmilkproductioninEthiopiafacesanumberofchallenges.Feedandforage(seed)fordairycowsareexpensiveandscarcelyavailabletosmallholders.Farmersfacealackofaccessibilitytolandwhichcanbedifficult toobtain fromthegovernment.Cross-bredcowsaremoreexpensive.AIservicesareoften difficult to access and may be of variable quality. Private AI providers (i.e. Addis LivestockProductionandProductivity ImprovementService)areperceivedtoofferhigherqualityservicesoverpublicAIserviceproviders(i.e.NationalArtificialInseminationCenter).

Milk quality is often poor due to inadequate adoption of hygienic practices as well as a lack ofadequateequipmentformilking,storageandtransport.Milkmaybedilutedtoincreasethequantityforsale.

3CIAWorldFactbook,2017.4InvestmentopportunitiesintheEthiopianDairysector(2015)5InvestmentopportunitiesintheEthiopianDairysector(2015)6InvestmentopportunitiesintheEthiopianDairysector(2015)

EVALUATIONREPORT

33

Collectionandprocessing

Dairy cooperatives collectand sellmilk toprocessors, sell rawmilkdirectly to consumersorprocessmilkin-house.Privatemilkcollectorsandprocessorsmayalsocollectmilkfromfarmersdirectly.Most,milkprocessingcompaniesinEthiopiaareconcentratedaroundAddisAbaba.Theprocessingcapacityoftheirplantsexceedstheavailablerawmilksuppliesofnearbycollectionsites.

Milk collection and processing is variable depending on the expertise/experience of staff, theavailability of equipment and access to road infrastructure.Many organisations do not have chilledstorageortransportequipment,insufficientqualitychecksandlackqualitybasedpaymentsystems.

Consumption

Humanconsumptionofmilkproducedis68%wherebytheremainderiswastedorutilisedbycalvesinconsumption.Only6.6%ofmilkproducedactuallyenterstheformal(e.g.viacooperatives)orinformaloutput market (urban sales or sales to neighbors). The majority of milk is consumed (48.5%) orprocessed(44.6%)bythemilk-producing-householddirectly.7

ThepriceperlitreofprocessedmilkinsupermarketsinAddisAbabais1.02EURandaveragepercapitaconsumption51.9 litresannually.Milkconsumption issignificantly lower inruralareaswith1.3 litresconsumedperyear,manymiddleandlowincomeconsumersconsiderpricestoohigh.Percapitamilkconsumptionhasbeenincreasingby2.2%peryearfrom2010-2015.8

Growth of the urban middle class with greater purchasing power will likely lead to an increaseddemandfordairyproductsoverthenextdecade.9ConsumerdemandformilkfluctuatesaccordingtotheOrthodoxfastingperiods,whilst,mediareportssuggestingthenegativehealthimpactsofAflatoxincontaminationinmilkin2014/15ledtoadecreaseindemandformilk.

7InvestmentopportunitiesintheEthiopianDairysector(2015)8Hemme,T.(ed.),2016.IFCNDAIRYREPORT2016.IFCN,Kiel,Germany.9PracticeBriefDairyBISSproject(2017)

EVALUATIONREPORT

34

4.EvaluationfindingsEvaluationfindingpresentedbelowaredonesoaccordingtokeycomponentsofEDGETproject’sdairyvaluechain interventions.TheoverviewsectiondescribesthedairyvaluechainsetuprelevanttotheEDGETproject,indicatingthedifferentcomponents.Thefollowingsectionprovidesanoverviewofthesocio-demographicprofileofhouseholdssurveyedatthebaselineandendline.

Subsequent sections (1) provide an overview of the component, (2) provide a summary of theimplementation of activities and achievement of outputs drawn from EDGET project’sM&E system,annualreportsand,insomecases,internalstudies;(3)presentthefindingsfromthebaselineand/orhouseholdsurveyspertinenttothesectioninquestion;and(4)highlightfindingsfromthequalitativecase studies. These individual pieces of evidence compiled together provide the evidence used toarriveatanassessmentof therelevance,effectivenessandsustainabilityofeachcomponent.A finalsection provides an overall assessment of the contribution of the project to strengthening thefunctionalityofthedairyvaluechainasawhole.

4.1OveralldairyvaluechainsetupThe figure below provides an overview of the dairy value chain, with a particular emphasis on theworeda level.ThegreyboxeswithdottedbordersrepresentkeysubsystemsofthedairyvaluechainthattheEDGETprojectengageswith,albeittovaryingextents.Theseinclude:

1. Theregulatoryandpolicyframework2. Regionalandnationalactorshigherupinthedairyvaluechain3. Extensionsystem4. Foragesystem5. Institutionalsupport6. AgroInputDealers7. DairyCooperativesandDPUs8. Householdproduction,processingsaleandconsumption

Key actorswho have a role to play are indicated in boxes 3 to 8. Actor boxeswith dotted bordersindicate that either the actor has a role to play but is not directly a part of the EDGETproject (e.g.AnimalHealthServiceproviders,Privatemilkbuyers/traders)orelseispartoftheEDGETprojectbutisnot found in all locations (e.g. Milk Collection Centres). Due to the high level of variation andcomplexity within the dairy value chain and across woredas, the diagram below offers a simplifiedpicturethataimstoconveythemainfocusareasoftheEDGETprojectandthekeyactorsrelevanttoEDGET project’s intervention. Amore detailed view of each component is included in the narrativesectionswithineachcomponent.

EVALUATIONREPORT

35

Figure2OverviewofthedairyvaluechaininEthiopiaOverviewofthehouseholdsurvey

Socio-economicprofileoffarmersatbaselineandendline

Table 5 and Table 6 provide an overview of the socio-economic profile of the comparison andinterventiongroupforbaselineandendlinesurvey.Estimatesforthepopulationoftargetfarmersandcomparisongrouparenotreportedbutsignificantdifferencesbetween interventionandcomparisongroup,includingovertime,arereported.

At baseline, themajority of household headsweremale (88-89%) and on average 45-46 years old.Therearenostatisticallysignificantdifferencesbetweeninterventionandcomparisongroupsforageandsex.Wecanseea10%reductioninproportionofmaleheadedhouseholdattheproject’sendline(77-80%)whileaverageagestayssimilartotheproject’soutset.Thenumberofhouseholdmembersishigherintheinterventiongroupthaninthecomparisonforboth,baselineandendline,andthereisan

EVALUATIONREPORT

36

overallincreaseinhouseholdsizeovertime.Theproportionofhouseholdswithchildrenundertwoisthesameforinterventionandcomparisongroup,butthereisa5%increasefrombase-toendline.

Withregardtotheeducationlevelsofhouseholdheads,atbaselinetherearemoreilliteratepeopleinthecomparisonthanintheinterventiongroup.Theproportionofilliteratehouseholdheadsdecreasesat the endline in favor of a higher percentage with primary education. These changes are morepronouncedforthecomparisongroup.

Weseeseveraldifferencesacrosstime;areductioninthepercentageofmaleheadedhouseholdsbyapproximately 10% as well as a reduced percentage of illiterate household heads. The number ofhouseholdmembersandtheproportionofhouseholdswithchildrenundertwogoesup.

Table 5 Socio-demographic background data for comparison and intervention group at base- andendline

Baseline Endline

Comparison(n=400)Intervention(n=1200)

Comparison(n=220)

Intervention(n=432)

Sex-headofhousehold 88%Male 89.1%Male 79.5%Male 76.9%Male

Age - head of household(meanaverage,min,max)10

45years,min=20,max=85

46years,min=21,max=90

45.9,min=22,max=80

47.5years,min=20,max=92

Education-headofhouseholda

Illiterate/noeducation

45.8% 33.4% 23.6% 27.3%

Primaryschool(1-4and5-8) 36.0% 43.4% 44.6% 45%

Secondaryschool 8.2% 10.3% 11.4% 10.2%

AdultBasiceducation NA NA 15.5% 13%

Other(specify) 9.8% 5.9% 5% 4.7%

Number of householdmembersa,c,d 5.5 6 5.9 6.3

Number of households withchildrenunder2,in%c,d 14.2% 15.8% 20% 20.6%a Significant differences in mean averages or cell distribution for baseline intervention and comparison, Chi-square teststatisticorTwo-sampleT-Test,p<0.05cSignificantdifferencesinmeanaveragesorcelldistributionbetweenbaselineandendlineforinterventiongroup,Chi-squareteststatisticorTwo-sampleT-Test,p<0.05dSignificantdifferencesinmeanaveragesorcelldistributionbetweenbaselineandendlineforcomparisongroup,Chi-squareteststatisticorTwo-sampleT-Test,p<0.05

10Therearesignificantdifferencesbetweenmaleandfemaleheadedhouseholdsintermsofage.Womenareinaverage47.8andmen45.8yearsold.Similarly,statisticallysignificantdifferencesexistforeducation.

EVALUATIONREPORT

37

Differences between intervention and comparison households surveyed can be seen in respect toeducationatbaselineandnumberofhouseholdmembersespeciallyforbothbase-andendline.

Thehouseholds’main sourceof income is crop farmingat thebaseline aswell as at theendof theEDGETproject.Dairy farmingasasourceof incomewasonlyaskedat theendline,buttheapparentdifferencesbetweencomparisonandinterventiongrouparenotstatisticallysignificant.

With regard to landownership,we see that farmers in comparisongroupownmore total land thanintervention farmers at end- but not at baseline. These differences at endline derive fromdifferentland sizes dedicated to crops and will be further discussed under section ’Forage and Forage SeedProduction-Landallocation’.

Table6Socio-economicdataforcomparisonandinterventiongroupatbase-andendline

Baseline Endline

Comparison(n=400)

Intervention(n=1200)

Comparison(n=220)

Intervention(n=432)

Mainsourceofincomeofthehousehold11

Dairyfarming NA NA 3.6% 8.3%

Cropfarming 95% 92% 95% 88.9%

Other 4.8% 7.9% 1.4% 2.7%

Landownership12

Averagetotalcultivatedland,inhab 2.35 2.3 2.59 2.21bSignificantdifferencesinmeanaveragesorcelldistributionforendlineinterventionandcomparison,Chi-squareteststatisticorTwo-sampleT-Test,p<0.05

Theresultsforcattleownershipshowsomesurprisingtrendsbetweenthebaselineandendline:

• Overall cattle herd size appears to have reduced marginally, albeit more so for interventionhouseholds;

• Theinitialproportionofhouseholdswithcrossbredcowsinthecomparisonworedasisjust4%,which appears to be strikingly low, both when compared with the intervention woredas atbaseline and when compared with the comparison woredas at endline. It is not possible toascertainwhythis isthecasebutpossiblereasonscouldinclude:poorselectionofcomparisonworedasatbaseline(i.e.notactuallysimilarto interventionworedas), issueswiththebaselinedatacollection(errors)and/orpresenceofintensive(non-EDGET)AIinterventionsincomparisonworedas.

11Categoriesbetweenbase-andendlinesurveydifferslightly.12Duetoamisunderstandingbyoneenumeratorregardinglandownershipquestion,weexcludedhis/herinterviewsfromtheanalysisforthisvariableresultinginadecreaseofthesamplesizeby59;Comparisongroupn=200,Interventiongroupn=393

EVALUATIONREPORT

38

Table7Cattleanddairycowownershipcomparisonandinterventiongroupatbase-andendline

Baseline Endline

Comparison(n=400)


Comparison(n=220)

Intervention(n=432)

Averagecattleherdsize,perhousehold 6.39 6.73 6.28 6.10

Averagenumberofcowsthatproducedmilkinthelastyear,perhousehold 1.74 1.72

Averagenumberofcrossbredcowsperhousehold 0.05 0.37 1.6 1.4

%ofhouseholdswithcrossbredcowsa,b 4% 28% 43% 31%a Significant differences in mean averages or cell distribution for baseline intervention and comparison, Chi-square teststatisticorTwo-sampleT-Test,p<0.05bSignificantdifferencesinmeanaveragesorcelldistributionforendlineinterventionandcomparison,Chi-squareteststatisticorTwo-sampleT-Test,p<0.05

Gendereddivisionofrolesindairyactivities

ThedistributionofdairyrelatedactivitiesamongwomenandmeninahouseholdcanbeseeninTable8 below. Both,men andwomen, are involved in looking after the cow (56%) butwomen only to agreater extent than men only. Women are more likely than men to be exclusively responsible forlookingafter thecows(35%womencomparedto9%men)milkingandprocessingmilk (around60%compared toaround8-9%ofmen)and for transporting themilk tomarkets (around55%ofwomencomparedto8%ofmen). In justover30%ofcases,householdsreportedthatbothmenandwomenare involved in theseactivities.Compared tootherpractices, thepurchaseof inputshas thehighestpercentageofhouseholdsinwhichmenareexclusivelyresponsible(29.1%).Thissuggeststhatwomenare generallymore involved thanmen in dairy related activities exceptwhen related to investmentdecisions(i.e.purchaseofinputs)whereinvolvementisthesame.

Table8Responsibilitiesfordifferentdairyrelatedactivitiesinthehouseholds13

Responsibilityfor:...looking

afterthecow...purchasingdairyinputs

...milkingandprocessing

...fortransportandmarketingmilk

...fortransportandmarketingprocessed

products

Womenonly 34.5% 27.0% 60.9% 59.3% 55.4%

Bothmenandwomen 56.1% 43.8% 30.5% 32.0% 36.3%

Manonly 9.2% 29.1% 8.6% 8.5% 8.1%

Hiredlabour 0.2% 0.2% - 0.2% 0.2%

13 Significant differences between intervention and comparison group farmerswere not foundwhich iswhy they are notreportedhere.

EVALUATIONREPORT

39

4.2StrategicObjective1:Toenhancesustainabledairyproductionandproductivity,inputsupplyandrelatedservicesStrategicObjective1coversaseriesofextensionrelatedEDGETprojectinterventionstopromotedairydevelopment.Theseinterventionsfocuson:

1. Theextensionsystem2. AgroInputDealers(AgIDs)3. Foragesystemdevelopment

Thediagrambelowprovidesanoverviewofthetimelineforimplementationoftheseactivities.

Figure3TimelineofSO1implementation14

Theextensionsystem

Overviewofcomponent

EDGETproject’ssupportonstrengtheningtheextensionsystemcanbedividedintothefollowingsub-components:

1. Strengtheningthecapacityofgovernmentextensionserviceprovidersattheworeda(WoredaLivestockOffices)andkebele(DevelopmentAgent)levels.

2. EstablishingandstrengtheningDairyFarmerExtensionGroups (DFEGs)asa farmer-to-farmerextensionmodel

3. Developing and distributing extension materials (manuals) as well as other key inputs (calffeed,MilkTransportationSystem(MTS)andforageseed).

The extension component of the EDGET projectworks primarily through the government extensionstructures.Regional livestockexpertsprovidedtrainings togovernment livestockexpertsat thezone

14PleasenotethatactivitieswithrelationtodevelopingaknowledgebasearereportedunderSO4.ButsomekeyfactmaybereportedunderthisSO1,too.

EVALUATIONREPORT

40

andworedaandzonelevelsaswellastoprojectZonalDairyCommunityMobilisers(ZDCM)andDairyExtensionPromoters(DEPs).TheseactorstogetherprovidedtrainingstothegovernmentDAs,whointurn-andwithtechnicalandfacilitationsupportfromtheZDCMs,DEPs,andworedalivestockofficers-deliver trainings and coaching/follow-up support to the farmers through theDairy FarmerExtensionGroups (DFEGs). The trainings encompassed orientation on the project and technical capacitydevelopmentfortheimplementationofprojectactivities.TheEDGETprojectalsodevelopedaseriesofillustrated trainingmaterials, translated into Amharic and Afan Oromo for use by extension serviceproviders as well as DFEG members. The ZDCMs, DEPs and Woreda Livestock Officers also play afacilitationroletosupportthedistributionofprojectrelatedmaterialsandinputs(suchasmanualsandforageseed)throughDAsandDFEGs.

DFEGsprovideamechanismforreachingalargernumberoffarmersthanwouldotherwisebepossibleandleveragesthepotentialofpeerlearningamongstfarmers.EachDFEGcomprisesapproximately25dairy farmermembers. Fiveof themembers inagroupare ‘lead farmers’ (oftenbutnotnecessarilycoincidingwiththemoreinstitutionalised‘modelfarmers’)andtheyaretheprimaryrecipientsofthetrainings provided by the DAs and DEPs. Each DFEG lead farmer is then expected to exchangeknowledgewithothersintheirgroup(sometimesreferredtoas‘followfarmers’)andtakeotherstepstopromoteandstrengthendairydevelopment.TheyalsosupporttheDAstocoordinateandcarryoutactivities such as trainings and exchange visits for their group members and play a role in thedistributionofprojectinputs.

In SNNPR and Amhara, DFEGs are federated into Kebele Dairy Development Committees. Thecommitteesarecomposedof3electedmembersinmostcases,drawnfromthepoolofDFEGleaders,and includebothmenandwomen. The roleof these committees is to facilitate coordinationacrossDFEGsandtofacilitatetheexchangeofinformationamongstmembers.Theymayalsoplayadditionalrolesrelatedtomilkcollectionandaccessing/purchasinginputs.

EVALUATIONREPORT

41

Figure4Actormapoftheextensioncomponent

Implementation:plannedvsactual

Table9below,basedondatafromEDGET’sM&Esystemandreportssummarisestheachievementofoutputsrelatedtotheextensionsystem.

Table9Extensionsystemactivities:achievementofoutputtargets

Outputdescription

Indicator Achievementendofproject

Endofprojectrevisedand(original)targets15

DairyFarmerGroupspromotingmilkproductionandmarketingorganizedandstrengthened

Numberofdairyfarmergroupspromotingmilkproductionandmarketingorganizedandstrengthened

>3,236/124% 2,600(2600)

NumberofDairyExtensionserviceproviderswhoreceivedTOTTrainingondifferentdairytrainingpackages

1,476/301% 490(490)

15 Some targets were repeatedly revised downwards, some indicators have been removed or added. Herewe report therevisedtargetsaspertheEDGETPerformanceM&Edata2014-2017.Originaltargetsfromreportsareinparentheses.

EVALUATIONREPORT

42

Outputdescription

Indicator Achievementendofproject

Endofprojectrevisedand(original)targets15

Numberofdairyfarmerswhoreceivedtrainingandextensionsupportondairydevelopment

56,107/86% 65,000(65,000)

Buildingcapacityofdairyfarmersandextensionsystem

In2014, theEDGETprojectconductedTrainingofTrainers (TOT) trainingson foragedevelopmentaswellascalfandcowmanagementfor486DAs,122woredaandzonallivestockexpertsand51projectDEPs. Over 30,200 farmers subsequently received a 1-day training, 110% as per the target for thatyear.TheprojectalsoinitiatedtheestablishmentandtrainingoftheDFEGs.

TOT trainings continued in 2015, with 1,162 DAs, 266 woreda and zonal livestock experts and 49projectDEPsandZDCMs.Subsequentlyover56,000dairyfarmers(86%aspertarget)receivedtraining.This year, EDGET project implemented its extension strategy for field level coaching and advisoryservicestofarmers.Asaresult,farmersstartedreceivingtechnicalfollow-upsupportandcoachingondairydevelopmentthroughtheDFEGs.

By 2016, once the EDGET project had become better established, themodel for extension supportfocused onworking through the DFEGs - i.e. trainings were delivered to DFEG leaders who in turnreachedout to theotherDFEGmembers.By thisyear,a totalof2,600DFEGshadbeenestablished.1,433DAs,livestockexpertsandDCMs/DEPstrainedover12,690leadfarmers(including1,462femalelead farmers). The lead farmers with support from the DAs and DEPs provided further extensionsupporttomorethan54,600farmers(84%asperplannedtarget).

By2017, cumulativeachievements forextension serviceproviders receivingTOT trainingswas1,476(DAs,livestockexperts,DCMs,DEPs)-abovethetargetsoriginallyset.However,thenumberofuniquehouseholdsreachedthroughthismodel,asreportedbytheEDGETproject,was56,107,whichamountsto86%ofthetotalEDGETprojecttarget.Duetoissueswithaccuratelytrackingthenumberoffarmersreachedandavoidingduplication,thefigureusedisthehighestnumberreachedinasingleyear.

ResultsfromanevaluationconductedbytheBureauofFinanceandEconomicDevelopment(BOFED)in2016, indicated that farmers who attended the EDGET project training sessions adopted good calfmanagement practices and showed changes in terms of Knowledge Attitude and Practices (KAP)relatedtofeedmanagementandoveralldairymanagement.However,performanceoftheextensiondeliverywasfoundtobevariableacrossworedas.

Findingsfromthehouseholdsurvey

The household survey gathered data on respondent’s participation in various types of dairy-relatedextensionactivities,includingtrainingandcoachingsupport.

Trainings

Overall, intervention group farmers received the training activities to a greater extent than farmersfromcomparisongroup.68%ofinterventiongroupfarmersreportedparticipatinginatleastonedairy-related training or exposure visit activity, as compared to 11% of comparison group farmers. These

EVALUATIONREPORT

43

figuresarethesameforbothfemaleandmaleheadedhouseholds.However,whenaskedwhointhehousehold(i.e.men,womenorboth)participatedinthetraining,only23%ofhouseholdsrespondedwith ‘woman’.This isdespitethefact thatwomenaretoagreaterextent involved inmilkcollectionandprocessingactivities(seeTable8above).

Thetrainingtopicsthatfarmersrecalledwereforagedevelopmentandfeedimprovement(26%),calfandcowmanagement(19%),hygienicmilkproduction(17%),feedandfeedingmanagement(16%),aswell as housing and manure management (14%). Farmers reported participating in dairy businessmanagementtrainings(4%)andexperiencesharingvisits(2%)toalesserextent.

The fact that11%comparisongroup farmers reported receivingdairy-related trainings suggests thatdairydevelopmentactivitiesarealsoongoinginthecomparisonworedas,albeitwithamuch-reducedintensity16.

Adviceandfollow-upsupport

47% of intervention group farmers received advice and follow-up support as compared to 6% ofcomparisongroupfarmers.Thisisregardlessofwhetherthehouseholdwasfemaleormaleheaded.Intermsof sourcesof adviceand follow-up, interventiongroup farmers reportedSNV /DEP in78%ofcases and government or public service providers in 21% of the cases. The content included foragedevelopmentandfeedimprovement(28%),calfandcowmanagement(19%),hygienicmilkproduction(17%), feeding and feedingmanagement (18%), aswell as housing andmanuremanagement (13%).Farmers received advice and follow-on support in dairy businessmanagement (3%) and experiencesharingvisits(2%)toamuchlesserextent.

The perceived benefit of advice and follow-up support is predominantly improved knowledge (56%)andinformation(36%)ratherthanchangeinpractices(2%)ornothing(7%).Whilethefarmersinthecomparison group were asked this question, too, they referred to support received in differentinstancesorotheractors.Notably, thissupportseemstobeperceivedas lesseffectivewith11%forcomparisongroupinsteadof1%forinterventiongroupfarmerssayingtheygained‘nothing’fromthesupport.Adoptionresultsarepresentedinmoredetailinthesectionon‘Householdadoptionofinputsandpractices.’

Table10Receiptoftrainings,adviceandfollow-upsupport

Endline

Comparison(n=220) Intervention(n=432)

Exposuretotrainings

%ofhouseholdreceivedtrainingondairyorexperiencesharingactivitiesoverlastfouryearsb

11.4% 68.1%

%offemaleheadedhouseholdreceivedtrainingondairyorexperiencesharingactivitiesoverlastfouryears

- 67.9%

16Thismaycauseconfoundingeffectswhenanalysingresultsonoutcomesbetweenthecomparisonandinterventiongroups.

EVALUATIONREPORT

44

Endline


%ofwomenparticipatingintrainings 22.6%

Exposuretoadvisorysupport

%ofhouseholdsreceivingadvisoryorfollowupsupportondairyproduction

6.3% 47.1%

...bygovernment/publicserviceprovider17 94.9% 20.6%

...bycooperative/farmergroup 2.6% 1.6%

…bySNV/DEP - 77.6%

…byother 2.6% 0.2%

Perceivedbenefitsoftrainingandadvisorysupport18

%reporting‘none’ 10.5% 1.2%

%reporting‘improvedknowledge’ 65.8% 55.0%

%reporting‘newinformation’ 18.4% 37.1%

%reporting‘improvedpractice’ 5.3% 6.7%

bSignificantdifferencesinmeanaveragesorcelldistributionforendlineinterventionandcomparison,Chi-squareteststatisticorTwo-sampleT-Test,p<0.05

Groupmemberships

At theendline,68%of intervention farmerswere found tobemembersofDFEGs.Atbaseline, sinceDFEGshadnotyetbeenformed,farmergroupmembershipreferstomembershipofacooperativeoranyothertypeof farmergroup/association.Membership levelswere47%forthe interventiongroupand50%forthecomparisongroup.

Table11Groupmemberships

Baseline Endline

Comparison(n=400)


Comparison(n=220)

Intervention(n=432)

%ofhouseholdsfarmergroupmembership 49.5% 47.2% NA NA

%ofhouseholdswithDFEGmembership NA NA NA 68.3%

17Pleasenotethatfiguresforcomparisongrouphavetobecaveatedbyaverysmalln=13forthisvariableandbelow18Pleasenotethatfiguresforcomparisongrouphavetobecaveatedbyaverysmalln=38

EVALUATIONREPORT

45

Projectinputssupply:Forageseed,calffeedandMTS

The EDGET project provided households with forage seed, calf feed and MTS. A total of 64.4% ofhouseholdsintheinterventiongroupreportedreceivingtheMTS,comparedto0.6%inthecomparisongroup.33%of interventiongrouphouseholds reported receiving forage seed (compared to13.9% inthe comparison group) and 32% reported receiving calf feed (compared to 8% in the comparisongroup). These figures indicate that inputprovisionactivities arebeing carriedout in the comparisonworedas, albeitwith a reduced coverage. Please see respective sections on forage and forage seedproduction,andAgIDsformoreagranularanalysis.

Table12Percentofhouseholdsreceivingvarioustypesofinput

Endline

Comparison(n=220) Treatment(n=432)

%ofhouseholdsreceivingMTSatleastonceb 0.6% 64.4%

%ofhouseholdsreceivingforageseedatleastonceb 13.9% 32.6%

%ofhouseholdsreceivingcalffeedatleastonceb 7.8%19 32%


Findingsfromqualitativeassessmentbysub-component

The table below presents the findings from the qualitative assessment by sub-component. Unlessspecific woredas are stated, the strengths and weaknesses are generalised across all the woredas.

19Asreportedbythesefarmers.ThesemaybecompletelyunrelatedtoSNV’sinterventions.

EVALUATIONREPORT

46

Table13Qualitativefindingsregardingtheextensionsystem

Sub-component Strengths Weaknesses

Developing

capacityof

extension

providers(WLO,

DA)

Thereiswidespreadappreciationacrosstheworedasfortheroleplayedby

theDEPsinrelationtoextension.Thisencompassesgroupformation,

distributionofinputs(forageseeds,trainingmaterials),technicaltrainings

andfollow-up.

AlmostallDAshavebeenplayingtheirrolesproactively.

Overall, DAs feel that their knowledge on dairy related activities has been

enhancedthroughthesupportthattheywereprovided-particularlyinterms

of technical trainingsrelatedtodairying.Theyreportedthat thishashelped

themtoplaytheirrolemoreeffectivelyandbringaboutpositivechangesfor

farmers. They also reported finding the extension materials they were

providedwithtobeuseful.

Continuityinthesupplyofinputs(Wuchale,WLO)

ThemodelofDAsvariesacrossregions. InOromiya,DAsdividetheirkebele intothree

areasandservethefarmersintheirassignedarea.AsaresultDAswhoarenotlivestock

specialistshavetoprovideadviceondairyingandreportedhavinglessconfidence.

More generally, across theworedas,DAswere reported tobeoverstretched (an issue

validated bymultiple actors), having to cover a very large geographical area, thereby

limitingtheirabilitytoprovidetherequiredsupport.OtheractorsoccasionallyfoundDAs

tobeexcessivelydrivenbypoliticalincentivesratherthanservingfarmers.

DFEGmembers reportedmixed views about the roles of DAs,with some appreciating

theirroleinstrengtheningdairyactivities(e.g.AletaWondo)andothersfeelingthatthe

DAs either lacked knowledge or were too busy with other work to play their role. In

somecases(e.g.LemuBilbilo),itappearsthattheDEPwasseenbyDFEGmembersasa

muchmore important contributor than theDA to extension activities, suggesting that

theDEPstepsintofillingapsintheDA’srole.

InsomecasesitwasnotedthatDAsrequireincentivestoplaytheirroles.Insomecases

DAswerealsoreportedtohavedistributed inputs (forageseed/cuttings/splits) tonon-

targethouseholds. Inothercases, followupby theDAsandDEPswasperceived tobe

verylimited(e.g.LemuBilbilo,Wuchale).Astheyareperceivedtohavemoreknowledge

thanDFEGleaders,DFEGmembersfeltthatrelianceonDFEGleaderswaslimitingtheir

accesstoaccurateandqualityknowledge/advice.

DAsarealsoexpectedtoplaya role inbringingAI services;however, theyareseenas

beingsomewhatindifferentinthisregard.

Establishingand

strengthening

DFEGs

DFEGsappeartohavebeensuccessfullyestablishedacrossthecasestudy

woredas.TheDFEGsdoappeartobedeliveringanumberofbenefitsto

farmers,thoughthereisconsiderablevariationinhowtheyoperateandthe

extenttowhichtheyareeffective.

Morespecifically,thereisamixedpatternacrossandwithinworedasinterms

ofhowDFEGmembersengagewithDFEGleaders.InsomecasesDFEGshold

regularmonthlymeetings,inotherstheydonotandinteractionsaremoread

hocorinformal.Inmostcasesthereisacombinationofboth.DuringDFEG

meetings,extensionmaterialsappeartobeused/discussed-thoughinsome

casesthisdoesnothappen.ItwascommonlyreportedthatDFEGmembers

wouldapproachtheirleaderstoseekadvicewhentheyneedit.

WhereDFEGsareworkingwell,membersreportedvaluingthegroupasa

meansofgainingknowledgeandlearningfromotherfarmers.Exposurevisits

toothers’farms,particularlyDFEGleaders,werefrequentlycitedasasource

ThepositiveattributesofDFEGsnotwithstanding,anumberofissuesandchallenges

wereidentifiedbydifferentactors.

Insomecases,DEP’sraisedconcernsabouttheselectionofhouseholds,whichtheyfelt

shouldhavebeendonebyexpertsduringdesignofproject.InLemuBilbilo,forexample,

somehouseholdshavenotevencollectedtheMTSyet.Inothercases,DFEGmembers

donotowncrossbreedcows.

DFEGleadersplayacrucialroleinthefunctioningoftheDFEGs.Insomecases,DFEG

leadersdidnotappearclearabouttheirrolesandresponsibilities(e.g.Wuchale).Across

theworedas,DEPsandinsomecasesDFEGmemberstoo,reportedunwillingnessof

DFEGleaderstoshareknowledgewith,provideadvicetoorshowtheirfarmstoother

farmers.

VariablelevelsofknowledgeandpracticeamongstDFEGleadersmaylimitthequalityof

knowledgetransferamongstfarmers(e.g.Dangila,LemuBilbilo).WhileDFEGleaderscan

EVALUATIONREPORT

47

Sub-component Strengths Weaknessesofmotivationandlearningaboutnewpractices.

Members also associateDFEGswith the ability to gain access to key inputs

suchvarioustypesofforageseedsaswellasequipmentsuchastheMTS.

ForserviceproviderssuchastheWLOandtheDA,DFEGleadersprovidea

simplifiedpointofcontactandmakeservicedeliverymoremanageable.DFEG

leadersserveasdistributeddemonstrationsitestohelpspreadknowledgeto

DFEGmembersandserveasabridgebetweenDAsandDFEGmembers.The

groupshavealsoallowedalargenumberoffarmerstobereachedwithina

shortamountoftime.

Inasmallnumberofcases(e.g. inoneDFEGinAletaWondoandanotherin

Lemu Bilbilo) , often due to individual leadership, DFEGs have created a

dynamicrelationshipwiththeirgroupandusetheircollectivestrengthtobe

more efficient in accessing government services (e.g. AI services, forage

seeds) as well as other inputs (e.g. clubbing together to get feed or

medicines).However,thispracticeremainssomewhatlimited.

DFEGmembersgenerallyratedthesupporttheyhadreceivedfromtheDEPs

quitehighlyacrosstheworedas.

behelpfulforexperiencesharing,extensionrequiresmorethanthis.Insomecases,

DFEGleaderswereperceivedasrequiringfurthersupportandnotplayingtheirroles.

WhereDFEGswerelessfunctional,memberscouldnotrecallactivitiesthatwerecarried

outandsomewerenotclearwhotheleaderoftheirDFEGwas.Inanumberofcases

DFEGmeetingswerenotheldregularly,andDFEGmembersfeltthattheirleaderswere

notreachingouttothem.

DFEGmembersoftenfeelthat‘somemembers’(thiswasnotmadeexplicit,butwe

presumetheywerereferringtoDFEGleaders)getmorevisits,training,supportand

benefitsfromtheDEPandDAthanothers.Theyfeelthatthisisnotfair.Thisreinforces

thenotionthat,ontheonehand,thattheydonotunderstandthepurpose/functionof

theDFEGsand,ontheother,thatleadersarenotplayingtheirroleeffectively.

OneDAreportedthatthefactEDGETapproachdidnotusethegovernment’sexisting1:5

networkwasanissue.However,itwasnotclearwhythiswasanissue.

Farmerscontinuetoexpectfreeinputsinsomecases.

Insomecases(e.g.inonekebeleinAletaWondo),DFEGmembershavenotreador

engagedwithanyofthetrainingmaterials.

CollectiveactionamongstDFEGmembers-suchasjointlypurchasingfeedormedicines-

appearstobesomewhatlimited,exceptwhereDFEGleadersareparticularlydynamic.

Insomecases,DAsfindthatDFEGmembersareunwillingtoattendmeetingswhen

called.Inthesamecase,DFEGmembersfindtheDAunhelpful.Thissuggeststheremay

beunderlyingissuesintherelationshipsbetweenactorsthathampertheexpectedroles

ofbothgroups.

Establishingand

strengthening

KDDC

KebeleDairyDevelopmentCommitteeswereoperationalin3outof5

woredas,asexpected(i.e.inalltheAmharaandSNNPRworedas).These

bodiesplayacoordinationrolewithrespecttoDFEGs.

Wheretheyhavebeenestablished,theKDDCsappeartobefunctional.

Insomecases,thedistinctionbetweentheKDDCandtheDFEGleaderswasunclearto

DFEGleadersandmembers,suggestingthatthereremainssomeambiguityaboutthe

respectiverolesofeach.

SomeKDDCmembersreportedthattheywereoverloadedwithworkand

responsibilities.

Extension

materials,content

andadoptionof

practices

DFEGmembersfrequentlynotedthefactthatextensionmaterialsareclear

anduseful.Inparticulartheyfindtheillustrationshelpfulaswellasthefact

thatthematerialsareavailableinlocallanguages.

OverallDFEGmembersreportedthatthenewknowledgeandinputssupplied

throughtheprojecthavehelpedthemtoincreasetheirmilkproduction.

Anincreasedtrendintheadoptionofimprovedpracticesrelatedtocalf

management,housing,feedingpractices(zerograzing,improvingqualityand

typeoffeed/forage),cleanmilkproduction,useofcrossbreedshasbeen

SomeDFEGmembersfeltthatthecostsofadoptingcertainpracticesoraccessingthe

requiredinputs(suchasimprovedfeed)couldbeprohibitive.

Insomeplaces(e.g.LemuBilbilo),dairyisarelativelynewactivityforfarmers-andthis

slowstheuptakeofnewpractices.

EVALUATIONREPORT

48

Sub-component Strengths Weaknessesreportedacrossalltheworedasbyalltypesofactors.

Acrosstheactorsandworedastherewasarecognitionofincreased

awarenessaboutbiogas.Somefarmersstartedusingbiogasandcomposting

forfertilizermethodsduetothetraining.Opengrazingisstillcommon

practice.

Overallextension

approach

Theoverallextensionapproachisseenbymostactorsacrossalltheworedas

tohavebeenpositivelychangedasaresultoftheEDGETproject.Ithas

shiftedfromatheoreticaltoapracticalfocus,particularlyasaresultofthe

focusonensuringavailabilityofinputs.Thenumberoffarmersreached

increasedasaresultofworkingthroughDFEGsandthecapacityofallkey

extensionplayershasbeenseentoincrease.

Someactorsfeelthatthetrainingandawarenesscomponentisthemost

significantcontributortochangeasithassignificantlychangedpeople’s

attitudestowarddairyfarminginapositivemanner.

Theoverallextensionactivitiesarealsoseenashavinggainedconsiderable

momentumoverthecourseoftheprojectasthebenefitsbecameevident

andmorefarmerscameonboard.

Farmers knowledge of and attitude toward dairy farming has changed

significantly

Across the board, respondents view the extension activities as having

contributedtoincreasedmilkproductionandmilkquality(lessrejection,also

becauseofsanitaryconditions).

Calves are growing faster and reaching maturity at an earlier age; calving

intervalshavealsoreduced,ashastheageoffertility;

Wheregoodmanagementandhousingpracticeshavebeenputinplace,the

problemofdiseaseshasreduced

Knowledge of dairy farming as a business (costs and returns) has also

improved

Farmers are shifting from a crop-based farming system to amixed (dairy +

crop)basedfarmingsystem

Milksaleshaveincreased

Provisionofinputsbytheprojectwasperceivedtobeinsufficientinquantitybya

numberofactors.

SomerespondentsfeltthattheEDGETprojectshouldreachouttoalargernumberof

beneficiaries.

Somefarmersdon’tyethavecross-breedcows,limitingtherelevanceofsomeproject

activities.

Inadequatesupplyofveterinarymedicineswasalsoidentifiedasanissuebyanumberof

actors,includingbothinstitutionalactorsandDFEGmembers.

Competitionofdairywithotheragriculturalactivities-e.g.coffeefarminginAleta

Wondo(DFEGmembersbecomeunresponsiveduringcoffeeharvesttime)-risks

underminingdairyvaluechainactivities.

FarmerTrainingCentreswereoftenfoundtolackresourcesandinputs,limitingtheir

utility.OnerespondentnotedthattheEDGETmodelwashelpingtoaddressthisandthat

thegovernmentshouldlearnfromtheSNVapproach.

WaterscarcitywasidentifiedasanimportantconstraintinDangilaandMachakel

woredas,whichimpactsnegativelyondairyactivities(e.g.washingcows,forage

production).

Diseaseoutbreaksinsomecases,posedchallengesandimpactedonmilkproductionand

cattlehealth.Appropriatemeasuresforaddressingdiseaseoutbreaksarerequiredto

managetheproblemwhenitarises.

EVALUATIONREPORT

49

Assessmentofrelevance


Overall the support todeveloping the capacityofpublic extensionproviderswas found tobehighlyrelevant. According to WLOs and DAs, the trainings and training material provided to them wererelevantandhelpedthemplaytheirrolesmoreeffectively.DAsreportedknowledgegainsmorethanWLOs,whereasWLOs tended toemphasise theadvantagesof theextensionapproachpromotedbytheEDGETproject.


GiventheconstraintsintheabilityofDAstoreachtheirtargetfarmers,DFEGswerefoundtoprovideaviablemechanismforreachingalargernumberoffarmerswithimprovedknowledge.However,theirrelevance depends heavily on the demand for new information by members and the capacity andmotivationofleaderstoplaytheirenvisionedroles.Itappearsthatthismayvaryconsiderablyacrossworedasandkebeles.

Trainingmaterials

ThetrainingmaterialsproducedbytheEDGETprojectwerefoundtobeusefulandrelevant.Theuseofillustrations was particularly appreciated. Where training materials have been translated into locallanguages - i.e. Amharic in Amhara and SNNPR, and Afan Oromo in Oromia - their relevance isperceivedtobehigher.However,somerespondentsnotedthatiftherearechangestotheextensionapproachinthefuture,thenthematerialsmayneedtogetupdatedagain.

Overall

The relevance of the extension component may be highest in those areas that are not alreadyrelatively well-established in terms of dairy development. In some cases, farmers felt that theextension supportwas not very useful to them and that their primary needwas support in gettingaccesstotherequiredinputs.

Assessmentofeffectiveness


Whileobjectivemeasuresofperformance forextensionserviceprovidersdidnot formapartof thisevaluation, recipients of the training interviewed through the qualitative work frequently reportedgains in knowledge - both with respect to the overall dairy development approach and on specifictopics.


Overall,theDFEGmodelappearstohaveworkedrelativelywellonthewhole,despitethechallengesand limitationsdiscussed in the findings fromthequalitative study. Inparticular, the introductionoftheDFEGmodelextendedthereachofDAs,allowingafargreaternumberoffarmerstobenefitfromthe dairy extension activities thanmight otherwise be possible. The evaluation team considers theinclusionofnon-model farmers(i.e.notdrawnfromthepoolofexisting ‘model farmers’used inthegovernmentextensionsystem)asDFEGleadfarmerstobeapositivemove,allowingmorefarmersto

EVALUATIONREPORT

50

play a role and helping to overcome entrenched power relations associated with existing modelfarmers who are often politically selected. The key constraints to effectiveness relate to weakleadershipoftheDFEGinsomecasesandalackofmotivationorwillingnesstoshareknowledgewithother farmers.Gaps in the understanding of the function and purpose of theDFEGs amongst somemembersarealsokeyissuesinthisregard.

Assessmentofsustainability


TheEDGETproject’sapproachofworkingthroughandbuildingthecapacityofthewell-establishedgovernmentextensionsystem-i.e.throughWLOsandDAs-ensuresacertaindegreeofsustainability.TheDAsconstitutealong-term,paid,skilledworkforcethatcanplayacriticalroleinstrengtheningdairydevelopmentbeyondthelifeoftheproject.TheknowledgeacquiredbyDAsandtheiraccesstotrainingmaterials,meansthattheyshouldbecapableofprovidingextensionadvicebeyondthelifeoftheproject.

However,anumberofriskstosustainabilityinclude:

• High rates of DA turnover, excessiveworkload and in some cases low levels ofmotivation allcombinetomitigatethecontributionofDAs.Thisisawell-recognisedandenduringissueintheEthiopiancontextanditmaybebeyondthescopeoftheprojecttoaddress.

• SomeDAs do not appear to have fully embraced their roles in dairy development,which haspresentedachallengetothesuccessfulimplementationofactivities.

• WLOsandDAshaveclearlydependedsignificantlyontheSNVDEPsinordertobeabletoplaytheir roles. This has not been limited only to the provision of trainings and the supply ofequipmentbuthas includeddirect involvementwithDFEG leadersandDFEGmembers.This isgenerally the case, but particularly sowhereDAs have not performed aswell. In some cases,DFEG members and leaders rate the DEPs as having played a much more central role insupportingthemthantheDAs.WhilethismaybeparticularlythecasewherecapacityofDAsisweak,itdoesraiserisksforsustainabilityandscalingpost-EDGETprojectsupport.


While there are clearly examples of DFEGs functioningwell, the overall sustainability of thismodelfacessomesignificantrisks.Keyissuestoconsiderinclude:

• AnumberofDFEGleaderssaidtheywouldcontinuetoplaytheirrolesiftheprojectcontinuestosupport them(ambiguouswhether this refers tomaterial supportorgeneralcoordinationandbackstopping support), but in some cases, they said that without support they would notcontinuetoplaytheirrole.WhileWLOsandDAsgenerallysaidtheywouldcontinuetoprovidesupportafter theprojectends, theyalsocitedvarious constraints (suchasbudget) thatmightrestricttheirabilitytoprovidesuchsupport.

• IncentivesforDFEGleaderstoplaytheirenvisionedrolesappeartobeweak.Thereappearstobelittleformal/directrecognitionoftheirroleamongstDFEGmembersandtheredonotappeartobeanyclearfinancialreturns.

EVALUATIONREPORT

51

• TheDFEGmodelhingescriticallyon themotivationandcapacityofDFEG leaders toplay theirroles in facilitatingknowledgeexchangewithandamongstDFEGmembers.Themotivation forDFEG leaders to play their roles aremostly articulated in terms ofwanting to support fellowfarmers.Atamorefundamentallevel,however,allfarmersstandtobenefitfromanincreaseintheproductionofqualitymilkasthisincreasestheoverallviabilityofthedairyvaluechain.

• WhereDFEGleadersarenotplayingtheirroles(orDFEGsarenotfoundtobefunctional)anin-depthappraisalneeds tobecarriedout.Thismayresult, forexample, inchanges to theDFEGleadership.

Having acknowledged the risks above, inmany casesDFEG leaders appear to bewell-established intheir communities,haveenduring relationshipswith theirneighboursandotherdairy farmers in thekebele/DFEG and are motivated to support their fellow-farmers. This signals a significant level ofvariationintheattitudesandrolesofDFEGleaders.

Keyissuesforconsiderationinasecondphase

Whatisthelong-termvisionforDFEGsandDFEGleaders?IstherescopeforintroducingsomeformofrotationamongstleadersthatprovideotherhighperformingdairyfarmerstoplaytheroleandrelievesthepressureonexistingDFEGleaders?Howwillthesegroupsbecontinuouslymotivatedandsustainedwithoutadditionalprojectsupportbeingforthcoming?

Reliance on theDEPs. To date, SNV has had a significant field presencewith a DEP posted at eachprojectworeda.Asnotedabove,theseDEPsarefrequentlyseentohaveplayedaveryimportantrole,insomecasescomplementingDAsandinothersevenhavingtocompensatefortheirshortcomings.IftheEDGETprojectextensionapproachistobereplicatedandscaledup,theviabilityofhavingsuchalargecadreofDEPsmaycomeintoquestion.

The combinationof theoretical trainingwith theprovisionof inputs is highlighted as a key to thesuccessoftheextensionmodel.Intheabsenceofanadequatesupplyofaffordableinputs(includingthroughthemarket)interestinthetrainingsmaybelimited.

Forageandforageseedproduction

Overviewofcomponent

TheEDGETprojecthassoughttoincreasetheavailabilityanduseofimprovedandappropriategreenforagetoenhancemilkproductionandquality.Thishasbeenpursuedthrough:(1)directprovisionofvarioustypesofforageseed,cuttingsandsplitstofarmers;(2)promotionofforagecropcultivationbydairy farmers (through technical trainings and motivating farmers); (3) promotion of forage seedmultiplicationbyindividualfarmers,throughseedproducergroupsandattheFarmerTrainingCentre(FTC) sites. The direct provision of forage seed was accompanied by trainings on both foragedevelopmentandforageseedmultiplication.

EVALUATIONREPORT

52

Figure5Actormapoftheforageproductionandforageseedsystem(requiresreview)


The table below, based on data from EDGET project’s M&E system and reports, summarises theachievementofoutputsrelatedtoforagedevelopmentandforageseedmultiplication.

Table14Provisionofinputsonforagedevelopmentandforageseedmultiplication

Output Indicators Target(%)achieved) Originaland(revised)target

Betterqualityinputs&servicestotargetedfarmersandVCactorsavailable

NumberofDairyHHsbenefitedfromforageinputsupplysupport 53,950/83% 65,000(65,000)

QuantityofDairyinputsdistributedtoDairyHHs

Forageseed:406,565kgCuttings&splits:22,325,596 Notargetsforquantityset.

Feedandforagesolutions

The EDGET project employed different strategies to address the constraints in feed and foragedevelopment indairyproduction.Householdswere supported togrowdifferent forage types, adoptimproved feeding techniques, use supplementary concentrate and other industrial by-products andimprovetheirfeed,forexample,withureaoreffectivemicroorganism(EM)treatment.

EVALUATIONREPORT

53

In 2014, EDGET project supported 27,693 households through free-of-charge distribution of forageseed, seedlings, cuttings and splits to grow different forage types (101% of target). As a result,householdsplantedanestimated1578haofforage(SNVEDGETprojectAnnualReport2014).

In2015,40,148households(i.e.96%oftarget)weresupportedwithforagedevelopment,plantingatleasttwotothreetypesofforageon3785ha.SeedinputsupplywasdifficultfortheEDGETprojectdueto the limited number of seed suppliers, quality issues and the absence of amarket based plantingmaterialsupplychain.Hence,individualfarmers,10farmergroupsand32kebelelevelFarmerTrainingCentreswereengagedforseedmultiplicationandthedevelopmentofplantingmaterialtoaddresstheforageseedsupplybottleneck.

The third roundof foragedevelopment in2016reached30,008households (133%ofannual target),resulting in6,753hacoveredunder foragedevelopment.DFEG’s tookonan important role in foragedevelopment support, i.e. forage seed and splits, cutting exchange and free provision. According toSNV’sowndata,80%oftargetedHHswerereportedabletoproduceatleastonetypeofforagebytheendof2016.Thisrepresentsa47%increaseonthebaselinefigure,resultingintheprojectachievingits5-yeartargetoneyearearly(SNVEDGETprojectAnnualReport2016,Page9).Thisisreportedtohaveledtoanimprovementinfeedsupplyresulted.

Additionally,SNVreportedthatfarmershavebeenareabletosavemoneypreviouslyusedforbuyinghayandadditionalfeedsupplements;andgeneratednewincomebysellingforageseedsandsplitstootherfarmers(SNVEDGETprojectAnnualReport2016,P9).In2016,EDGETprojectalsosupported36individualfarmers,6farmergroupsand57FTCsforseedmultiplication.

By 2017, the EDGET project had supported the production and distribution of 406,566 kg of forageseedand22,325,596foragecuttingsandsplits.

FindingsfromtheHHSurvey:

The household survey gathered data on respondent’s allocation of land, uptake of various foragedevelopmentandseedmultiplicationactivities.

Landallocation

With regard to landownership,we see that farmers in comparisongroupownmore total land thanintervention farmers at end- but not at baseline. These differences at endline derive fromdifferentlandsizesdedicatedtocrops.Allocationoflandforforageproductionandgrazingisthesamebetweenthe groups. Over time we see a reduction in the average area of land allocated to the followingcategories:(1)fallowandgrazing;(2)pasture;and(3)foragecropproduction.

Table 15 Land ownership and allocation for base- and endline data grouped by intervention andcomparisongroup

Baseline Endline

Comparison(n=400)


Comparison(n=220)

Intervention(n=432)

Averageareacoveredbyforageproductionfouryearsago,inhab NA NA 0.05 0.16

EVALUATIONREPORT

54

Baseline Endline

Comparison(n=400)


Comparison(n=220)

Intervention(n=432)

Averageareaofhomesteadorbackyard,inha 0.11 0.13 0.18 0.19

Averageareaoffallowland,inha 0.17 0.20 0.01 0.02

Averageareaoflandcoveredbyforagecrops,inhaa 0.07 0.19 0.06 0.09

Averageareaofgrazingorpastureland,inhaa 0.71 0.47 0.30 0.22

Averageareaoflandcoveredbycrops,inhab 1.29 1.31 2.03 1.65

Averagetotalcultivatedland,inhab 2.35 2.3 2.59 2.21

a Significant differences in mean averages or cell distribution for baseline intervention and comparison, Chi-square teststatisticorTwo-sampleT-Test,p<0.05bSignificantdifferencesinmeanaveragesorcelldistributionforendlineinterventionandcomparison,Chi-squareteststatisticorTwo-sampleT-Test,p<0.05

Regarding forage crop production, this decrease over time for the intervention group is consistentwhencomparingdatafrombaselinetoendline,butalsowhencomparingperceptionofrespondentasto ‘Averagearea coveredby forageproduction four years ago’. Pleasenote thatwearenot able tocalculatestatisticaldifferencesforeffectsbetweenbase-andendline.Investigatingthis issuefurther,welookedatthenumberofhouseholdswhoactuallysaidtheycultivatedforagecrops.Wecomparedpercentagesbetween‘Averageareacoveredbyforageproductionfouryearsago’andgrowingforagecropsnow.

Figure 6 shows that for intervention vs comparison group the percentage of respondents growingforage crops is 33% vs. 23%. When asked about cultivation of forage crops four years ago, thepercentagesare20%vs.13.2%respectively.Thedatashowsan increasingtrend inforagecultivationforbothgroups.

EVALUATIONREPORT

55

Figure6Bar-plotwith interval estimateofpopulationproportions (CI95%) forproportionof farmersallocationlandtoforageproductionnowandfouryearsback

Forageproduction

The survey found that 47.2%of intervention group householdswere producing forage legumes andgrassesduringthelast12months,comparedto40%inthecomparisongroup(statisticallysignificantatthe 90% level). 27.5% of intervention group households reported engaging in backyard forageproduction,comparedto18.2%inthecomparisongroup(statisticallysignificantatthe5%level).

Ofthe47.2%ofinterventiongrouphouseholdsinvolvedinforageproduction,32.4%ofwerefoundtobeproducingmorethanonevariety,14.6%morethantwovarietiesand2.3%producingmorethan3varieties.Whilethebaselinedidnotincludedataonthenumberoffarmersgrowingmorethan2or3varieties,theresultsforfarmersgrowingmorethan1foragecropshowsasignificantchangebetweenbaseline and endline from 6.5% to 32.4% in intervention woredas and from 6.4% to 20% forcomparison woredas. This suggests that the project has contributed to increased diversification offorageproduction.

Table16Overviewofadoptionoffeedingpractices20

Baseline Endline Comparison

(n=400)Intervention(n=1200)

Comparison(n=220)

Intervention(n=432)

%ofhouseholdsengagedinfarmer-to-farmerseedsexchangeb

0.0% 9.3%

20Pleasenotethatbaselinemeasureswerenotasked inthesamewayas inendlineandcomparisonsovertimearehencedifficult.

EVALUATIONREPORT

56

Baseline Endline Comparison

(n=400)Intervention(n=1200)

Comparison(n=220)

Intervention(n=432)

%ofhouseholdsengagedinforageseedmultiplicationb 3.2% 19.4%

%ofhouseholdsengagedinforageseedproductionb 8.6% 32.6%

%ofhouseholdsgrowingatleast1foragecrop 23.0% 47.1% 40.0% 47.2%

%ofhouseholdsgrowingmorethan1foragecrop 6.41% 6.5% 20% 32.4%

%ofhouseholdsgrowingmorethan2foragecrops 5.5% 14.6%

%ofhouseholdsgrowingmorethan3foragecrops 2.7% 2.3%


Forageseedproduction

In termsof forage seedproduction, 19%of households in the intervention groupwere found to beinvolved in seed production, compared to just 3.2% in the comparison group.While the practice offarmer-to-farmerseedexchangewasnotobservedatallinthecomparisonworedas,9%ofhouseholdssurveyedintheinterventionworedasreportedengaginginthispractice.

Figure7Bar-plotwithintervalestimateofpopulationproportions(CI95%)foradoptionofforageseedproductionpracticesincomparisonandinterventiongroup

AsshowninTable12,32.6%ofhouseholdsintheinterventionworedasreportedreceivingforageseedat leastonce in thepast fouryears, compared to13.9% in thecomparisongroup.Therewere some

EVALUATIONREPORT

57

notabledifferencesbetweeninterventionandcomparisongrouphouseholdsintermsofthesourceofforageseeds. Interventiongrouphouseholds involved inforageseedproduction,weremore likelytoreportAgIDsas a sourceof forage seed (24.1%) than in the comparisongroup (0% reportedAgIDs).However, for both groups, own production (36.5% to 43.7%) and government agents21 (31.1% to46.8%)werethemostcommonsourcesofforageseed.2.5%offarmersreportedgettingtheirforageseed from a cooperative or farmer group in the intervention woredas compared to 0% in thecomparisonworeda.Privatedealerswerereportedasthesourceforaround5%ofhouseholdsinbothgroups.

Inordertobetterunderstandtheprojectcontribution,respondentswereaskedtoreportthenumberofyearsthattheyhadadoptedeachoftheabovepractices.Respondentsweremostlikelytocite2or3yearsago,whichcoincideswiththeprojectperiodandsuggeststheprojectplayedaroleinpromotingthesepractices.

Whenaskedaboutperceivedchangesinforageseedproduction,householdsintheinterventiongroupweremorelikelytoreportanincreaseoverthelast4years(52.6%)thanhouseholdsinthecomparisongroup(36.8%).

Households from the comparison group did not report earning any income from the sale of forageseed, improved forage or natural grass/pasture. Only 11 households, i.e. less than 3%, from theinterventiongroupdid. Theaverage incomeearned fromselling forage seed (n=1), improved forage(n=5)ornaturalgrass/pasture(n=5)forthese11households,is2951.5Birr.Thereisagreatvariationintheaverageearningsfromeachofthesesources.

Findingsfromqualitativeassessmentbysub-component

The table below presents the findings from the qualitative assessment by sub-component. Unlessspecificworedasare state thepositiveandnegative findingsaregeneralisedacross all theworedas.

21ThequestionnairedidnotdistinguishwhethertheinputsprovidedbygovernmentagentswerefromtheEDGETprojectorsomewhereelse.

EVALUATIONREPORT

58

Table17Qualitativefindingsontheforageproductionandseedsystem


Forageseeddistributionandtraining

Forage seed distribution was successful. Farmers received different varieties offorage seed (Desho grass, elephant grass, mulato, alfalfa) and used them toproduce forage. The varieties were generally perceived (e.g. by the woredalivestockofficeinLemuBilbilo)tobeappropriateaspertheagro-ecologicalzone.VarietiessuchasDeshograsswerealsoappreciatedfortheirabilitytobegrownyear-round (provided moisture is available). Trainings were also provided onforage development and thesewerewell received. All stakeholders appreciatedthecombinationofthetrainingwiththeprovisionofinputs.

In a number of woredas, respondents noted in particular the willingness offarmers to use the new seed varieties, perhaps signalling a higher than usualreadinessoffarmerstoadopt/trialnewvarieties.

Overalltheproject isseenashavingledtoincreasedforageseedavailabilityandforage production, although in Wuchale there appears to be some questionregardingtheproject’scontributiontothis.

Farmershavebeenexchangingforageseedwitheachothertosecureaccesstoasufficientdiversityofforagetypes.Thishappenseitheroneortwotimesperyear(e.g.AletaWondoandDangila).Thisappearstobearelativelynewpracticethatisgraduallygainingacceptance.

Initially through training and then subsequently through direct observation andexperience sharing, farmers recognise the value of producing and providingappropriate varieties of forage to their cows, particularly in terms of increasedquantity and quality of milk production. Moreover, the forage varieties werefound tobepalatable for calvesaswell as cows. Inaddition tomilkproduction,improved foragewasseenasacontributor toanearlierageof fertility forcowsandreductionindisease.

Asaresultfarmersaremoreeagertocontinueproducingforageontheirlandandtofeedittotheircows.

For some farmers, seed production has also become an income generatingactivity, as they are able to sell seeds to other farmers. In Machakel woreda,forage production was seen as one of the most successful components of theproject, and a trend in shifting land allocation from crop production to forageproductionwasalsonotedbysomerespondents.Theincreasedallocationoflandtoforageproductionwasalsohighlightedinallworedas,althoughthereappearstobesomevariationacrossfarmers.

Theincreasedforagedevelopmentisalsoseenasapositivetrendwithrespecttoenablingfarmerstoshifttoazero-grazingmodelforfeedingtheirlivestock.

In Lemu Bilbilo, the practice of drying and storing forage appears to have been

Althoughtherehavebeenclearpositivedevelopmentsrelatedtoforageproduction,thekeychallenge-reportedacrossallworedas-isrelatedtotheavailabilityofseed.Insomecases, thiswasframed intermsofthetotalquantityofseeddistributedbytheproject (e.g.AletaWondo), theprojectnot continuing to supply improved seedvarieties (e.g. Dangila, Machakel) or the unavailability of seeds in general (e.g.Machakel, Wuchale). The lack of sustained seed supply availability has created aconstraint to forage production by farmers. This is exacerbated by the fact thatfarmers do not always keep seeds for replanting (despite being trained on this).Rathertheycontinuetoexpectthatnewseedswillbeprovidedbythegovernmentortheproject.

Asaresultofconstraintsintheavailabilityofforageseed,somefarmers(e.g.Dangilaand Machakel) have reported challenges in providing their cows with adequateforage.Thiswasreportedtohaveledtoareductioninmilkproductionandassociatedincome.

Whilefarmersareengagingintheexchangeofseed,somedidnotseethispracticeasalong-termsolution.

Resource constraints were also highlighted as an issue when it comes to forageproduction,particularlylandandwater(e.g.Dangila,Machakel)

InWuchaleandLemuBilbiloworedas, thesuitabilityof theseeds to the localagro-ecologywasraised. InWuchale, inparticular,coldweatherandtherequirementforcontinuousmoistureavailability ledto foragecropsdying. Inothercases (e.g.LemuBilbiloandMachakel), farmershighlighteda lackofavailabilityofsufficientvarietiesofseedandchallengesfacedparticularlyduringthedryseason.

In one of the kebeles inWuchale,members from one of the surveyed DFEGs (themajorityofwhomhadjoinedin2017)reportedthatworkonforagehadnotyetbeencarriedout.Theynotedthattheworedalivestockofficehaddonesomeworkbutthatit remained theoretical and hard to access asworeda officials did not come to theKebele.Inthisworeda,thelackofavailableseedandlimitedsuitabilityofthevarietyhasmeant that farmers have not continuedwith forage production. AnotherDFEGfromthesameworeda,reportedthatforageseedshadbeenintroducedbutwerenotwellsuitedtothelocalconditionsandwerenolongeravailable.

In Lemu Bilbilo one respondent noted that seeds had been provided by the DA tonon-target farmers.TheDAalsonoted that the seeddidnotperformwelland thatthis had created some resistance from farmers. In the same woreda the WoredaLivestock Office reported thatmelilotus had been incorrectly distributed as alfalfa.Some DFEG members also felt that DFEG leaders had not provided a sufficientquantityofseedtotheirmembers.

EVALUATIONREPORT

59

Sub-component Strengths Weaknessestakenup. Inareaswherethere isagoodmarket fornon-foragecrops-e.g.barleyforbeeror

potatoesforthelocalmarket-farmersinterestinallocatinglandtoforageislimited.

Forage preservation for dry seasons has proven to be challenging, with limitedadoptioninsomecases.

Accordingtosomeof theregional inputsuppliers, farmers’demandfor forageseedonthemarketisunderminedbythefactthatfree/subsidizedmaterialsareprovidedbyNGOsandgovernmentprojects.

Forageseedmultiplication

Insomecases(e.g.LemuBilbilo,Machakel),FTCshavebeenengagedinallocatingland for forage seed multiplication. This is seen as having contributed to animprovementintheavailabilityofforageseeds.

However, it appears that in Lemu Bilbilo,while forage seedsweremultiplied in2016,thiswasnotthecasein2017asaresultofcrop-rotationpracticeattheFTCsite.

ContinuityintheproductionofforageseedbyFTCshasprovedtobeanissueas(a)the FTC sites competewith other kinds of (non-dairy) demonstration activities; (b)croprotationispracticedsofoddercropcultivationgetsdiscontinued.

Moregenerally,themarketforforageseedremainsveryunderdeveloped.

EVALUATIONREPORT

60


Theforageseedandforagedevelopmentcomponenthasacriticalroletoplayinthedairyproductionsystem. Thanks to project activities, there is growing and relatively widespread recognition acrossvaluechainactorsofthecontributionthatfeedingappropriatetypesofforagecanplay,particularlyinrelationtoincreasingthequantityandqualityofmilk.


Thereare relatively successful casesof forage seeddevelopment throughFTCsand through farmersthemselves.InthecaseofFTCs,institutional,financialandmanpowerlimitationscurbstheviabilityoftheapproachatscale-andwhilethisapproachmaybeusefulfordemonstrationpurposes,itisunlikelythat it can adequately address the demand. The practice of farmers themselves multiplying,exchangingandsellingseed,either individuallyor throughseedmultiplicationgroups,holdspromiseandiscontributingtotheavailabilityofforageseed.However,forthesefarmersandtheirgroupswhoare involved inproducingforageseed,accesstoqualityseedwillbeessential forthemtobeabletoreplenishthegeneticstockoftheirforageseedovermultipleyears.


Overall theestablishmentofasustainableforageproductionsystemispremisedonanumberofkeyelements:

1. Continuous supply of improved/certified forage seeds appropriate to the agro-ecologicalconditions fordecentralisedmultiplication.This is theprimarybottlenecksince thesupplyofforageseedsistheprimaryconstrainttoproduction.

2. Demand for forage fromdairy farmers.Thisappears tobewellestablishedsince theprojecthasenhanceddairyfarmers’appreciationofthebenefitsofprovidingforagetocattle,throughbothtrainingsandexchangevisitsandreflectionsonexperience.

3. Capacityandmotivationtoproduceand/orbuyforageamongstfarmers.Thisisseentobeonan upward trend. Farmers have produced forage from the seed provided to them, oftenallocatinglandthatwasformerlyusedforfoodcropproductiontoforagecultivation.

4. Functional systems for forage seedmultiplication,whether throughFTCs, individual farmers,seedproducergroupsorothercommercialsetups:

o FTCs face institutional barriers (e.g. competing pressures with other land uses, croprotation),thatmaylimittheirviabilityasseedmultiplicationsites.

o Farmersfacelandandwaterconstraints,limitingproductioncapacity

Consideringtheabove,theoverallsustainabilityoftheforagecomponentremainsinquestion.Whilesignificantprogresshasbeenmadeandthemodeloffarmer/groupbasedmultiplicationappearstobeasuccessfulmodelofaddressinggreen foragesupplyconstraints in theshort term(provided farmershaveaccess to seed, landandwater) a larger scale solutionwill be required to create a sustainablesystemforforageseed.

EVALUATIONREPORT

61


However,todate,theforageproductionsystem,includingtheproduction,distribution/sale/exchangeanduptakeofforageseedshassomecriticalbottlenecksthatlimittheextenttowhichthefullbenefitofimprovedforagevarietiesisbeingrealised.Thisinturnplaceslimitsonthefurtherdevelopmentofthedairyvaluechain.Thesebarrierswillneedtobeovercomeinorderforfurtherdevelopmentofthedairy value chain to functionand tobe sustained. InDangila, theDPU is consideringplayingamoreproactiveroleinforageseedproduction.

Otherkeyissuestoconsiderinclude:

• Focusonthesupplyofforageseedsandthediversificationofmodelsforproducingitaffordably• Thecostofforageseed/forageproductioncanbeprohibitiveforsomefarmers• Continueexistingactivitiesrelatedtoforagemultiplicationandincreasethescaleofthem• Considerhowtoenhanceforagecultivationintheoveralllanduseandnaturalresourceplanning

processesintheKebele

AgroInputDealers

Overviewofcomponent

The EDGET project has sought to support the establishment and/or strengthening of Agro InputDealers in order to increase the supply of quality, affordable dairy farming inputs for smallholderfarmers.TheAgIDscouldbeprivatebusinessesorcooperativesandcouldbenewagenciesorexistingones. The project invited proposals from interested parties and used these to select suitablecandidates.SupportprovidedtotheAgIDsincludetrainingsandguidelines,theprovisionofequipmentandmaterials, exposurevisits andbusiness tobusiness linkages.By routing the supplyof key inputs(e.g.calf feed)andequipment (e.g. theMTS) throughtheAgIDs, theprojectsought tohelpAgIDstoestablishtheirnetworksanddistributionchannelsinordertoreachfarmers.

EVALUATIONREPORT

62

Figure8ActormapoftheAgroInputDealersystem


Table18AgroInputDealers

Output Indicator Achievementoftarget(as%) Revisedtarget(originaltarget)

Betterqualityinputs&servicestotargetedfarmersandVCactorsavailable

NumberofInputsuppliers/dealerssupported/strengthened

50(98%) 51(51)

Agro-InputDealers

In order to create a sustainable solution to addressing the shortage of quality and affordable dairyinputs, the EDGET project supported a total 50 Agro Input Dealers (AgIDs). Suitable AgIDs wereidentifiedin2014,andtheircapacitywasbuiltthroughvarioustypesoftrainings,materialsupport(e.g.display tables, shelves, signboard and uniforms), business advice/coaching and the facilitation ofbusiness to business linkages. 51AgIDswere selected in 2015, oneofwhichwithdrew in 2016. TheselectionofoneAgIDinGozamnworedawascancelledsincetherewereirregularitiesintheselectionprocedureandtheprojectcouldnotidentifyasuitablereplacement.

Aspartofthebusiness linkagesupport, theproject facilitatedBusinesstoBusiness(B2B)networkingevents(alsoatnationallevelin2017)betweenAgIDs,nationalandregionaldairyinputsuppliers,andmicrofinanceinstitutions(MFIs).Reportedly,MFIsshowedaninterestindevelopingacreditfacilityforAgIDs.

EVALUATIONREPORT

63

AgIDsalsoservedasanimportantmechanismforprovidingfarmerswithaccesstokeyinputs,suchasforageseedandcalffeed(neitherofwhichwerewidelyutilisedintheprojecttargetarea).ByworkingthroughAgIDs, theEDGETprojecthelpedtheAgIDs toestablish relationshipswith farmersandbuildtheirdistributionnetworkswhilealsoprovidingaccess tokey inputs.Afterestablishing linkageswithsuppliers, theEDGETproject introducedapro-poorvouchersystemforsupplementarycalf feed, theMilk Transportation Systems (MTS), forage seeds and other inputs (see Table 12 for details). Thevouchers were supposed to incentivise and subsidize risk-averse farmers to try out and adopt newtechnologiesandpractices.AgIDsreceivedcommissionfromtheEDGETprojectforprovidinginputstofarmers,redeemableuponsubmissionofthevouchers.DetailsontheMTSandcalffeedarediscussedbelow,whereasforageseeddistributionwascoveredintheprevioussectionontheforageproductionsystem.

MilkTransportationSystem(MTS)

In 2013, EDGET project secured a sub-licensing agreement to manufacture 500,000 units of MTS(locally referred to as ‘Mazzican’) - a high quality food grade plastic container with lid, filter andmeasurement gauge - to improve the hygienic collection and transportation of milk for farmers.Universal Plastics inAddisAbabawas contracted in 2014and received anup-frontpaymentby SNVEDGETprojecttobuyamachinethatcouldproducethecans.Itwasnotuntil2016that26,271MTSswere produced and distributed through AgIDs to the target households. Non-project householdsexpressed interest inMTS aswell. As a consequence and to create additional demand for theMTS,35,000unitswere tobedistributed tonon-targeted farmersaswell.By2017,a totalof95,000MTSweredistributed(95%ofthenewtarget).

Thedelay inproductionanddistributionofMTSduring the firstyearsof theprojectwererelatedtoseveralchallengesfortheproducerUniversalPlastics.Theseincludeashortageofforeigncurrencyforbuying the appropriate machinery and raw materials from international markets, secure timelytechnicalinputfromanItaliancompanyforthenewlyacquiredmachinery,aswellasfrequentelectricpowercuts.

Interview data suggests that the Mazzican is a high-quality product fulfilling quality standards andperceived as such by companieswhen exhibited. But for farmers, the benefitsmay not be quite soobvious.TheMTSqualitycomeswithahigherprice(e.g.comparedtosimplebucketsorcanswhichcostafourthoftheMTS)andataweightof900g.Especially,thelidoftheMTSwassaidtobetoolargeandmayrequireadjustmentinfutureprototypes.Asaresult–asreportedbyUniversalPlastics–thecurrentproductwouldbedifficult to sellon themarketand therefore reliesongovernment/projectsupportforongoingproductionanddistribution.

Feedsupplementstrategy

Under the aforementioned scheme to use AgIDs as a distribution system for calf feed, AgIDsdistributedthecalffeedtheyreceivedfromsupplierstohouseholdswithfemalecross-breedcalvesinexchange for vouchers. EDGET project extension staff, DAs and AgIDs were trained on assessingconditions of calves, ear tag applications, and other topics to be able to identify eligible calves for

EVALUATIONREPORT

64

supplementary feed. Households then received vouchers for feed supplements promoting also thelinkagebetweenthedairyfarmersandtheagro-inputdealerintheworeda.

In 2016, 13,755 households received 70-100 kg supplementary feed per calf for a total of 16,492calves.DAsandDEPs carriedoutweekly calf girth andheightmeasurements subsequent to the calffeeddistribution.ThiswasreportedbySNVtohavedemonstratedthattheuseofsupplementarycalffeedhadbroughtfastgrowthandimprovedhealthandbodyconditionand,thereby,tohavechangedtheperceptionamongstdairyfarmersofthevalueofpropercalffeedingandmanagementpractices.Aninformalsurveycarriedoutbytheprojectfoundthattheageatwhichfemalecalveswerereadyfortheir 1st AI service had reduced from around 24-36 months to 14-18 months (SNV EDGET projectAnnualReport2016,P10).

The EDGET project faced some delays with regard to the distribution of supplementary calf feed,missingthetargetof32,500householdsfor2015.Thedelay,duetotheprocessofmappingsupplyanddemandofcalffeedinearly2015,entailedinventarisingcrossbredcalvesatthehouseholdlevel,andthenidentifyingcalffeedsuppliersthroughatenderingprocess.CalffeedwasthendistributedthroughthenetworkofAgIDsstartinginearly2016.By2017,theEDGETprojecthadsupportedthedistributionof14,176quintalofcalffeedto14,683households(30%oforiginaltarget).Inadditiontothedelaysin2015, thehigher thanexpectedcostof calf feed in2017alsomeant that coveragehad tobe scaledback

FindingsfromtheHHSurvey:

Aseriesofquestionswereincludedinthehouseholdsurveyfocusingspecificallyoncalffeed.78%ofrespondents intheinterventiongroupreportedowningcalvescomparedto68.6%inthecomparisongroup. Of these a total of 27.9% of intervention group farmers reported that they had ever usedsupplementarycalffeed,comparedto9.3%inthecomparisongroup.Farmerswhousedthecalffeedin both groups mostly reported that this had a very positive change (67.9% across both groups),followedbyanunsatisfactorychange(29.5%).78%offarmersintheinterventiongroupreportedthattheyplannedtocontinueusingsupplementarycalffeed,comparedto66.7%inthecomparisongroup.When those who did not plan to continue were asked to provide reasons for this, statisticallysignificantdifferencesinthedistributionofresponseswereobserved.Amongsttheinterventiongroup,respondentsweremorelikelythantheircomparisongroupcounterpartstocitelackofmoney(30.9%vs 22.9%) andwaiting for free supply (14.7% vs 7.6%). On the other hand, theywere less likely toreportlackofinformation(17.4%vs32.5%).

EVALUATIONREPORT

65

Findingsfromqualitativeassessmentbysub-componentTable19QualitativefindingsonAgroInputDealers


SelectionandcapacitydevelopmentofAgIDs

SNV has provided support to various types of AgID, including existingbusinesses,newbusinessesandcooperatives.

SNV’s support included technical training (on running a business andbookkeeping; paying commission for the distribution of calf-feed andequipment(MTS);andprovidingsomebasicequipment(e.g.shelves).

ThissupportwasperceivedbyAgIDstohaveledto:

● Better management of the business in terms of inventory,bookkeeping,andshoporganisation

● Morehygienichandlingofanimalfeed

● Abilitytopayrent(e.g.thankstocommissionfromEDGETproject)● Introductiontosuppliersandenhancedbargainingpower● Abilitytoassessqualityofinputsandensurequalityofproduce● Technical knowledge regarding cow feed and associated practices,

whichenabledAgIDstoprovidetechnicaladvicetofarmers● Supporttodiversificationofforagesupply

Thisinturnisseentohavecontributedtoincreasedprofitability(seefollowingsection).

From KII of DAB DRT national / regional stakeholders (Altaseb Mekbib Feedsupplier):

Calf-feedandMTS canbe stored for a long time inour store,without relyingtoomuchonframerscollecttheinputsontime.

In caseof SNV, staffwereperceived tobe veryefficient in theirperformancebecause they are directly located in the woreda to facilitate and supportstakeholders.ThisisamuchbettermodelthanotherNGOswhojustremaininAddis.

OneofthelargerfeedsupplierscommentedthattheremayhavebeenissueswiththeselectionofAgIDs,assomeofthosetheyliaisedwithappearedtobelackingintheneededfinancialcapacityandbusinessacumen.

CommissionAgIDsgetforsupplyofinputstofarmersistoolow(Wuchale)

AtleastoneoftheAgIDsreportedlydevelopedadependencyonEDGET,expectingthat theywould continue to receive commission from supplying project inputs tobeneficiariesanddiscontinuingthesupplyofcalffeedafterprojectsupport.

Linkageswithregionalfeedsuppliers

EDGET project supported the establishment of business to business (B2B)linkagesthroughintroductionsandthroughnetworkingeventswithagroinputprocessors/suppliersandexposurevisitsforAgIDs.

All three of the regional feed suppliers interviewed reported increased salesduetolinkageswithproducersandAgIDsandtheincreasedlevelofdemandatthefarmerlevelresultingfromtrainingsandawarenessraisingactivitiescarriedoutthroughEDGETproject.TheyalsocitedtheincreaseinthenumberofAgIDs(duetoEDGETprojectsupport)ascontributingtothisgrowthinbusiness.This

Delayswerefacedinthedistributionofcalffeed.

Oneof the feedsuppliers felt that theAgIDselectionhadsome issuesas someofthemwerequiteweakonbusinessandfinancialmanagement.

Increasesinpricesaredrivenbyincreasesinthepricesofrawmaterials.

Oneofthefeedsupplierscitedanumberofchallenges.Thesewererelatedto:

•Shortageofrawmaterials

EVALUATIONREPORT

66

Sub-component Strengths Weaknessesalsoledtoanincreasedinthenumberofagentsworkingforthem.Oneofthesuppliersreportedthat‘thewholeofSidamaandGedozoneshavebecomemyagents’.Anotherreported“upstreamwewerelinkedwithfeedprocessorsanddownstreamtoproducers.WearenowalsoknownintheWoredaandinotherregions in Ethiopia, like Tigray and Oromia. This a good opportunity for thefutureexpansionofourbusiness.”

Oneofthefeedsuppliersreportedanincreasedinthenumberofstaff.

Oneof the interviewed feedsuppliers reported that their salesofother (non-dairy)livestockfeed(e.g.forcattlefatteningandpoultry)viatheEDGETprojectsupportedAgIDshadincreased.

BothoftheinterviewedsuppliersreportedthattheB2Blinkagemeetingswereveryusefulforgettingtoknowtheotherstockholders.

One of the regional feed suppliers reported that the businesses would beaffectedbytheclosureoftheproject,theyfeltconfidentthatsufficientdemandhadbeencreatedtokeepthebusinessprofitable.

ThereareNGOs involved indairyworkingwithus,EDGETproject isunique increatingmarketlinkagewithproducersandAgID.

•Qualityofrawmaterialsispoor

•Increaseofpriceofrawmaterials

•Shortageofforeigncurrencyforimportofpremix

Oneofthefeedsuppliersreportedthattherecouldbearisktothebusinessiftheproject stops, since the same level of demand may not be maintained in theabsenceofthesamescaleandintensityofprojectactivities.

AnotherreporteddelaysinthecollectionofMTSandcalffeed.

Decline in the demand for forage seedwas also reported, though no reasonwascited.

AgIDoperations As presented above, the five EDGET project-supported AgIDs reportedbenefitingfromtheirinvolvementintheEDGETproject.Forthemostpart,theyappeartohavesuccessfullyexpandedtheirscaleofoperationsandtherangeofinputsthattheyprovide.Overall,theAgIDsfeelthatfarmersarehappywiththeproductstheyareselling(notethatthisviewinotalwayssharedbyfarmersandotheractors).

According tosecondarydatagathered fromtheAgIDs, threeof the fiveAgIDsreportedanincrease inthequantityofcalffeedconcentratesoldafterEDGETproject support. Four of the AgIDs reported an increased in the number offarmerscomingtobuycalffeedafterEDGETprojectsupport.

DairyrelatedinputsarethemainsourceofrevenueforfouroftheAgIDs,whilefor one (Lemu Bilbilo) it is crop related inputs. The AgID inMachakel is alsoinvolvedinbuyingmilk,whichisprovingtobeaprofitablebusinessforthem.

AlloftheAgIDsrecruitedadditionalstaffduringtheprojectperiod.

Someof thekeybenefits reportedbyotherstakeholders related to theAgIDsandtheirservicesaresummarisedbelow:

● Thereisgreateravailabilityofforage/feedatthekebelelevel;butisalsoaccessibletofarmerswhocangodirectlytotheshop.

● The AgID shop is open more and they are providing increasedquality, range and quantity of inputs (e.g. including ureas and

In some cases, price, quality and variety of feed supplied by the AgIDs are notperceived positively by some actors (DAs, DEPs, DFEGmembers).More generally,however,thepriceoffeedisoftenperceivedtobequitehighforsomefarmers(E.g.Machakel, Lemu Bilbilo, Dangila), particularly when considered in relation to theprice ofmilk. Increases in the prices charged by the raw input suppliers and theagro-dealers contributes to this (e.g. Lemu Bilbilo). At times, specific feed/inputssoughtbyfarmers-e.g.molasses,calffeed,etc.-arenotavailablewiththeAgID.

In some cases (e.g. Dangila) it was reported that the demand from farmers issomewhat weak or that some farmers do not buy their forage from the EDGET-supportedAgID(Machakel).

Inmanycases, farmers continue to relyonnon-EDGETprivate feedprovidersandtraders. Although they often provide door-stop service and their products arerelativelycheap,thequalityofthefeedisgenerallyperceivedtobequitepoor.

Transportationof feedto farmers,particularly those living farawayfromtheAgIDshopcanbeanissueinsomecases(e.g.LemuBilbilo).InMachakel,coveragebytheSNV-supportedAgIDwasseentobeweak.

Some respondents reported that certain AgIDs were not strictly following theSNV/EDGETprojectguidelines-e.g.regardingwarehouse,shelvingandstorage(e.g.LemuBilbilo,feedproduction,Dangila).Thisisseenashavinganegativeimpactonthe quality of the feed mixes which has financial and potentially reputational

EVALUATIONREPORT

67

Sub-component Strengths Weaknessesmolasses)

● TheAgIDprovidesfeedoncredit● In some cases the AgIDs are not only providing feed but also

technicaladvicetofarmersonuseofthefeed.● Theavailabilityof improvedforage/feedhasalso ledtoan increase

inthepracticeofstallfeedingandzero-grazingDFEGleadersreportedsatisfactionwiththeavailabilityoffeed.

Theintroductionofcalf-feedhasalsobeenanimportantinnovation.

There are also a number of other providers of forage/feed, includingconcentrate, molasses, urea, etc., operating in the woredas who are notsupportedbyEDGET.

implicationsfortheAgIDs.Forexample,theAIDinDangilareportedlosingmoneyasaresultofhavingtodisposeofdamagedfeed.

AnumberofAgIDsalsonotedthathightaxeslimitthemarginsthatAgIDscanmake.

OtherissuesnotedbyprojectstaffincludethehighrentalcostoftheshopusedbytheAgIDandinstancesofnon-targetedfarmerstryingtotakeconcentratecalffeedthatwastargetedforDFEGmembers.

EVALUATIONREPORT

68


Feedsupply isaprimaryconstraint fordairy farmersandthedairyvaluechain,beingoneofthekeydeterminantsofmilkproductioninmilkingcowsandhavingstronglinkageswithcalfdevelopmentandthe age at which calves become fertile. The prevailing context in terms of feed supply can becharacterisedbylowendtraderswithcheaperproductsthatarewidelyperceivedtobeofpoorqualityandhighendsupplierswithgoodproductsthatareprohibitivelyexpensiveforsmall-scalefarmers.Assuch,theAgIDcomponentappearstofillagapinthemarketbyprovidingbetterqualityfeed(thanthetraders) at amore affordableprice (than the established, high end feedbusinesses - such asAlemaKoudijs).


TheEDGETprojectappearstohavebeenquiteeffectiveintheestablishmentandfurtherdevelopmentofAgIDs,usingavarietyofappropriateinterventionstoachievethis.

• Businesstobusiness linkages-alltheAgIDs interviewedfoundtheB2Bnetworkingeventstobeaveryusefulmeansofstrikingbusinessdealswithsuppliersofrawmaterialsatcompetitiveprices. In somecases,groupsofAgIDsoperating in similargeographiesare reported tohavecollaboratedtosecuredeliveriesof inputsatevenmorecompetitiveprices.Thisapproachtonetworking emerges as a strong,market-based (i.e. facilitative)means of strengthening theinputsupplyofthedairyvaluechain.

• TechnicalandbusinesstrainingswerealsoperceivedpositivelybytheAIDswhoreceivedthem.They reported increased ability tomanage their stocks, handle feedproperly, organise theirshopsbetter, and to run their businessmore effectively (including through improved recordkeeping).

• AgIDsasadistributionchannel (establishinga relationshipwithcustomersandadistributionnetwork)

AlloftheAgIDsinterviewedreportedthattheirbusinessesweregrowingsuccessfullyintermsofscaleof production, expanding customer base and profitability. They envision growing demand for theirinputsinthefuturefollowingthefurtherdevelopmentofthedairyvaluechain(increasedprevalenceofcross-breeds,growingappreciationofdairyasaviablefarmbusiness/livelihoodoption,etc.).

However, a set of key issues related to either actual or perceived problemswith the quality, price,variety and availability of inputs provided by agro-input dealers do indicate some limitations in theextent to which the AgIDs are proving to be an effective mechanism for meeting the demand forsufficientquantity,qualityanddiversityoffeed.Otherconstraintstofurtherbusinessdevelopmentwillalsoneedtobeconsidered


Themarket-basednatureof thiscomponentandtheconsiderablesuccessthathasbeenachievedtodatebodeswellforthesustainabilityofthisapproach.Withmarket-basedinterventionsitisimportanttorecallthatmarketdynamicscanbeunpredictableandthatwhilethesuccessorfailureofindividual

EVALUATIONREPORT

69

businessesmayvaryconsiderably, the long-termconcern iswiththeestablishmentof functionalanddynamicmarketsthatareabletoadaptinordertorespondtochangingcircumstancesandthenatureandpatternofdemand.

OveralltheAgIDcomponentappearstoholdconsiderablepotentialintermsofsustainability.• Evidence illustrates a viable business proposition for SMEs to fill a genuine gap (between

tradersandlarger,well-establishedproviders)inthemarketforfeed.• B2Bnetworkingisaprovenlow-costsustainableapproachtofosteringmarketdevelopment.• AgIDsare innovating, addingadditionalproduct lines (e.g. forage seed)and serviceofferings

(e.g.milkcollection)totheirbusinesseswhichbodeswell.


• Longer-termdynamicswithinthefeedmarket.WhilethesupporttoindividualAgIDsthroughtheprojecthasbeencommendable indemonstratingtheviabilityofSME-typeAgIDsfocusedonthedairysectorinthetargetgeographies,thisapproachmaynotproveviableforEDGETatscale.Assuch,itwouldbeusefulforEDGETtounderstandthemainbarriersandincentivestoentryandgrowthfornewandexistingAgIDsandtofocusoncreatingtheconditionsthatwillenableAgIDsthatservethedairysectortomultiplyandthrive.

• Key constraints to further business development, particularly access to finance/credit wereraised in anumberof cases. This limits theabilityofAgIDs toupgrade theirwarehouses forsafestorageof feedsand for increasing thecapacityof thewarehouse. In the future furthereffortsmayberequiredtofacilitateaccesstofinance/credit.

• Getting thepricing rightcontinues topresentachallengeasDFEGmembers frequentlycitedhighpricesasabarriertopurchasingmorefeed.

• Farmersarefrequentlyconcernedthattheydonothaveaccesstoqualityfeed.Activatingthegovernment’s role in checking quality and price of feed products (for example, through acertification system) couldplayan important role inmanaging issuesofmistrust in the feedsector.

• Inadditiontotheabove,theexistingsuiteofactivitiestargetingAgIDs–particularlybusinesstobusiness linkagesand technical trainings shouldbecontinued. If itdoesnotalreadyexist,thenanindustry/commercialassociationofdairy-focusedAgIDsandAgroInputSupplierscouldalsobeestablishedforbroadergovernanceanddevelopmentofthesector.

Householdadoptionofinputsandpractices

Drawing on evidence from the household survey, this section presents key differences between theinterventionandcomparisongroupsattheendline.Comparisonwithbaselinedatawasonlydoneforselectedpracticesinthissectionwhereweareabletousethesamequestionstructureasthebaselineor else transform thebaseline data in amanner suitable for comparison.Where a comparisonwithbaselinedata isnotpossible,dataon ‘fromwhomdidyou learnthispractice’,and ‘sincewhenhaveyouadoptedthispractice’areusedtoinfertheprojectcontribution.

EVALUATIONREPORT

70

Improvedforageandanimalfeed

Differences in theuseofdifferent typesofanimal feedwere,however,much lesspronounced.Hay(80%), crop residue (64-71%)are themostused improved feeds, followedbyby-productsof cerealsandoilseeds(25%).Multi-nutrientblocks,molasses,mixtureofforageorcerealbransareusedbylessthan 12% respectively. However, statistically significant differences were only found between thegroupsintheuseofforagecrops.

The EDGET project sought to bring about changes in the feeding practices including zero-grazingfeedingforcross-bredandlocalcows,useofacutandcarrysystemforfeed,concentratesupplementsfor pregnant andmilking cows, and similar practices. For intervention group farmerswith crossbredcows(ncomp=164,ninter=387)theadoptionrateofzero-grazing ismarginally lessthanforcomparisongroup farmers (19% vs 26%). However, intervention group farmers were marginally more likely topracticemainlygrazingwithsomestallfeeding(24%vs16%).Around15%inbothgroupsreportedonlygrazingfortheircrossbredcows.

Looking however into the reasons given for the adoption of zero-grazing, we can ascertain thatdifferent feedingpracticesmoreoftenderive fromtraining (referringmainly to trainings fromAgIDs and/or cooperatives supported by SNV/EDGET)andtoamuchlesserextentpersonalexperience.Forcomparison group farmers it is the other way around - personal experience is cited much morefrequentlythantrainingreceived.

Table20Overviewofadoptionoffeedingpractices22

Baseline Endline

Comparison(n=400)


Comparison(n=220)

Intervention(n=432)

%ofhouseholdsadoptingzerograzing(cross-bredcows)b 25.6% 18.9%

%householdsusingcutandcarrysystemforfeeda 79.4% 95.9% 56.6% 61.6%

%of householdsusing concentrate supplements for yourpregnantandmilkingcowsandheifersa

25.3% 49.9% 22.3% 29.7%

%ofhouseholdspreparingyourownimprovedfeedssuchasureastrawtreatment,silage,multi-nutrientblockb

- 11.6% 25.6%

%of households varying the feeding depending on stageoflactationa,b

61.2% 43.4% 20.2% 34.9%

%ofhouseholdsmonitoringcows’productiona 67.1% 51.1% 23.3% 31.4%

%ofhouseholdsprovidingenoughwatertocattlea 92.1% 96.8% 58.4% 66.3%a Significant differences in mean averages or cell distribution for baseline intervention and comparison, Chi-square teststatisticorTwo-sampleT-Test,p<0.05bSignificantdifferencesinmeanaveragesorcelldistributionforendlineinterventionandcomparison,Chi-squareteststatisticorTwo-sampleT-Test,p<0.05

22Pleasenotethatbaselinemeasureswerenotasked inthesamewayas inendlineandcomparisonsovertimearehencedifficult.

EVALUATIONREPORT

71

Withregardtootherfeedingpractices,themostwidelyusedpracticesarecutandcarrysystems(57%comparison vs. 62% intervention) and providing enough water for cattle (58% comparison vs. 66%intervention).Table20showsthatsignificantdifferencesbetweeninterventionandcomparisongroupfarmers exist only for ‘own improved feed preparation’ (12% comparison vs. 26% intervention) and‘varying feeding depending on lactation’ (20% comparison vs. 35% intervention). Adoption rates forthesepracticesarehigheramongsthouseholdsintheinterventiongroupthaninthecomparisongroup(seeFigure9).

Respondentswereaskedtoreportthenumberofyearsthattheyhadbeenadoptingeachpractice.Theaverage number of years of adoptionwas 3 years for ‘own improved feed preparation’ and over 4years for ‘varying feeding depending on lactation’ for both comparison and intervention groups.LookingatFigure10wesee that3years followedby2years ismost frequentlymentioned forbothpractices.Thissuggeststhatthemaincauseforthesechangeslieswithintheprojectperiod.

Figure 9 Bar-plot with interval estimate of population proportions (CI 95%) for adoption of feeding

practicesincomparisonandinterventiongroup

EVALUATIONREPORT

72

Figure 10 Histogram showing the count of years since adoption of practices for which we found

significantdifferencesbetweencomparisonandinterventiongroup(interventiongrouponlydisplayed

here)

Wealsoreportbaselinefigures(seeTable20)onthefeedingpracticeswhichareinallinstanceshigherthanendlinefigures.Wedonotsuspectthatasmallerproportionoffarmers isusingthesepracticesnow,butthatthewaythesequestionswereaskedwasdifferentorthatthereareerrors inthedata.Thisiswhywedonotreportneteffectsfromdifference-in-differenceanalysishere.

Milkingandmilktransportation

Usageofmilkingequipment isverycommonamongbothcomparisonand intervention farmers.Wecansee10%morecomparisongroupfarmersusingmilkstorageequipmentandcirca20%moreoftheintervention farmers usingmilk transportation equipment (see Table 21 and Figure 11). In terms ofwhatequipment isused, claypotsandplastic jarsare citedmost frequently formilking, storingandtransport.Therearenosubstantialdifferencesbetweencomparisonandinterventiongroup(seeTable21below).

Table21Overviewofequipmentusedbycomparisonandinterventiongroupsforstorage,milkingand

transport

Comparison(storage/transport/milking)

Intervention(storage/transport/milking)

Claypot/Gourd(Kill) 37%/13%/45% 36%/22%/45%

Plasticjar/vessel23 59%/79%/52% 63%/72%/54%

Other 4%/8%/2% 1%/6%/1%

23manufacturedlocallyforthepurposeofliquidcontainerincludingmilk

EVALUATIONREPORT

73

Figure 11 Bar-plot with interval estimate of population proportions (CI 95%) for usage of milking

equipmentincomparisonandinterventiongroup

Basedontheendlinesurvey,64%ofinterventiongroupfarmersreceivedtheMTSatleastonceduringthecourseof theproject.67%of interventiongroup farmers (i.e.277 farmers fromthe interventiongroup)reportedusingtheMTSformilkingand47%reportedusingitfortransportation.

Regardinghygienicmilking practices, we can see an increase in the adoption of practices over theprojectperiodforboth,comparisonandinterventiongroup.Wecanseethatatbaselinealready,thereweresignificantdifferencesbetweengroupswith regard tocleaningofmilkingequipmentwithsoap(34% comparison vs. 40% intervention), cleaning hands after milking (33% comparison vs. 42%intervention)andcleaningofcows’teatbefore/aftermilking(2%comparisonvs.10%intervention).At endline we find significant differences for cleaning hands before (52% comparison vs. 63%intervention)andaftermilking(46%comparisonvs.55%intervention)aswellascleaningthemilkingarea(63%vs.77%).

EVALUATIONREPORT

74

Table22Overviewofadoptionofhygienicmilkingpractices24

Baseline Endline

Comparison(n=400)


Comparison(n=220)

Intervention(n=432)

% of households cleaning milking equipment andutensilswithsoapa

33.9% 39.7% 48.2% 48.3%

% of households cleaning of hands aftermilkingwithsoapa,b

32.5% 42.1% 46.4% 54.6%

%ofhouseholdscleaningofhandsbeforemilkingwithsoapa,b

27.4% 40.1% 52.3% 63.3%

%ofhouseholdscleaningofmilkingareac,d 1.4% 4.8% 63.2% 77.4%

% of households cleaning the cows' teat before andaftermilkingwithsoapa

1.7% 10.0% 35.0% 36.0%

a Significant differences in mean averages or cell distribution for baseline intervention and comparison, Chi-square teststatisticorTwo-sampleT-Test,p<0.05bSignificantdifferencesinmeanaveragesorcelldistributionforendlineinterventionandcomparison,Chi-squareteststatisticorTwo-sampleT-Test,p<0.05cSignificantdifferencesinmeanaveragesorcelldistributionbetweenbaselineandendlineforinterventiongroup,Chi-squareteststatisticorTwo-sampleT-Test,p<0.05dSignificantdifferencesinmeanaveragesorcelldistributionbetweenbaselineandendlineforcomparisongroup,Chi-squareteststatisticorTwo-sampleT-Test,p<0.05

Hygienic milking practices are performed every day rather than after every milking. But there aredifferencesbetweeninterventionandcomparisongroup:Interventiongroupfarmersperformhygienicmilkingpracticestoagreaterextentaftereverymilkingthancomparisongroupfarmers.

We calculated net effects for intervention using difference-in-difference analysis looking at baselineandendlinedata.Apositiveneteffectof10.8%canbe found for cleaningof themilkingarea.Withregardtootherhygienicmilkingpracticesthenet interventioneffect isnegative,butnotsubstantive(rangingfromareductionof1%to7%).

24Pleasenotethatbaselinemeasureswerenotalwaysaskedinthesamewayasinendlineandcomparisonsovertimearedifficult.

EVALUATIONREPORT

75

Figure12Bar-plotwith interval estimateofpopulationproportions (CI95%) foradoptionofhygienic

practicesincomparisonandinterventiongroup

Looking intothereasonsbehindadoptionoftheabovementionedpractices,wecanascertainfortheintervention group that theymore often derive from training and to amuch lesser extent personalexperience. For comparison group farmers it is the otherway around - personal experience is citedmuch more frequently than training received. In terms of non-adoption of hygienic practices, lackknowledgeortechnicalskillsisbyfarthemostcitedreason.

Forpracticeswith statistically significantdifferences inadoption ratesbetween the interventionandcomparisongroups(namely,washinghandsbeforeandaftermilking,aswellascleaningmilkingarea),2, 3 and 4 years since adoption were the most frequently cited. This lies within the projectimplementationperiod(seeFigure13below).

EVALUATIONREPORT

76


significantdifferencesbetweencomparisonandinterventiongroup

Animalhealth

Asforanimalhealthpractices,wecanseeamajorityoffarmersinbothinterventionandcomparisongroupsconsultingaveterinarianforprevention(57%),treatingsickanimalswithantibiotics(62%)andundertaking regular vaccination (85%). It is only in the case of usage of antibiotics that we see astatistically significant difference in the adoption, with 66% for intervention farmers and 58% forcomparisongroup.

EVALUATIONREPORT

77

Figure 14 Bar-plot with interval estimate of population proportions (CI 95%) for adoption of animal

healthpracticesincomparisonandinterventiongroup

Table23Overviewofadoptionofanimalpractices25

Endline


% of households consulting a veterinarian to control, prevent andtreatdiseases 54.1% 58.9%

%ofhouseholdsdewormingeachofdairycowatleastonceperyear 38.2% 42.5%

%ofhouseholdstreatingsickanimalswithantibioticsb 57.7% 65.8%

% of households undertaking regular vaccination programs forinfectiousdiseases 83.6% 85.5%

Calfmanagement

Forcalfmanagementpractices(calffeedingpracticesarediscussedintheAgIDcomponentoftheSO1section)thereareupto30%moreinterventionfarmersallowingthecalftosucklethemother(34%vs.53%),carryingoutheartgirthmeasurementsandeartagapplications (4%vs.33%),cleaningthecalfafterdelivery(41%vs.50%),feedingconcentratefeedtothecalf(25%vs.37%),bucketfeeding(23%

25Pleasenotethatbaselinemeasureswerenotavailableforthesepractices.

EVALUATIONREPORT

78

vs.29%)andvaryingtheamountofmilkprovidedtothecalfwhenbucketfeeding(21%vs.30%,seeFigure15).Allofthesedifferences(exceptpracticingbucketfeeding)arestatisticallysignificant.

Table24Overviewofadoptionofcalfmanagementpractices26

Endline


%ofhouseholdsallowingthecalftosucklethemotherb 33.8% 52.6%

% of households carrying out heart girth measurements and ear tagapplicationsb 4.0% 33.2%

%ofhouseholdscleaningthecalfimmediatelyafterdeliveryb 30.5% 50.2%

%ofhouseholdsfeedingthecalfwithconcentrateb 25.2% 37.2%

%ofhouseholdspracticebucketfeeding 22.5% 28.5%

% of households varying the volume of milk depending on the stage oflactation(ifbucketfeeding)b 21.2% 30.1%


Amongst the intervention group, reasons given for the adoption of these practices are most oftentrainingandtoamuchlesserextentpersonalexperience.Forcomparisongroupfarmersitistheotherway around - personal experience is cited muchmore frequently than training received. Coaching,learningfromotherfarmersandincomeopportunitiesareveryrarelycitedasreasonsforadoptionofcalf management practices. When asking more specifically about who provided the training,intervention farmers cite government extension providers (DAs, woreda experts, etc.), AgIDs, DFEGleadersandother farmerswhereascomparisongroupfarmersmainlyrefertogovernmentextensionproviders and other farmers. When asked about reasons for non-adoption, both intervention andcomparison group farmersweremost likely to cite lack of knowledge (ca. 60-80% of non-adopters)followedbylackofrequiredinputs(15-35%).


EVALUATIONREPORT

79

Figure 15 Bar-plot with interval estimate of population proportions (CI 95%) for adoption of calf

managementpracticesincomparisonandinterventiongroup

Looking intowhen calfmanagement practiceswith significant differences between comparison andintervention groups were adopted, Figure 16 below shows that 2 and 3 years ago - i.e. within theimplementationperiodoftheproject-werecitedmostfrequently.

EVALUATIONREPORT

80



Housingandmanuremanagement

In terms of housing and manure management, more than two thirds - in both intervention andcomparison groups - keep their cattle in confined and clean areas, and that more than 40% keepdifferent types of cattle in separate housing. Only about 20-25% use a constructed feeding troughaccessible from inside and outside. Statistically significant differences between comparison andintervention farmers exist for ‘providing adequate ventilation and lighting for cows’ barn’ (39% vs.53%)andadequatestorageofmanureforcropapplication(25%vs.34%).

EVALUATIONREPORT

81

Table25Overviewofadoptionofhousingandmanuremanagementpractices27

Endline


%ofhouseholdskeepingcattleinconfinedareafreefrommudandmanure 65.5% 70.8%

%ofhouseholdsstoringmanureadequatelyandthenapplyittocropsb 24.5% 34.0%

%ofhouseholdshavingseparatehousingtypefordifferentcattletype 41.4% 45.8%

%ofhouseholdshavingcowsbarnwithadequateventilationandlightingb 39.1% 53.0%

%ofhouseholdsusingconstructedfeedingtrough(insideandoutside) 25.5% 20.6%


Reasons for adoption of practices are most often training over personal experiences for theinterventionfarmersandviceversaforcomparisongroupfarmers.Whenaskedaboutnon-adoption,both, interventionandcomparisongroup farmers refer toa lackofknowledge, rather thana lackofrequired inputsor lackofadviceandfollow-upsupport.Asmallerproportionalsoclaimedhighcosts(<10%)asareasonfornon-adoptionofthedifferentpractices.

Figure 17Bar-plotwith interval estimateof populationproportions (CI 95%) for adoptionof housing

andmanurepracticesincomparisonandinterventiongroup


EVALUATIONREPORT

82

Looking into when practices were adopted (in which we found significant differences betweencomparisonandinterventiongroup),Figure18belowshowsthat3yearsagowascitedmostfrequentlyforhousingandmanuremanagementpractices.ThisiswithintheimplementationperiodoftheEDGETproject.



Climatesmartpractices

With regard to climate smart practices, a very small proportion of farmerswere found to be usingbiogas (6% comparison vs. 3% intervention) while a much larger proportion reported enrichinglivestockfeedwithagriculturalbyproducts(49%comparisonvs.44%intervention),andusingmanureto fertilize the farm (61% comparison vs. 71% intervention). There were no statistically significantdifferences detected between the groups. Other practices, such as the cultivation of green forage,adoptionof the cut-and-carry systemandamoveaway fromopengrazingalso represent importantstepstowardmoreclimatesmartdairypractices.

EVALUATIONREPORT

83

Figure 19 Bar-plotwith interval estimate of population proportions (CI 95%) for adoption of climate

smartpracticesincomparisonandinterventiongroup

Table26Overviewofadoptionofclimatesmartpractices28

Endline


%ofhouseholdsusingmanuretofertilizethefarm 60.9% 71.1%

%ofhouseholdsenrichinglivestockfeedwithagriculturalby-products 38.6% 43.5%

%ofhouseholdsusingbiogas 5.9% 3.2%

Insights from the qualitative data collection suggest that across actors and woredas there is arecognitionof increasedawarenessaboutbiogas.Somefarmersstartedusingbiogasandcompostingforfertilizermethodsduetothetraining.

Summaryforadoptionofpractices

Atotalof34differentpracticesrelatedtoanimalhealth,hygienicmilkingpractices,forageproduction,housing and manure management, calf management and feeding were assessed. Statisticallysignificantdifferenceswerefoundforatotalof16specificpractices,i.e.whereahigherproportionofinterventionthancomparisongroupfarmerswerefoundtohavereportedadoptionof thepractices.Thesepracticesinclude,forinstance:farmersengagedinforageproduction,forageseedmultiplication,


EVALUATIONREPORT

84

cleaningofhandsbefore/aftermilkingwithsoap,measuringheartgirthofthecalfandprovidingcalveswithconcentratefeed.Generalizationstothepopulationofallfarmerscanonlybemadeforthese16practicesforwhichdifferenceswerefoundtobestatisticallysignificant.29

Thedatashowsthatalthoughthemeannumberofyearsapracticewasadoptedisgreaterthan3,themajorityof adopters reportedadopting thepracticewithin the last 3-4 years, i.e.within theprojectimplementation period. Themean, however, is skewed upwards by a smaller proportion of farmerswhohaveadopted thepractice for a largernumberof years. Theoverview figurebelowalso showsthat the median, which separates a sample into its higher and lower half, is often lower for theinterventiongroupthanforthecomparisongroupandthatitissmallerthanfiveyearsforallpracticesintheinterventiongroup.

Figure 20Median year since adoption of practices for respondents for comparison and intervention

group. Only showing practices in which we found significant differences. The median is the value

separatingthehigherhalfofadatasamplefromthelowerhalf.

Acloserlookatthereasonsforadoptionrevealsthat‘training’iscitedmuchmoreoftenand‘personalexperience’ toamuch lesserextent for the interventiongroup.This is theotherwayaround for thecomparisongroup.This suggests that theprojectwasa significantdriverofadoptions in theproject

29Pleasenotethatwevisualised ineachFigurearangewithinwhichthe ‘true’populationestimate lies. Itmaynotbetheexactpointfigureasreportedinthetableabove.

EVALUATIONREPORT

85

area. Looking into the providers behind the training, EDGET project supported actors (e.g.AgIDs/Cooperatives) werementionedbytheinterventiongroupbutnotbythecomparisongroup.

Milkproduction

Table27:Milkproduction

a Significant differences in mean averages or cell distribution for baseline intervention and comparison, Chi-square teststatisticorTwo-sampleT-Test,p<0.05cSignificantdifferencesinmeanaveragesorcelldistributionbetweenbaselineandendlineforinterventiongroup,Chi-squareteststatisticorTwo-sampleT-Test,p<0.05dSignificantdifferencesinmeanaveragesorcelldistributionbetweenbaselineandendlineforcomparisongroup,Chi-squareteststatisticorTwo-sampleT-Test,p<0.05

Bytheendoftheproject,householdsfrominterventionandcomparisonworedaswerefoundtohavean averagemilk production of 953 and 1068 litres respectively (the differencewas not found to bestatisticallysignificant).Forthesubsetofhouseholdsthathadmilkingcows,thefiguresarehigher,butmoresoforcomparisonvillages.Thedifferencesbetweenthesefigureswerenot,however,foundtobestatisticallysignificantattheα=95%level.Analysisatthelevelofindividualcrossbredcowsshowsasimilarpatternasatthehouseholdlevel,statisticallysignificantdifferencesbetweenthegroupsexistonlyatthebaseline.

Looking at data onmilk production by sex of the household head at endline, there is a noticeablyhighervolumeofmilkproductionfor female-headedhouseholds (1523 litresperyear) thanformaleheadedhouseholds(1100litresperyear).Thisdifferenceisstatisticallysignificant.

DIDestimationmilkproduction

We calculated net effects for the intervention using difference-in-difference analysis looking atbaselineandendlinedataformilkproductionperhouseholdandmilkproductionperhouseholdswitha milking cow. As the results for the two groups are not substantially different, the analysis herefocuses only on milk production per household. While overall milk production increased betweenbaseline and endline for both groups, taking into account the different starting points for milkproduction between comparison and treatment at baseline, we see that milk production in the

Baseline Endline

Comparison(n=400)


Comparison(n=220)

Intervention(n=432)

Averageannualmilkproductionpercow(onlyforcrossbreedmilkingcows)a 519.2 971.1 1208.4 1001.9

Averageannualmilkproductionpercow(onlyforlocalmilkingcows)a 428.3 455.3 332.9 359.2

Averageannualtotalmilkproduction(litres)perHHa 259.3 438.3 1067.8 952.6

Averageannualtotalmilkproduction(litres)perHHthatproducedmilk

a,c,d 437.6 655.8 1243.0 1098.7

EVALUATIONREPORT

86

comparisongroup increasedbya largeramount than for interventiongroup30.TheneteffectorDIDestimate for being in the intervention group is significant and negative with milk production goingdownby294litres.Duetotheviolationofnecessarypre-conditionsforperformingDID-estimationandissueswiththebaselinedata,theseresultsshouldbetreatedwithgreatcaution.SeeandFigure21formoreinformation.31

Figure 21 Differences between groups and over time for milk production per household. The 95%

confidenceintervalsdepicttherangeofthemeanaveragemilkproductioninthepopulation

Perceivedchangesinmilkproduction

Lookingatperceivedchangesinmilkproductionoverthelastyears,38%forcomparisonand47%forinterventiongroupsaiditincreased,34%versus31%saiditstayedthesameand28%versus22%saiditdecreased(differencessignificantat10%level).Thesedifferenceswerenotfoundtobestatisticallysignificant at95%.Reasons for increasewerebirthof calves (88%vs. 92%)andpurchaseof animals(12.2vs.4.4%).Reasonsfordecreaseweresaleofcalves(38%vs.39%),deathofanimals(43%vs.51%andother(17%vs.11%).

30Weestimatedasignificantpositivemaineffectfortime(i.e.overallmilkproductionisgoingupby809litresfrombase-toendline forbothgroups)and forgroupmembership (i.e.milkproductiononaverage is179 litreshigherwhenbeing in theinterventiongroup)31InordertoestimatetheDIDorneteffect,pre-conditionssuchashomogeneityofvariancesandnormaldistributionhavetobecheckedfor.Thesepre-conditionsarenotestablishedformilkproduction.Thereareseveraloutliervaluesandvariancesofmilkproductionbetweenbase-andendlinearenotequal.Thismaybedueto the factofsamplinganewatendlineratherthanhavinga longitudinalapproachwhere the samehouseholdsare surveyedatbase-andendline. Itmayalsobedue todifferentcleaningofbaselineandendlinedata.

EVALUATIONREPORT

87

4.3 Strategic Objective 2: To increase processing and marketing ofdairyproductsStrategicobjective2focusedonincreasingtheprocessingandmarketingofdairyproductsbothatthehouseholdlevelandbroadlywithinthedairyvaluechain.WhileStrategicObjective1providesthebasisforhouseholds toachievehigher levelsofmilkproduction,StrategicObjective2 focusesonenablinghouseholdstoearnhigherincomeseitherbysellingtheirmilktoamoreremunerativemarketorelsebyproducingprocessed/value-addeddairyproductsthatcanbesoldatapremium.Thissectionbeginswith a presentation of findings from the baseline and endline surveys related to the sale andprocessing of milk and milk products and then proceeds with an analysis of EDGET project’sinterventions to support and strengthen Dairy Processing. The diagram below provides a high-leveloverviewofthetimelineofactivitiesforStrategicObjective2.

Figure22TimelineforimplementationofSO2

Householdmilkprocessing,consumptionandsale

Milkprocessing

The household surveys at baseline and endline gathered a variety of data on household level milkprocessing. This covered a proportion of households involved in processing butter, cottage cheese,souredmilkandthequantitiesproducedandsoldofeachprocessedproduct.

EVALUATIONREPORT

88

Table28%Householdsproducingdifferenttypesofdairyproducts32

Baseline Endline

Comparison Intervention Comparison Intervention

Butter(%ofHH) 54.3% 53.3% 64.1% 66.7%

Cottagecheese/Ayib(%ofHH) 12.0% 20.2% 49.1% 46.5%

Souredmilk(%ofHH)a,b 45.5% 36.5% 27.2% 38.7%a Significant differences in mean averages or cell distribution for baseline intervention and comparison, Chi-square teststatisticorTwo-sampleT-Test,p<0.05bSignificantdifferencesinmeanaveragesorcelldistributionforendlineinterventionandcomparison,Chi-squareteststatisticorTwo-sampleT-Test,p<0.05

Endline analysis illustrated that the proportion of households involved in processing milk was verysimilar inboth interventionandcomparisonvillages (82%and86%respectively),withnostatisticallysignificantdifferencebetweenthem.Overall,over60%ofhouseholdsreportedproducingbutter,withno statistically significantdifferencebetween comparisonand interventionworedas. Cottage cheesewas produced by almost half of comparison and intervention group households. Comparisonhouseholds were less likely to produce soured milk (27% vs. 39%). Overall the proportion ofintervention group households producing all three products increased between the baseline andendline,thoughsubstantiallyonlyforbutterandcottagecheese.Theproportionofcomparisongroupfarmers processing butter (54% to 64%) and cottage cheese (12% to 49%) increased over time butdecreasedforsouredmilk(46%to27%).

Looking at the quantities produced (kg per household per year) of each type of dairy product, nostatistically significant differences were detected between comparison and intervention groups atendline. But significant differences in annual production were found at the baseline betweencomparisonandinterventiongroupforcottagecheese(39kgvs.79kg)andsouredmilk(66kgvs.103kg). Differences between the baseline and endline data could not be tested due to problems withmergingdatasets.

a Significant differences in mean averages or cell distribution for baseline intervention and comparison, Chi-square teststatisticorTwo-sampleT-Test,p<0.05bSignificantdifferencesinmeanaveragesorcelldistributionforendlineinterventionandcomparison,Chi-squareteststatisticorTwo-sampleT-Test,p<0.05

32Onlyhouseholdsreportingaquantitygreaterthan0werecountedasproducers.Proportionsarebasedoncompletesamplesize(withoutNAs)forsubgroups.Significancestatisticscouldnotbeconductedforbaselineandendlinecomparisonsduetodifficultiesinmergingbaselineandendlinedata.

EVALUATIONREPORT

89

Table29Averagequantityofdifferenttypesofprocesseddairyproductsproducedbyhousehold33

Baseline Endline


Butter(kg/HH) 27.2 32.6 38 37

Cottagecheese(kg/HH)a 39.0 78.6 78 77

SouredMilk(kg/HH)a 65.93 102.09 167 147a Significant differences in mean averages or cell distribution for baseline intervention and comparison, Chi-square teststatisticorTwo-sampleT-Test,p<0.05

Perceivedchangeinprocessingofdairyproducts

Whenaskedaboutperceivedchanges intheprocessingofmilkproducts,thedifference inresponseswas found to be statistically significant. In particular, households in the interventionworedasweremore likely to report an increase (51%) than households in the comparison group (39%), andhouseholdsinthecomparisongroupwerealmost10%morelikelytoreportthatproductionhadstayedthesame.

Saleofrawmilk

Accordingtotheendlinesurvey,32.4%ofinterventiongrouphouseholdshadsoldmilkascomparedto21.5%ofcomparisongrouphouseholds.Forthebaselinethiswas29.5%forcomparisonand51%forinterventiongroup.

Of the households who reported selling raw milk, households in the intervention woredas weremarginallymore likely to report selling to individuals (49% vs. 40%) and traders (22% vs. 20%) thanhouseholds in the comparisonworedas.Households in the comparisonworedasweremore likely toselltoaprivatecompany(15%vs.5%).Amuchsmallerdifferencewasfoundintheproportionsellingto cooperatives, with 24.5% of those in the comparison group selling to cooperatives compared to21.6% in the interventionworedas.Of thoseselling tocooperatives in the interventionworedas, themajority(around60%)reportedsellingtheirmilktotheDPUcooperativessupportedbyEDGET.

On average, no significant difference was found in terms of the volume of raw milk sold by theintervention and comparison groups (1524.4 litres vs. 1505.3 litres on average). Average pricesobtained during both the fasting and non-fasting seasons were also found to have little variation,generallybeingintherangeof10.9to11.4Birrperlitre.

33Significancestatisticscouldnotbeconductedforbaselineandendlinecomparisonsduetodifficultiesinmergingbaselineandendlinedata.

EVALUATIONREPORT

90

Saleofprocesseddairyproducts

Table30%ofhouseholdsinvolvedinsaleofrawmilkandprocesseddairyproducts

Baseline Endline


Rawmilk(%ofHH)b 23.3% 34.0% 21.5% 32.4%

Butter(%ofHH)b 47.5% 17.5% 62.7% 53.7%

Cottagecheese(%ofHH)b 35.0% 4.5% 22.2% 14.3%

Yoghurt(%ofHH) 6.0% 0.7% 1.6% 4.1%


Butterwasthemostfrequentlysoldmilkproduct(comparisonn=89,interventionn=154),followedbycottage cheese (comparison n=24, intervention n=30). Only 1 household in the comparison groupreportedsellingsouredmilkcomparedto6intheinterventiongroup.

Table31Quantityofdairyproductssold34

Baseline Endline

Product Comparison Intervention Comparison Intervention

Rawmilk(litressold)a 540.2 852.0 1505.3 1524.4

Butter(kgsold) 64.5 33.9 39.6 34.3

Cottagecheese(kgsold)a 118.6 35.8 67.8 91.8

Souredmilk(kgsold) 98.5 47.4 48.0 254.3a Significant differences in mean averages or cell distribution for baseline intervention and comparison, Chi-square teststatisticorTwo-sampleT-Test,p<0.05

Overallwhothehouseholdssoldtheseproductstodidnotvarysignificantlybetweencomparisonandintervention group households. While there were some differences in who these processed milkproductswere sold to, the numbers of households involved in sale of thesemilk productswas toosmall todetect any statistically significantdifferences in thepricespaidbydifferentbuyers.Overall,butterwassoldatapriceof130Birrperkg,cottagecheeseat46Birrperkgandsouredmilkat13Birrper kg. None of the respondents reported selling any of these products to a cooperative/DPUsupportedbytheEDGETproject.Ratherthemainbuyerswereeitherindividualsortraders,withafewhouseholdsreportingthattheyhadsoldtheirproductstoaprivatecompanyoracooperativethatwasnotsupportedbytheEDGETproject.Individualsandtradersalsoappearedtopaythehighestpricesforbothcomparisonandinterventiongroups.

34Significancestatisticscouldnotbeconductedforbaselineandendlinecomparisonsduetodifficultiesinmergingbaselineandendlinedata.

EVALUATIONREPORT

91

Incomefromthesaleofmilkandprocesseddairyproducts

Table32belowprovidesanoverviewofrevenuefromrawandprocesseddairyproductsperhouseholdfor the endline survey only. The total revenue is based on adding the revenue from each type ofproduct sold. This differs from the baselinemethodology which used an estimate of the aggregatefigure, yielding a different figure for income. This total estimate was replicated at the endline (seeTable33below)toallowformethodologicalcomparabilitybetweenbaselineandendline,despitethelimitations of thismethod. The baseline data for income from different types ofmilk products hadmultipleerrorsandcouldnotbecomputed reliably. In subsequentcalculationsofnet income, thesefigureshavebeenexcludedinordertoensurecomparabilitybetweenbase-andendline.

The table shows that comparison group generated a total of 7093 Birr from sale of raw milk andprocesseddairyproductscomparedto7671Birrintheinterventiongroup(differenceisnotstatisticallysignificant,p>0.1). It isobservedthat themostof therevenue ismadeupof revenue fromrawmilk(3780vs.5294Birr)followedbybutter(2883vs.1910Birr).Cottagecheeseandsouredmilktakeupamuchsmallerproportion.

Table32Incomeearnedfromsaleofmilk,dairyproductsanddairyrelatedactivities

Endline

Comparison Intervention

Revenuefromrawmilk(Birr) 3780.0 5293.9

Revenuefrombutter(Birr) 2883.1 1910.0

Revenuefromcottagecheese(Birr) 426.9 402.3

Revenuefromsouredmilk(Birr) 3.6 64.6

Totalrevenuefromsaleofmilkanddairyproducts 7093.5 7670.9

The Table 33 below presents the average income earned from sales of all dairy and dairy-relatedservices.Asmentionedabove,theincomefromdairyisbasedonanestimateofthetotalratherthanthe addition of revenue from each type as presented in the table above. The only statisticallysignificant difference between the comparison and intervention woredas at endline is the higheraverageearning fromthesaleofanimalmanureordung in thecomparisonworedas.Thedifferencebetweencomparisonaninterventiongroupregardingtotalincomeearnedfromdairyrelatedactivitieswas not statistically significant, with an average of 6959 Birr per household per year earned incomparisonworedasand6798Birrperhouseholdperyearearnedininterventionworedasatendline.

Notethatincomefromsaleofdairyanimalswasnotconsideredintheincomecalculationsatbaseline.Furthermore,dairy incomeatbaselinewasnotbasedonthesumof incomefromdifferentproducts.Forcomparability,thesamemethodhasbeenusedatendline.Consequentlytheincomefromrawmilkandprocessedproductsinthetablebelowdoesnotadduptothetotalincomerowfromsaleofmilkanddairyproductsrow.

EVALUATIONREPORT

92

Table33:Incomeearnedfromsaleofmilk,dairyproductsanddairyrelatedactivities35

Baseline Endline


Totalsaleofmilkanddairyproductsa,c,d 1439.3 3165.0. 5611.7 5856.9

Revenuebullservices 0.0 3.9 81.9 251.1

Revenuedairyanimals NA NA 3160.5 2755.0

Revenueanimalmanureordungb,d 6.6 16.4 437.0 147.7

Revenueother 830.7 528.6

Totaldairyincomea,c,d 1445.8 3185.3 6959.2 6798.0

Totaldairyincomeincludingdairy

animals 10119.7 9553.0

a Significant differences in mean averages or cell distribution for baseline intervention and comparison, Chi-square teststatisticorTwo-sampleT-Test,p<0.05bSignificantdifferencesinmeanaveragesorcelldistributionforendlineinterventionandcomparison,Chi-squareteststatisticorTwo-sampleT-Test,p<0.05cSignificantdifferencesinmeanaveragesorcelldistributionbetweenbaselineandendlineforinterventiongroup,Chi-squareteststatisticorTwo-sampleT-Test,p<0.05dSignificantdifferencesinmeanaveragesorcelldistributionbetweenbaselineandendlineforcomparisongroup,Chi-squareteststatisticorTwo-sampleT-Test,p<0.05

There are significant increases over time for total sale ofmilk and dairy products, sale ofmanure /dungforcomparisongroupandtotaldairyincomeforcomparisonandinterventiongroups(pleasealsoseeDIDestimationfurtherbelow).

Costsofproduction

Thehouseholdsurveyscollecteddataoncostsassociatedwithdairyproduction.Thecostsconsideredin this calculation include breeding services, hired labour, interest payments on loans and healthservices.At theendlinean ‘other’ categorywas includedaswell.Themajordriver for thechange incosts is feed,which sawa substantial increase in both comparison and interventionworedas, albeitmoresoincomparisonworedas.Itshouldbenotedthatthebaselinedatacontainsunrealisticallyhighvaluesforcostsassociatedwithbreedingservices.

Table34Costsofdairyproduction.Pleasenotefairlysmallnforsomeofthevariablespresented36

Baseline Endline

Typeofcost Comparison Intervention Comparison Intervention

Breedingservice(Birr)c,d 941.2 1103.1 12.1 45.5

35Forcomparabilitypurposes,baselinedatahasbeentransformedsothatallNAsreceived0values.Pleaseseemethodologysectionforthelimitationsandproblemswiththisapproach.Further,totalsaleofmilkanddairyproductsdonotaddupfromrevenue of individual products. This is because the total revenue has been asked on the one the one hand as a generalquestion(forcomparabilitypurposeswithbaselinedata)andinamoredetailedquestionforeachdairyproductrespectively.Itisduetoinconsistenciesintheseresponsesthatfiguresdonotaddup.Baselinefigureswherecompletelyun-realisticwhichiswhytheyarenotreportedhere.36Forcomparabilitypurposes,baselinedatahasbeentransformedsothatallNAsreceived0values.Pleaseseemethodologysectionforthelimitationsandproblemswiththisapproach.

EVALUATIONREPORT

93

Baseline Endline

Typeofcost Comparison Intervention Comparison Intervention

Hiredlabour(Birr)c,d 2.0 70.7 325.3 260.4

Interestpaymentonloan(Birr) 0.0 1.5 0.0 192.6

Othercostsduringtheyear(Birr) 19.6 97.1

Healthservice/VET(Birr)a,d 41.5 154.0 272.7 257.5

Feed(Birr)a,c,d 206.2 1064.6 2965.0 2471.5

Totalcost(Birr)a,d 1191.0 2393.7 3594.8 3324.5a Significant differences in mean averages or cell distribution for baseline intervention and comparison, Chi-square teststatisticorTwo-sampleT-Test,p<0.05cSignificantdifferencesinmeanaveragesorcelldistributionbetweenbaselineandendlineforinterventiongroup,Chi-squareteststatisticorTwo-sampleT-Test,p<0.05dSignificantdifferencesinmeanaveragesorcelldistributionbetweenbaselineandendlineforcomparisongroup,Chi-squareteststatisticorTwo-sampleT-Test,p<0.05

Perceivedchangestocostsandincome

Whenaskedtocommentonreasonsforperceivedchangesincostsoverthelastfouryears,differencesintheproportionreportingthattheirincomeincreased,stayedthesame,ordecreasedwerefoundtobestatisticallysignificantatthe90%levelonly.41.6%ofhouseholdsinthecomparisongroupreportedanincreaseascomparedto50.0%intheinterventiongroup.

Wherepeoplereportedincreasedcosts,significantdifferenceswerenotfoundinthereasonsforthechange given. Reasons comprisemore cows (40% vs. 36%), buying better quality feeds (55% versus57.3%)andother(4%versus7%)forincreasedcostsandlesscows(67%versus57%),reducedqualityoffeed(28%versus25%andlessveterinarycosts(6%versus14%)fordecreasedcosts.

Netincomefromdairy

Averagenetincomefromdairyattheendlinewasfoundtobe6269Birrforinterventionand6128Birrcomparisongroups.Thedifferencebetweencomparisonandinterventiongroupswasnotfoundtobestatisticallysignificant.However,asignificantdifferencewasfoundinthenetincomebetweenbaselineand endline for both groups. The average costs incurred in the comparison group increasedsignificantly from1192Birr perhousehold to3595Birr perhousehold in the comparisongroup,buthadnotchangedsignificantlyfortheinterventiongroup(2393.7to3324.5Birr)overtime.

Averagenetincomeininterventiongroupincreasedfrom792to2937Birrperhousehold,equivalenttoa total increaseof3.7 times.For thecomparisongroup,net incomewas found tohave increasedfrom 254 to 2534 Birr, equivalent to a total increase of 10 times. Please note that the net incomecalculationisbasedontheestimateoftotaldairyincome,nottheadditionofincomefromeachtypeofdairyproduct.

EVALUATIONREPORT

94

Table35NetincomeinBirrfromdairyrelatedactivities37

Baseline Endline


Totaldairyincomea,c,d 1445.8 3185.3 6128.5 6269.4

Totalcosta,d 1191.0 2393.7 3594.8 3324.5

Netincomec,d 254.8 791.6 2533.8 2937.2a Significant differences in mean averages or cell distribution for baseline intervention and comparison, Chi-square teststatisticorTwo-sampleT-Test,p<0.05cSignificantdifferencesinmeanaveragesorcelldistributionbetweenbaselineandendlineforinterventiongroup,Chi-squareteststatisticorTwo-sampleT-Test,p<0.05dSignificantdifferencesinmeanaveragesorcelldistributionbetweenbaselineandendlineforcomparisongroup,Chi-squareteststatisticorTwo-sampleT-Test,p<0.05

DIDestimationnetincome

We calculated net effects of net income for the intervention using difference-in-difference analysislooking at baseline and endline data.We estimate a significantmain effect of time (i.e. net incomeincreasedby2279Birrfrombase-toendline).Wedonotseeasignificantmaineffectforbeinginthetreatment group, nor can we see a significant effect for the DID-estimator (i.e. net effect for netincomechange).Duetotheviolationofnecessarypre-conditionsforperformingDID-estimation,theseresultsshouldbetreatedwithcaution.38Pleasesee

Table35andFigure23formoreinformation.

37Forcomparabilitypurposes,baselinedatahasbeentransformedsothatallNAsreceived0values.Pleaseseemethodologysectionforthelimitationsandproblemswiththisapproach.38InordertoestimatetheDIDorneteffect,pre-conditionssuchashomogeneityofvariancesandnormaldistributionhavetobecheckedfor.Thesepre-conditionsarenotestablishedfornetincome.Thereareseveraloutliervaluesandvariancesofmilkproductionbetweenbase-andendlinearenotequal.Thismaybeduetothefactofsamplinganewatendlineratherthanhavingalongitudinalapproachwherethesamehouseholdsaresurveyedatbase-andendline.OrduetosubstitutionofNAvaluesinthebaselinedatasetwith0orinconsistentcleaningofend-andbaselinedatasets.

EVALUATIONREPORT

95

Figure 23Differences betweengroups andover time for annual net incomeper household. The 95%

confidenceintervalsdepicttherangeofthemeanaveragenetincomeinthepopulation.

CooperativeswithDairyProcessingUnits

Overviewofcomponent

EDGETproject’sapproachtodevelopingoutputmarketsformilkandmilkproductsfocusesprimarilyontheestablishmentofDairyProcessingUnits(DPUs)-attheworedaandinsomecases,atthekebelelevel. Dairy Processing Units are facilities that are attached to a cooperative and managed by adedicated management committee. Dairy farmers in the woreda can become members of thecooperative, whether they are members of a DFEG or not. EDGET project provides training to themanagement committees (on management, bookkeeping, hygienic milk production, milk qualitytesting, marketing, etc.) and equipment for milk collection, storage, testing and processing. EDGETproject alsoprovides trainings to theworeda cooperativeagencyand theworeda livestockoffice toorientthemontheDPU’sandgettheirsupportinkeytechnical,legalandoperationalmatters.

Where the catchment of the DPUs are large, EDGET project has promoted the establishment ofdecentralisedMilkCollectionCentres(MCCs)tofacilitatetheaggregationofmilkfromindividualdairyfarmerstothecooperative.CooperativeswithDPUsselleitherrawmilkorprocessedmilkproductstoprivate sector or institutional buyers, including othermilk cooperatives/milk unions and larger scalemilkprocessors.EDGETprojectandtheWoredaCooperativeAgenciesplayaroleinfacilitatinglinkagesbetweentheDPUsandtheseotheragencies.Farmersaretypicallypaidforthemilktheyprovideonamonthlyortwo-weeklybasisandinsomecasesalsoreceiveannualorbi-annualdividends.

EVALUATIONREPORT

96

Figure24ActormapoftheDPUcooperativeandMCCcomponent


Table36AchievementofprojectoutputsforcooperativesandDPUs

Output

description

Indicator Achievement end of

project

Endofprojectrevisedand

(original)targets

Milkcollectionandcooperativeenterpriseprocessingcentresestablished

Numberofmilkcollectionandprocessingcentresestablished

86/95%

(52DPUs+34MCCs)

91,i.e.53DPUs+38MCCs(160)

Numberofmilkcollectionandprocessingcentresleaders/workerswhoreceivedtrainingonmilkqualityandhandling

378/118% 318(NA)

Developmentoftechnologies&

Numberofdairyfarmergroups/FOsreceivedprocessing

70/77% 92(170)

EVALUATIONREPORT

97

Output

description

Indicator Achievement end of

project

Endofprojectrevisedand

(original)targets

strategiesinprocessingandmarketingsupported

andmarketingtechnologysupport

SmallscaleProcessingTechnologysupport

70/78% 91(160)

Inpacketpasteurizationsupport 0/0% 1(10)

Businessrelationships&investmentinproduction,processingandmarketingsupported

Numberofprocessingunitsaddressedinbusinesslinkagefacilitationsupport

44/83% 53(160)

NumberoftargetedDairyHHslinkedwithformalmilkmarket

3,198/20% 16,000(16,000)

Key focus areas for the EDGET project under this strategic objective were the establishment andupgradingofcooperativeswithDairyProcessingUnits(DPU)andmilkcollectioncentres.

MilkCollectionCentres(MCCs)andDairyProcessingUnits(DPUs)

TheEDGETprojectwanted to setup38milk collectioncentresand53processingunits.A collectioncentre (DPU)servicesaround100 farmersand ismanagedthroughacooperative. It is intendedasacentralcollection formembers tosupply theirmilk to. Itwas intendedthatqualitycontrolwouldbeadministeredandthatmilksupply is recorded.Payment formilkwouldbeperiodical (e.g.every twoweeks)orasagreedamongthemembers.TheDPUismanagedbyacommittee.Collectioncentrescansell rawmilkon retailor inbulk towholesalers/processors.Processingcentres (MCCs)arecollectioncentreswithanadditionalfacilityforskimmingmilk,butterchurningorsimilarfacilities.Thesecentrescan sell raw skimmedmilk, butter and cottage cheese (cf. Strategy for EstablishingDairy ProcessingUnitsfortheEDGETproject,2014)

The EDGET project conducted market studies in Woredas in which DPUs and MCCs were to besupportedwithcapacitydevelopment,storageandprocessingtechnologyandmarketlinkagesupport.TheresultingstrategywaspilotedinOromiain2015.

Byendof2015,atotalof13DPUshadbeenestablishedacrossOromia,AmharaandSNNPR(38shortoftarget).In2016,18DPUswerefurtherestablishedand/orre-organised(17shortoftarget,nomilkprocessing centres established). A total of 22 DPUs received processing equipment and relevanttrainingby2016.SixDPUsreceivedsupportonbusiness linkages. In2017,34MilkCollectionCentres(MCCs)andfurther21DPUswereestablished.

The reasons for thedifficulties inestablishingand supportingexistingmilk collectionandprocessingcenterslieintheslowprocessofmembershipmobilizationandregistrationasacooperative,fulfillingthe necessary prerequisites for a dairy processing unit (such us housing, etc.) and foreign currency

EVALUATIONREPORT

98

problemswhenimportingcreamseparatorsandchurnersfromabroad.EstablishingorsupportinganexistingDPUtooksignificantlymoretimeandresourcesthananticipatedduringinitialprojectplanningwhichiswhytargetswererevised.39

Overall, the number of establishedMCCs andDPUswas 86 (96%of revised target)with small-scaletechnologysupport for70of them(76%ofrevisedtarget).44DPUs(83%ofrevisedtarget) receivedbusiness linkages support. The first cooperative is expecting to receive a small-scale pasteurizationplant technology inearly2018.By theprojectend,a totalof3,198dairyhouseholds (20%of target)werelinkedwiththeformalmilkmarketthroughtheDPUsandMCCs.

39Targets forestablishingMCCsandDPUsaswell as for in-packetpasteurization supportwere significantly reviseddue tochanges in project approach, practical challenges such as delay of technology imports, and feasibility of establishing andsupportingDPUs.

EVALUATIONREPORT

99

FindingsfromthequalitativeassessmentTable37Qualitativefindingsoncoopsanddairyprocessingunits


Institutionallinkages(e.g.withworedalivestockofficeandworedacooperativeagency)

Aspartof itssupporttotheDairyProcessingUnits,SNVworkedwiththeWoredaCooperative Promotion Agencies (WCA) and Woreda Livestock and FisheriesResource Development Offices (WLO) in each woreda, to orient them on theproject,trainthemonhowtosupporttheDPUsandtofacilitatelinkages.

The WCAs play a crucial role in formation/registration of the cooperatives,establishingby-laws,auditingthem,resolvingdisputesandfacilitatingtheirlinkagestoothercooperativesandunions(e.g.LemuBilbilocoopagency linked6DPUstoAselamilkunion).

Inall5oftheworedas,theWCAisseen-atleasttosomeextent-tohaveplayedits role in supporting the DPUs and providing some of the key areas of supportoutlinedabove.

While the WLO’s have a less direct role with respect to the DPUs from anorganisational/operational perspective, their overall role in strengthening dairyvalue chainsmeans they are an important stakeholder. The project has engagedthemintrainings,orientedthemontheDPUsandhelpedtoestablishaplatformforthem to coordinate and liaise with the Woreda Cooperative Agencies (e.g. asmentionedinAletaWondoandLemuBilbilo)

Some actors felt that the WCAs still lack capacity, particularly in relation tohelpingtoestablishmarketlinkagesfortheDPU(AletaWondo,LemuBilbilo,etc.).InLemuBilbilo, theWCAwasreportedtohaveonlyconductedanaudit in2016andnotin2017.

Someactors insomeworedas (e.g.AletaWondoandLemuBilbilo) feel that theWLOsaremorefocusedoncrops(e.g.coffeeinAletaWondo),particularlyduringharvesttime,ratherthandairy.ThemultiplecompetingprioritiesareassociatedwithpoorperformanceoftheDPUs.

Establishmentandstrengtheningoftheworedadairycooperatives/DPU

EDGETprojectprovidedvarious typesof support toworedadairycooperatives tosupporttheestablishmentandstrengtheningofDPUs.This includedthefollowingitems:

Trainingson

● Milkqualityandhygiene● Businessdevelopmentforcoops● Book-keeping● Milkprocessing● Milkmarketing● ExposurevisitstootherDPUs

ProvisionofequipmentfortheDPU:

● Lactometer● Alcoholtestingkit

DPUsareofteninefficient.Lacksufficientsupplyofmilk.Lackofdecisionsduetointerests.Competitionwithtraders(e.g.Wuchale,LemuBilbilo40)

There was mention of committees overcharging on transport costs and takeadvantageperdiems.

40TheDPUsinLemuBilbilowerenotestablishedbyEDGETprojectrathertheywererehabilitated.TheyexistedbeforeEDGEprojectintervention.

EVALUATIONREPORT

100

Sub-component Strengths Weaknesses● MTS(notprovidedtoDPUbuthelpstheDPU)● Creamseparator● Butterchurner● Refrigerator● Refurbishmentofbuildings

Facilitatingbusinesstobusinesslinkages

● Throughnetworkingevents● Byscopingoutmarketsandmakingintroductions

Monitoringsupport,qualityinspection

Organisationaldevelopment

● Marketing● Membership● Establishment and capacity development of MCC (in SNNPR and

Amhara)Beyond this SNV’soverall support to strengthening thedairy value chain throughextension activities (training farmers on cleanmilk production, etc.), provision ofMTS,supporttoAgIDsandincreasingtheavailabilityofforageseedsareallseenasimportantcontributorstotheDPUsaswell,sincetheyhelptoincreasethesupplyofqualitymilk.

Informaltraders Informal traders provide an avenue for farmers to sell their milk often atcompetitive rates,withdoor-stepcollectionandwithoutmuch regard forquality.From the farmers’ perspective, they provide an important avenue for the sale ofmilk.

InmostoftheworedasthereisaverylargenumberofInformaltradersaswellassmall scale buyers like cafeterias and tea shops. Their lack of attention to thequality ofmilk and the goodprices they pay are perceived as a threat tomoreformal, regulatedmilk markets. Moreover, their sheer number means that thegovernmentinstitution,withtheirlimitedcapacity,areunabletoregulatethem.

Asaresult,notonlyDPUcooperativesbutalsomilkprocessorsareoftenunabletogetaccesstothevolumesofqualitymilkthattheyrequiredtooperate.

MCCs WhereMCCswereoperational,theyseemtobeworkingwellonthemostpartandfacilitatingthetimelyflowofmilkfromfarmerstotheDPUs.

PerceptionsofcheatinginmilkmeasurementandpricewascitedbyonegroupofDFEGmembersinDangila.

Delays in milk provision to the MCC by member farmers adversely affects thesupplyofmilkfromtheMCCtotheDPU.Thisisinpartlinkedtolimitationsonthesideoffarmersbutalsotolowlevelsofmilkproductionlinkedtothehighcostsofinputs

FunctioningoftheDPUs

4outof the5woredaswere reported tobe functional,withonehaving stoppedmilkcollectionduringthelastyearduetointernaldynamics.

DPU reported various benefits resulting from the support provided by EDGETproject.

NotallDPUsarereportedtobefunctionalintermsofbuyingandsellingmilkfromfarmers. The DPU in Lemu Bilbilo was perceived not to be playing its role bymultipleactorsasa functionofweekcommitmentbytheDPUcommittee,poormarket linkages anda lackof support from theWCAand theWLO. TheDPU inWuchalewasalsoseentobeoperatingquitepoorly.

EVALUATIONREPORT

101

Sub-component Strengths WeaknessesBenefitsreportedinclude:

● Increasedregularityofpaymentstofarmers(e.g.LemuBilbilo)● ImprovedworkingenvironmentforDPUcommittees● Reducedwastageandreturnofmilk● Regularmonitoringofmilkquantityandprice● IncreasedunderstandingofandcapacitytofulfiltheroleoftheDPU● Increaseinthesupplyofmilk● Improvedmanagementofthecooperatives● Improvement in thequality ofmilk suppliedby farmers (as a result of

trainingandtestingofmilk)● EquipmentintheDPUiswellmaintained● Insomecases(e.g.Dangila),theDPUreportedthatthedemandformilk

from buyers had increased as a result of recognition of the improvedquality

● DPUs are able to buy collect milk even during the fasting season(Dangila)

● IncreaseddiversityofmilkproductsarebeingproducedSomeDPUshavebetterreportingandmanagementpracticesinplace.(Machakel)

Thelackoflinkageswithmilkbuyersappearstobethemajorimpediment,thoughitisabiggerissueinsomeworedas(e.g.AletaWondo,LemuBilbilo)thanothers.In some cases this is because of limitations on the part of the DPU and otheragencies,inothersitappearstobeduetotheweakstatusofthemarketformilkinthearea.Oftenitisacombinationofboth.

Milk purchasing can sometimes reduce significantly during the fasting season,withobviousnegativeimplicationsfordairyfarmers.InWuchale,DFEGmembersreportedthattheDPUrigsthealcoholteststorejectmoremilkduringthefastingseason.

Transportation issues(i.e.gettingthemilkfromfarmerstotheDPU)alsoaffectstheabilityofDPUtocollectmilkfromallthedairyfarmersinthecommunity(e.g.Machakel,LemuBilbilo,Wuchale).

Similarly,poorroadconnectivityinMachakelisseenasasignificantconstrainttotheabilityoftheDPUtosellmilktooutsidebuyers.

InLemuBilbilo,itwasreportedthatfarmersweresupplyingadulteratedmilk.

Poor performance of the DPU (milk collection, expanding membership, makingpayments)negativelyimpactsontheDA’sabilitytoplaytheirrole.

Lowdemandformilkinthewidermarketswasreportedasanissue.

FarmersreportedabadexperienceinWuchaleandMachakelduetobankruptcyofthelocalmilkunion(notthecooperative/DPU)whichresultedinnon-paymenttofarmerswhohadsuppliedmilk.Thishasloweredfarmers’trustandinterestinworkingcollectively.

According to DPU committee representatives from Dangila and Wuchale, themainreasonsformilkrejectionwererelatedto:

● Healthconditionofthecow;● Distanceofmilkcollectingcentre● Hygieneofmilkequipment● Insufficientfatcontentand/orabsenceofcreaminthemilk● Milkcombinedwithpowderedmilk

Regionallevelmilkprocessors

TheSNNPRregionalmilkprocessorreportedreceivingtrainingsfromEDGETprojectand support in establishing linkages with farmers. These trainings helped theprocessors to collect quality milk and advise their suppliers on how to producequalitymilk.

Overallthedemandforprocessedmilkwasreportedtobeincreasingasaresultofincreasingurbanisation.

TheEmbetmilkprocessorreportedthatEDGETprojecthadtriedtolinkthemwithfarmersbutthatitwasuneconomicaltotransportmilkfromsofaraway.Asaresult

The regional milk processor from SNNPR reported challenges in securing asufficientsupplyofqualitymilk.Forexample,AlmiAsmamawreportedthattheirplanthasacapacityof20,000litresperdaybutiscurrentlyonlycollecting50%ofitscapacity.

Roadsareofpoorqualityandmakemilktransportationdifficult.

Suitablepackagingmaterialswerealsoreportedtobeverycostly,pushingpricesof milk up and reduces demand, with many consumers preferring raw milk or

EVALUATIONREPORT

102

Sub-component Strengths WeaknessestheyreportedthattheirbusinessdidnotbenefitfromEDGETprojectsupport. unpackedprocessedproducts.

Mediapropagandasurroundingaflatoxinissuesin2015wasreportedtohavehadanegativeimpactonmilkdemand.

Limitedoperationalbudgetsanddifficulties inobtaining loans frombanks limitsbusinessdevelopment.

EVALUATIONREPORT

103


DPUs address a key gap in the existing outputmarkets formilk andmilk products since there is an

absence of buyers with quality standards at the woreda and kebele level who can purchase from

smallholderdairyfarmers.

MCCsarealsorelevantinthattheyfacilitatethesupplyofmilkfromdairyfarmerstotheDPU,thereby

overcomingkeybottlenecksrelatedtomilkcollectionandtransportation.

The activities carried out by SNV to support the development of the DPUs (institutional linkages,

trainings,equipmentprovision,formationofMCCs,etc.)areallfoundtobehighlyrelevant.

Contextual factors (prevalence of crossbreed cows, existence of viablemilkmarkets, infrastructure,

etc.)howeveraremajordriversofthefunctionalpotentialofDPUs.


Overall SNV’s support in establishing and developing DPUs has been mostly effective. The key

challenge is that market development is not a linear process and contextual factors have posed a

challenge.

Delays in the provision of equipment have limited the extent to which the dairy cooperatives

supportedwithDPUshavebeenabletofunctionasintended.

DPUsarenotallfunctioningwell.Somewereevenfoundtohavestoppedcollectionatthetimeofthe

evaluation.

DespitetheseissuesDPUshaveledtoincreasedinterestindairyfarmingandanincreasedrecognition

oftheimportanceofmilkquality.Theircontributiontoincreasedincomesfordairyfarmers,however,

remainsquestionable.

ThemainissuesfacedbyDPUsare:

• Difficultyinestablishingmarketlinkages:LimitingtheDPU’sabilitytopurchaseandsell-onthe

milkthatisproducedbyfarmers

• Managementissues:ThesecaneasilyunderminethefunctionalityoftheDPUs.

• Lack of equipment: Without relevant facilities and equipment the DPUs cannot function

effectively


The evaluation found that theDPUswere heavily reliant on external support for their functioning -

bothintermsofmanagementandoperationsandtheestablishmentofmarketlinkages.

The capacity of the Woreda Cooperative Agencies to effectively play their roles in supporting the

developmentofDPUsappearstobesomewhatlimited,asperceivedbykeyprojectactors.

Even if market linkages are established, the development of dairy cooperatives into effective

organisations isa long-termprocess,besetwithrisksandchallenges.Withoutongoingorganisational

EVALUATIONREPORT

104

and institutional support, extending well-beyond the end of the EDGET project, the prospect for

sustainabilityofthecooperativesislikelytobequitelimited.

IfDPUsarenotabletopurchaseandsellmilkasperdairyfarmers’productioncapacity(allyearround),

the gap opens up the opportunity for informal traders to step in.While this may be beneficial for

farmersintheshort-termbecauseofconvenienceincollectionandlowmilkqualitystandards,italso

risksunderminingthevisionofahigh-qualitymilkvaluechainthatlendsitselftocommercialisation.


TargetingandselectionofworedasandkebelesforestablishmentofDPUs:careneedstobetakentoensure that DPUs are established in locationswhere they have sufficient potential to develop. This

meansthatissuessuchasroadconnectivity,electricitysupplyaswellasaccesstoviablemarketsanda

sufficientactualvolumeofmilkproductioninthecatchmentareashouldbecarefullyassessed.Failure

to adequately consider all these factors limits the ability of cooperativeswithDPUs to succeed and

riskscontributingtonegativeperceptionsabouttheirrelevance/utilityamongstdairyfarmers.

Linkages with buyers: Cooperatives with DPUs can only serve their purpose if they have stablerelationshipswithbuyerswhocan(a)absorbthevolumeofmilkthatfarmersneedtosupplyand(b)

offeracompetitiveprice.Thecapacityofthecooperativestoestablishtheselinkagesthemselvescan

bequitelimited.Athoroughmarketassessmentandclearcommitmentsfrompotentialbuyersmaybe

aprerequisitetosuccessfuldevelopmentofthecooperativeswithDPUs.

Organisationalcapacity:TheorganisationaldevelopmentofDPUsislikelytobealong-termchallenge.

Appropriate mechanisms for providing business, technical and management support to the

cooperatives is essential. However, the requirements my well be beyond what the Woreda

CooperativeAgenciesarecapableofproviding.Assuch,alternativesupportorganisationsmayneedto

be established that can play the required role across multiple woredas, creating alignment in the

approach and facilitating standards and best practices across these cooperatives. This is a role that

couldevenbeplayedbylargerprivateplayerswhodecidetolinkupwiththeDPU/cooperatives.

SequencingofDPUdevelopmentactivities:Inmanywaysareiterationofwhathasalreadybeensaid,

the sequencing of DPU development activities should be carefully managed. DPUs need to have

simultaneousaccesstoappropriateequipment(e.g.forrefrigeration)andlargerscalebuyersinorder

to offer a viable milk market for dairy farmers. Failures on either front can undermine the overall

functioningoftheDPU.

4.4StrategicObjective3:Tocontributetodevelopmentofinstitutions

andtodairysector-wideinitiatives

StrategicObjective3encompassesa rangeof interventions thatengagewithavarietyofdairyvalue

chain actors at the regional andnational levels aswell aswithworeda livestock offices. Preliminary

discussionswithrelevantinstitutions(e.g.SARCandzonalagriculturalbureausinSNNPRandAmhara)

wereheldin2014toexploreopportunitiesandneedsforcapacitystrengthening.Aneedsassessment

wasconductedin2014andledtothefollowingareasofsupport:

EVALUATIONREPORT

105

• Institutionalsupporttoworedalivestockoffices

• Engagementwithregional/nationalforageseedproducersandmultipliers

Institutionalsupporttoworedalivestockoffices

Overviewofcomponent

As part of its support to the development of pro-smallholder dairy value chains, EDGET project

provided institutional support to woreda livestock offices above and beyond the core support on

extensionservicedelivery.Thefocusofthesupportprovidedwasonaddressingsomeoftheimportant

constraints faced by the woreda livestock office. Particular attention was given to the role of the

woredalivestockofficeinprovidingArtificialInseminationservices-withtheprovisionofmotorcycles

and large and small liquid nitrogen flasks for storage at the woreda livestock office and for

transportationbymotorbike.Inadditiontothistechnicaltrainingsandcapacitydevelopmentsupport

wereprovidedtotheworedalevelAItechnicianstoenhancetheirabilitytoplaytheirrole.

Whileanimalhealthservicesarealsoakeyresponsibilityoftheworeda livestockoffice,thiswasnot

identifiedasapriorityareaforprojectsupportbytheworedalivestockoffice.

Figure25Actormapoftheinstitutionalstrengtheningcomponent(woredaservices)

EVALUATIONREPORT

106


Table38Institutionalsupporttoworedalivestockoffices-achievementofoutputtargets

Outputdescription Indicator Achievementendofproject

Endofprojectrevisedand(original)targets

Dairy sector stakeholders that

address critical constraints for

dairy sector development

supported

Numberofregionalandfederal

level dairy sector institutions

supported6/100% 6(9)

AI service access and inefficiency is one of themain constraints identified. For this, EDGET project

provided training support to 183 AI inseminators at woreda and kebele level and supply of liquid

nitrogencontainersinAmharaandSNNPRin2015.AIequipmentwasonlydeliveredinOromiain2016

asa resultofdelays causedby foreigncurrency shortageandother logisticalproblems facedby the

supplier.

Figure26TimelineofimplementationforSO3

In sum, more than 55 government offices (mostly woreda livestock offices) benefitted from AI

equipment, even though delivery of procured goods and services by contracted suppliers was

sometimes late.Atotalof183AItechniciansreceivedtrainingsupport.Sixregionalandfederal level

dairy sector institutions were supported (100% of revised target). The challenge in engaging more

institutionsweremainlyduetothelimitedbudgetallocationandtheengagement/capacityofregional

partners.

Assessmentbysubcomponent

Table39Qualitativefindingsontheinstitutionalstrengtheningcomponent(woredaservices)


Artificial

inseminations

Transportation problem for AI technicians

addressed

Equipment (liquid nitrogen flasks) is helping to

ensurequalityofthesemen

ImproveddeliveryofAIservicestofarmers(faster

service)

Increase in the number of AI technicians in the

Therearestillgapsbetweendemandandsupply

Some DEPs reported varying levels of skill and

knowledgeamongAItechnicians.

Despite thematerial supportprovided thereare still

issuesofsupply(e.g.availabilityofliquidnitrogenand

qualitysemen)

Concerns about the semen quality remain despite

EVALUATIONREPORT

107

Sub-component Strengths Weaknessesworeda (AletaWondo, LemuBilbilo), thoughnot

attributedtotheproject.

Trainings on fertility management are being

providedbyAItechnicians

Overall this is leading to an increase in the

numberofimprovedbreedcowsintheworedas

gains in service delivery efficiency. In some cases

farmers reported getting local breeds instead of

cross-breeds.However,othershavereportedthatthe

efficiency of AI (in terms of positive pregnancy

diagnosis) has not changed substantially and thatAI

techniciansstilltakealongtimetoreachthefarmers.

ThesupplyofsementotheWLOcanalsobeanissue

(e.g. Lemu Bilbilo had no supply between June and

Oct2017)

Some farmers report having to repeat insemination

and also cited costs in excess of the official

governmentfeepayingETB100insteadofETB5)

FarmersgetdemotivatedquicklywhenAIservicesare

notsuccessful

InWuchale,theissueoffuelcostswasalsoraised.AI

techniciansreportedonlyhavingenoughfuelbudget

to cover 50-60 km per day even though the

requirementismuchhigher.

InsufficientnumbersofAItechnicians(Wuchale).

Animalhealth Woredaofficesareplayingtheirroleinproviding

drugs,treatmentandimmunisationservices

Unavailability and high price of medicines for

livestockisreportedtobeanissuebyDFEGmembers

and other actors in a number of woredas (e.g.

MachakelandDangila).

Disease outbreaks were identified as an issue in

Dangila, leading to death and reduced milk

production. Mastitis was flagged as a major issue

LemuBilbilo

Lackofexplicit focusonanimalhealthseenasakey

gapinEDGETprojectbyanumberofdifferentactors

attheregional,woredaandkebelelevels.

Delaysinanimalhealthexpertsreachingfarmers


ThesupportprovidedbytheEDGETprojecttotheWLOaddressedclearneedsoftheWLOinrelation

to theprovisionofAI services.Equipmentprovidednotably included theprovisionofnitrogen flasks

andmotorcycles, critical forAI technicians tomaintain thequalityof semen, increase their response

timeandexpandtheircoveragearea.

ThesupportonAIprovidedisseentobeinsufficientwithrespecttothedemandbecause:

o BudgetconstraintsontheWLOsideforlogistics/transportation/fuel

o Irregularandinsufficientsupplyofqualitysemenandliquidnitrogen

AnimalhealthisanimportantissueraisedinMachakel,LemuBilbiloandDangilathatwasonlypartially

addressed through the project. More specifically, preventative health care was provided through

improved feeding practices, improved housing, hygienicmilk production, etc., whichmany farmers’

reportedhad ledto improvedanimalhealth.However,poorresponsetimesofvets, lackofandhigh

EVALUATIONREPORT

108

costofmedicinesmeantthatwhenanimalsdofallill,thereisahighlikelihoodthattheydonotreceive

treatmentinatimelymanner.


The support toAI technicians increased their coverageand there is aperceived improvement in the

deliveryofAIservices.

ThesuccessoftheAIsupporthasbeenfurtherbolsteredinareaswheretherewasanoverlapwithAGP

(e.g.AletaWondo),whichsupportedtherecruitmentofadditionalAItechnicians.

Despitethis,issuesstillremainforreasonsoutlinedabove.

Overall, thecapacity trainingandequipment supportwas seen tohaveplayedaconstructive role in

enablingtheongoingprovisionofAIservicesintheworeda.

However, thekeyconstraintsof limitedbudget for transportand the lackof inputs suchasnitrogen

andsemenwereseenasongoingchallengesthatlimitthescaleandqualityofAIserviceprovision.


The technical trainings and equipment provided are likely tomake a sustainable difference (good

quality,etc.)

• ThemotivationofAItechniciansisquestionable


The institutionalsupporttoworedasonArtificial Inseminationmighthavebeenbetterconsideredas

an integralelementof StrategicObjective1, since it concernskey inputsand services related to the

developmentofthedairyvaluechain.

• WhilethetechnicalandmaterialsupportprovidedtoWoredas is likelytobesustainable,the

activities carried out do not amount to a coherent and effective programme to enhance AI

servicedelivery.

• In the future, amore systematic approach to strengthening AI serviceswould be advisable,

eitherasastandaloneprogrammethatworksinsynergywithprojectslikeEDGET,orasafully

developed component, similar to the work with AgIDs and on strengthening forage seed

systems.

Ideasputforwardbyrespondents:

• IntroducebullbreedingserviceinsteadofonlyAI

• Semenqualitytestkitsshouldbesupplied

• Liquidnitrogencentre(Aselanitrogencentre/LemuBilbilo)shouldberenovated

• Useofmethodssuchassynchronisation

• DrivinglicensesforAItechnician(Machakel,aspertheDA)

• Allocationof(more)budgetbythegovernmentfortransportationandfuel

• Increasedfocusonanimalhealth

EVALUATIONREPORT

109

Improvethesupplyofqualitysemen

• Awiderangeofactors-woredaofficers,DAsanddairyfarmers-reportedissuesrelatedtothe

quantityandqualityofsemenavailable.Addressingsuchbottlenecksinthesupplychainneed

tobeaddressedinthefuture.

Improvethesupplyofliquidnitrogen

Engagementwithregional/nationalforageseedsuppliers

In 2014, discussions commenced with possible regional institutions and assessment of institutions

started.SouthAgriculturalResearch Institute (SARI)hasbeensupported inusingtheir forage landto

producehighlydemandedbasicforageseed.

Assessment of institutions finalised in 2015 and agreement was reached with regional livestock

agencies on the institutions to be supported and how to overcome challenges. Institutions (i.e.

livestockagenciesintheregions)hadtosubmitproposalstoSNVforapprovalofcapacitysupport.In

2016,engagementwithgovernmentregionalforageseedmultiplicationcentresstartedtoaddressthe

forage seed and plantingmaterials shortage. Despite repeated efforts and discussionswith regional

officials, only SARI came upwith a good enough proposal on forage seedmultiplicationwhichwas

approved.Supportfocusedonforageseedmultiplicationtoaddressforageseedshortage.Thisledtoa

totalof66.5quintalsofforageseedinthefirstroundofmultiplication.


Thesupplyofimproved/qualityforageseedconstitutesamajorconstraintinthedairyvaluechain.As

such,workingwithcredibleandreliable-supplyinstitutionstodevelopseedandplantingmaterialsisa

highlyrelevantactivity


WhiletheworkwithSARIhasbeenpositive,itisclearthatthequantityofforageseedproducedfalls

significantlyshortoftherequirements.

Assuchtheoveralleffectivenessofthiscomponentrevealssomesignificantgaps.Themechanismof

requestingproposalsfrominstitutionsseekingtocollaboratewiththeEDGETprojectmayhaveproved

tochallenginginlightofthelimitedcapacitywithinthesector.Assuchalternativeapproachesmaybe

requiredtoaddressthisgap.


Itisuncleariffarmers’demandorforageseediswellunderstood.Inlightofthisaclearassessmentof

thequantityofseedrequiredandaclearplantomeetthesupplyshouldbeundertaken.Ifprivate

sectorsuppliersand/orfarmersproducingseedmeettheseneeds,asustainablesupplyoftheseinputs

canbedevelopedtomeetthedemandarticulated.

EVALUATIONREPORT

110


A focused, in-depth analysis of the forage seed system carried out in close coordination with key

stakeholders in the system could help to underpin a clear strategy for addressing the demand for

forageseed.

Theroleoftheprivatesectorinseedmultiplication-e.g.throughoutgrowerschemesinvolvingAgIDs

orregionalfeedsupplierscouldbefurtherexplored

4.5 Strategic Objective 4: To develop a knowledge base on dairy

relatedissues


EDGET project developed and disseminated extension related and good practice materials in dairy

production, processing, marketing and development. The project also organised knowledge sharing

events.


The project developed its Learning and Knowledge Management Strategy in 2014. Project

representatives participated in EKN Learning Knowledge Event sharing their experiences. EDGEThas

alsobecomeanactivememberoftheLivestockBroaderPlatformandLivestockTaskForce.

In2015,theprojectorganised48(64%of2015target)dairygoodpracticeevents(46atworedaand2

at zonal level) for a total of 1080 farmers in forage development and calf / cowmanagement. 17

technical discussions were also facilitated among livestock experts and extension personnel from

government at zonal andworeda level. 3 farmer extensionmaterialswere developed (one short of

target).

Inthesameyear,theprojectcarriedout73reviewsessionsatbothcentralandregional levels.M&E

staffandtheEDGETprogrammemanagerattendedallAgri-proFocusandEKNlearningeventsandin

oneof thempresentingchallengesand lessons inworkingwith thepublicandagriculturalextension

system in Ethiopia. Regional managers attended various technical working groups and multi-

EVALUATIONREPORT

111

stakeholder meetings contributing to cross-organisational learning, collaboration and knowledge

sharing(2015AnnualReport).

In2016, theprojectorganisedexperience sharingvisits for2,631 farmers (228%of2016 target)but

due to the situation at field level, events planned could not be organised as expected (0 of 61

targeted). 240,000 farmer extension material was printed and distributed. The extension materials

prepared contained three packages - Forage Development, Calf and Cow Management, and Milk

Handling and Hygiene, with images and text in local languages (Amharic and Afan Oromo). 14

knowledgematerialswerealsoproduced(155%of2016target).

Table40Achievementofprojectoutputsforcreatingaknowledgebaseondairyrelatedissues


End of project revisedand(original)targets

Knowledgebaseof

extensionservicedelivery

systemstrengthen

NumberofDairyDevelopmentTraining

packages/modulesdeveloped10/167% 6(10)

NumberofFarmerExtensionMaterials

packagesdeveloped(numberof

packages)

8/1000% 8(8)

NumberofFarmerextensionmaterials

printedanddistributed(numberof

copies)

240,000/92% 260,000(520,000)

Knowledgebaseof"best

practices"indairy

production,processing

andmarketingdeveloped

anddisseminated

Numberofdairygoodpracticesharing

eventsorganized93/42% 224(224)

NumberofdairyHHsparticipatedin

experiencesharingvisits5,116/197% 2600(2600)

Numberofknowledgematerials

developed&disseminated20/71% 28(0)

In thesameyear,EDGETwiththehelpofan internationalconsultantorganisedthe2016knowledge

event.50EDGETandgovernmentstaffparticipatedoverthecourseofthefour-dayevent.Projectgood

practicesandthreepracticebriefsweredevelopedduringtheevent.Further,4workingpapersand16

farmer stories were written and later published. Similarly to the previous year, EDGET staff

participated in sector meetings and learning events at different levels and participated in the

NetherlandsTrademissiontoEthiopia.In2017,10dairydevelopmenttrainingmodules(125%of2017

target),5 famerextensionmaterials (125%of2017 target)and6knowledgematerials (31%of2017

target),including5practicebriefsand1synthesisreflectionpaper,werefinalised.

Data from the household survey and qualitative interviews regarding where beneficiary farmers

receivedtrainingandknowledgefromisreportedintheuppersectionsinadoptionofpractices.

EVALUATIONREPORT

112

Overallassessment

The EDGET project has made a substantial contribution to knowledge development concerning

inclusivedairyvaluechainsinEthiopia.

4.6 Strategic Objective 5: To improve nutritional status of children

throughdairyconsumption

Overviewofthecomponent

Dairydevelopmentandnutritionhaveanumberofimportantlinkages.Atthelevelofthesmallholder

household,dairydevelopmenthas thepotential tocontribute to increasedconsumptionofmilkand

processeddairyproductswithinthehousehold,withimplicitnutritionalbenefits.Atthesametime,it

posestheriskofreducinghouseholdconsumptionofdairyproductsasmoremilkissold(particularlyif

pricesarehigh)andlessretainedforhouseholdconsumption.Fromamarketsystemperspective,the

increasedproductionofmilkenablesanincreasedavailabilityofdairyproductsforconsumermarkets.



Initially,in2014EDGETprojecthadexploredtheideaofdevelopinganutritiousyogurtbasedchildren’s

drink. However, further research was required to address food safety aspects and there were

difficulties findingasuitable industrialprocessorpartner.Thenutritionstrategywas revised in2015,

pillarsofthestrategyinclude:

• Awareness raising concerningqualitynutritionand the importanceofmilkproducts todiversify

diets;

• Milkfortificationasasolutiontomicronutrientdeficienciesinchildrenunder2andpregnantand

lactatingwomen.

Basedoninsightsgainedin2015(cf.donorreportp.6),itwasdecidedfortheforthcomingyearsthat

the EDGET project will focus all its efforts on the implementation of the nutritional change

communications strategy and piloting of approaches that are directed at behavioural change at the

householdlevel.

EVALUATIONREPORT

113

Table41Achievementofprojectoutputsfornutrition

Outputdescription Indicator Achievement endofproject

End of project revisedand(original)targets

Awarenessraising

events/campaignsonvalue

ofdairyproductsforchild

nutritionorganized

Numberofnutritionawareness

raisingevents/campaigns

organized

19/12% 153(153)

Numberofpeopleaddressedin

nutritionawarenessraising

events/campaigns

6,178/? ?

Affordable&Accessible

dairyproductstargeting

childrendeveloped

Numberofaffordable&accessible

dairyproductstargetingchildren

developedandadopted

Activitywascancelled 0(1)

Awarenessraisingcampaignswereinitiatedin2016onasmallscaleandrepeatedagainin2017(e.g.

World milk day was organised in 2014 and 2015 with the Livestock Resource Development and

Promotion Agency and anNGO). Assessment of nutrition status in project areaswas undertaken in

2016.Thescaleupofthenutritioninterventionsishighlightedasapriorityforthesecondphaseofthe

EDGETproject.

Duetothesmallscaleoftheinterventionandcompetingevaluationpriorities,itwasagreedthatthe

evaluation would not gather primary data to assess the effectiveness of the nutrition awareness

campaigns.However,dataonthepatternsofhouseholdconsumptionofdairyproductswasincluded

in thehouseholdsurvey.Theprimaryanalysis in this section isbasedona reviewofdocumentation

regardingthenutritionpilot.Observationsarepresentedbelow.

Findingsfromthehouseholdsurveyrelatedtodairyconsumption

In terms of consumption of dairy products, the only statistically significant difference found was

related to the consumption of milk at the household level. More specifically, households in

interventionworedaswerefoundtohaveahigherdailyconsumptionofmilk(0.51litres)thanthosein

comparisonworedas(0.3litres)duringthefastingseason.

Pilotingthenutritioninitiative

In 2016, a desk review report on nutrition was published with the support of SAK Business and

PersonalDevelopmentPLC.

InMay 2017, EDGET commissioned a formative research study in intervention woredas on 8WHO

indicatorsforassessinginfantandyoungchildfeedingpractices(IYCF),MinimumDietaryDiversityfor

Women (W-MDD), barriers to practice appropriate maternal, infant and young children nutrition

(MIYCN) and missed opportunities or recommendations to design an effective nutrition Social and

Behavioral Change Communication (SBCC) intervention. The research provided useful insights into a

futurenutritionawarenessraisingcampaign:itshowedthatonly39.4%ofchildreninthesamplemet

requireddietarydiversity,thatW-MDDSwasbelowcriticallimitwithonly4.3%meetingtheminimum

of five food groups out of 10 food groups for consumption. 67.9% heard or received messages or

EVALUATIONREPORT

114

information on IYCF during the 1000 days of life, with 58%mentioning health education in health

facilities by health extensionworkers (HEW).While about half of the respondentswere engaged in

community organisations or social programs, IYCF related messages were not spread through such

gatherings.ThepreferredsourceofinformationtolearnaboutIYCFaretheHEWs.Lackofknowledge

and awareness about IYCFwas cited forHEWandmothers aswell as their supporting environment

(community,family,etc.).41

TrialofImprovedPractices

InJuly2017,EDGETandthesubcontractedfirmEUREKAHealthServicesconcludedaTrialofImproved

Practices (TIPs) to enhance better nutrition during the first 1000 days in three pilot Woredas of

Amhara,OromiaandSNNPRegions.TheTIPsallowsprogramplannerstopre-testtheactualpractices

that a program will promote by testing the messages on practices through close counselling and

guidance (four visits to mothers in total) for an intervention group, and by following up on which

practiceswereadopted.Thegroupdoesnotreceiveanyinterventionorsupport.42Thispilottestedthe

compatibility of SBCC message-materials-channels strategy, i.e. six major MIYC feeding

recommendations identified at the formative research stage. The following qualitative insights for

programmingandmessagingweregeneratedthroughthestudy:

• Key barriers to more optimal breastfeeding include mothers’ lack of information,

demonstrationsandaccesstoadvicefromeldersandpeers;

• Iftheporridgefedtochildrenisthick,motherswereafraiditwouldchokethebaby;

• Parentswerewillingtodoanythingtohelptheirchildreneatandgrowhealthy if theadvice is

complementedwithdemonstration;

• Mostmotherswhoweresuccessfulinaddingmilktotheirchildren’sdietsduringthetrialperiod

reported that theywerepleasedbecause themilk softens theporridge and their children eat

wellanditgavethemanewwaytopreparefood;

• Milk was the food of choice because many mothers recognized that since milk was readily

available,itwasnotexpensive.

Thestudyalsocollectedquantitativedatabasedonasubjectivemeasureofadoptionofpractices in

comparisontoagroup(SeeFigure29below).43

41FormativeResearchtoDeterminetheStatusandBarriersofOptimalNutritionduringtheFirst1000daysofLife:TheCaseof

SNV-EDGETDairyProjectAreas.May2017

42 TIPs is considered a formative and often qualitative technique with small sample sizes. Statistical analysis beyond

description of absolute cases was not part of the study. This means that little can be concluded about messaging

effectiveness.

43SNVEDGET:TrialofImprovedPracticesReport,July2017

EVALUATIONREPORT

115

Figure29Overviewofdata fromtheTIPcomparinguptakeof recommendedpracticesbymothersasreportedduringfollow-upvisits.Pleaseseethelimitationsofpossibleconclusionsfromthequantitativepartofthestudy44

The TIPs study design and possible conclusions based on quantitative data are prone to a range of

criticism:

• Objectivemeasures of effectiveness of themessages on the nutrition status of target groups

werenotcollected(thestudyreliedonlyrecalldataaboutpractices),whichiswhylittlecanbe

said about the impact of adoption of practices on children’s nutrition. Subjective perceptions

andrecalldataarepronetoarangeofbiases.

• Thesamplesizeofthecontrolgroupwas60andfortheinterventiongroupwas180.However,

sub-samplesfordifferentmessagesweremuchsmallerrangingfrom9to51incomparisonand

27to153intheinterventiongroupdependingonwhichrecommendedpracticewaslookedat.

Conclusionsbasedonsmallsub-samplesshouldbedrawnwithcautionandcarefullybackedup

withadditionalandmorerigorousevaluationsinfuture.

• Fromthestudyreportitisnotclearwhichquantitativeconclusionsweredrawnfromthestudy

andwhyacontrolgroupwasemployed.Inordertotestmessageeffectivenessandadaptitfor

theprogrammecontext,onewouldexpecttocompareat leasttwomessageswitheachother.

Those messages should target the same recommended practice (in order to ensure

comparability)butmaybeconveyed inadifferentwayorwithdifferent intensity, forexample

comparingtheeffectivenessofthreeroundsofcounsellingpermothercomparedtojustone,or

comparingdifferentwaysofframingthemessages.

44DiagrambasedonnumbersprovidedintheTrialofImprovedPracticesReport

67%

96%

90%

87%

63%

0%

20%

0%

24%

24%

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Refrainfromgivingliquidsandfoodsotherthanbreastmilkto

yourbaby

0–5monthsold.

Allinfants6-8monthshouldstartsemi-solidcomplementary

foodsbeginningat6monthsofage

Feedyourbaby6–11monthssoftandthickporridgeDonotgive

wateryporridge/gruel

Feedyourbaby6–23monthssoftandthickporridgeformulated

andenrichedfromcereal/rootandlegumeandenrichedwith

animalsourcefoodandvegetable

Increasetheamountoffoodandfrequencyoffeedingyougive

toyourbaby6–23monthsfollowingrecommendation specific

toage.

Successratecontrolgroup Successrateinterventiongroup

EVALUATIONREPORT

116

• The present study looked at sixmessages thatmainly target different groups of children, e.g.

infants 6-8 months or baby 0-5 months. The possibility of comparingmessages is, therefore,

limited.Further,thecontrolgroupdoesnotgetanysupportorintervention.It isthereforenot

surprising that no or only smaller adoption rates can be seen for the control group. It seems

advisable to assess effectiveness of future nutrition related interventions in a more rigorous

manner(e.g.RCTstyleassessments).

SocialBehaviourChangeCommunicationStrategyandScaleupPlan

In June 2017, EDGET project developed the SBCC Strategy and Scale up Plan for the promotion of

appropriatenutritionduringthefirst1000daysoflife.Goalsoftheplaninclude:45

• Enhancethenutritionbehaviorsfocusingonimprovingdietarydiversityofmothersand

infantsindairyproducingHHs.

• Improve the production and consumption of nutritious andmicronutrient-dense diet

forchildrenundertwoandpregnantandlactatingwomenintheprojectareas.

• Strengthen capacity to of Agriculture and Health extension workers on appropriate

nutritionduring1000daysoflifeandimplementingnutritionSBCCinterventions

Toachievethesegoals,theplansetsoutthefollowingSBCCstrategy:

1. Advocacytoincreaseresourcesandpolitical/socialcommitmentforchangegoals

2. Tailormade trainings onNutrition Sensitive Agriculture (NSA) for nutrition task force in the

projectarea

3. ContextualizedawarenesscreationofCommunityleadersandpartners“Learningbydoing”

4. Communitymobilizationforwiderparticipation,collectiveaction,andownership

5. Behaviorchangecommunicationforchangesinknowledge,attitudes,andpracticesofspecific

audiences(includingcounselling,peereducation,demonstrations)

6. Nutritionsensitiveagricultureextension

7. Bestexperiencesharingwithpositivedeviancefamily

Theproposedcross-cuttingstrategicmixwillleadto:

• Creationofdemand for,andutilizationofappropriatenutritionduring the first1000daysof

lifeandtopromotetheadoptionofpositive,healthy,protectivebehaviors.

• Improvementinthequalityofserviceprovisionatthelevelofqualitynutritionservicedelivery,

including capacity strengthening in Interpersonal communication skills, counselling skills and

communitymobilization.

45SNVEDGET:SocialandBehaviorChangeCommunication(SBCC)Strategyforthepromotionofappropriatenutritionduring

thefirst1000daysoflife.June2017

EVALUATIONREPORT

117

OverallassessmentoftheSBCCStrategyandScale-upPlan

Thescale-upplanbuildsonthelearningandinsightsgeneratedfromtheformativestudyandtheTIPs

report.Itappearstomakegooduseofsynergieswithexistingprojectinterventionsandoutcomes,i.e.

thatmoremilk fromdairyproduction isavailable forconsumptionbymothersandchildren.Mostof

theobjectivesandindicatorsreferencedintheplanaretime-boundtoDecember2017(cf.Behavioral

M&Eprocessobjectivesand indicators intheScale-upPlan),whichappearstobetooshort.Further,

giventhattheproportionofSNV-EDGETsupportedhouseholdswithundertwochildrenis20%,itwill

bedifficulttoreachthetargetof65,000householdswithawomen-child(<2years)pairswithMIYCN

messagesthathavealsoreceivedsupportondairyproduction,processingandmarketing-buildingon

the potential synergies of the project’s dairy and nutrition component. This target may, therefore,

need to be scaled back. Furthermore, we would have liked to see an external and more rigorous

evaluationofoutcomesandcampaigneffectivenessbeforelargescaleroll-outofactivities.


Maternal,neonatal, child,andadolescenthealthandnutritionareoneof themajorpriorityareas in

Ethiopia’sHealthSectorTransformationPlan(2016-2020).2016datafromtheEthiopianDemographic

andHealthSurveyshowedthat38%ofchildrenunder5areconsideredshortfortheirageorstunted.

EDGETprojectresearchdocumentsfurthersuggestthewidespreadproblemofchildrennothavingthe

requireddietarydiversityandthatW-MDDSwasbelowcriticallimits.Allthingsconsidered,nutritional

behaviorandchangecommunicationinterventionsseemtobeveryrelevantintheEthiopiancontext.

Giventhedairyfocusoftheprojectandthehighnutritiousvalueofdairyproducts,follow-onprojects

in collaboration with the Ethiopian extension system seem to be well positioned to deliver such

interventions.


Incasethenutritionawarenesscampaignisrolledoutinfuturewerecommendtotesteffectivenessin

amorerigorousmanner.TheTIPsstudyremainsinconclusiveinthatregardduetodesignlimitations.


Asperthecurrentscaleupplan, thenutritionstrategywillbe implementedbyEthiopianAgriculture

andHealthExtensionworkerswithclosesupportbySNVEDGETnational,regional,zonalandworeda

staffs as well as the Eureka Health consulting team. The collaboration with existing government

structureshasthepotentialforsustainabilityofcapacitybuildingefforts.

4.7Cross-cuttingStrategies

Crosscuttingissuesincludethepromotionofwomenandyouthentrepreneurshipandclimatechange.

This section focuses specifically on the women and youth entrepreneurship component since the

climatechangecomponentisaddressedinrelationtoSO1andtheadoptionofclimatesmartpractices.

EVALUATIONREPORT

118

Womenandyouthentrepreneurship

Inordertostrengthenthepositionofwomenandyouthindairyvaluechains,theEDGETprojecthas

soughttopromotewomenandyouthentrepreneurshipinthedairyvaluechain.Thefocushasbeenon

promotinglocalbusinessinitiativesthatinvolveunemployedwomenandyoungpeopleininputsupply,

seedmultiplicationanddairyprocessingactivities.


Table42Achievementofprojectoutputsforwomenandyouthentrepreneurship


Endofprojectrevisedand(original)targets

Women&Youthdairyregional

andenterprisesestablished

Numberofwomen/youthdairy

enterprisesandgroupsestablished43/84% 51(51)

Women&Youthparticipationin

andleadershipoffarmer

organizationsandenterprises

promoted

Numberofwomen/youthdairy

farmerswhoreceivedleadership

capacitydevelopmentsupport

Theoriginalplanwascancelledandtheproject

focuswasshiftedtoengagingyouth&womenin

dairyproduction&marketingenterprisessupport.

InterviewswithkeyprojectstaffrevealedthattheEDGETprojectdidnothaveacleargenderstrategy

attheoutset,resulting indelayedprogressonthis front.However, In2015,farmerhouseholdswere

askedtosendwomentofarmertrainingsessionsand13DPUshadatleastonefemaleboardmember

plus onewomenout of two technicians hired formilk processing. The requirement forDPUsput in

placein2015werestrictlyfollowedin2016.Astudywasconductedtoassessgenderintegrationinthe

projectandthisledtothedevelopmentofanewgenderandyouthstrategyforEDGET,whichwasput

forwardin2016.

In2017,theprojectgavefurtherinstructionstoinvolve/womenHHmembersinthetrainingssessions

(either to invite women HH members or both husband and wife). This has been successfully

implementedinSNNPRandAmhararegion.Womenhavealsoreceivedattentionduringadvisoryand

coachingsessions.

AlsofortheAgIDs,womenapplicationsweregivenpriority.Butduetolimitednumberofapplications

fromwomen,only6outof51AgIDswereactuallywomen-ledin2015.Thewomenandyouthstrategy

remained to be finalised. This iswhy for 2016, only threewomen groups and 6 youth groupswere

established and / or supported for forage seedmultiplication andmarketing. This was short of the

targeted36women/youthdairygroups.Byendof2016,anassessmentonhowtoengagewomen

and youth for the project was completed, showing for instance that women struggle to acquire

improved breeds (EDGET Programme Gender and Youth Mainstreaming Strategy report in SNNP

regional state, 2016). A pilot for gender and youth ran in 2017. By the project end, 43women and

youthenterpriseshadbeenestablished(86%ofthetarget)andalsoreceivedequipmentandtraining

support.

EVALUATIONREPORT

119

Findingsfromthehouseholdsurvey

Findings related to thegendereddivisionof roles indairyactivitieshavebeenpresented inprevious

sectionsofthereportbutaresummarisedhereforconvenience:

• Womenaremore involved in labour-intensivedairyactivitiesassociatedwith lookingafter the

cows, milking them and processing dairy products. They are less involved in market-related

activitiessuchasbuyinginputsandmarketingmilkproducts.

• Female headed households were less likely to be members of DFEGs than male headed

households(42%vs48%)

• While far fewerwomenwere reported tohaveparticipated in thedairy related trainings than

men, an equal proportion of both male and female headed households received trainings,

indicatingthatfemaleheadedhouseholdswerenotdisproportionatelyexcluded.

• Statistically significant differences in the receipt of project inputs betweenmale and female-

headed householdswere only found for calf feed,with a higher proportion of FHH reporting

receiptofcalffeedthanMHH.

• Intermsof theadoptionofpractices,severalkeypracticeswerereviewed.Thesepertainedto

the feeding system, preparation of improved feeds, varying feeding based on the stage of

lactation,monitoringcows'milkproduction,producing forage,accessingartificial insemination

services, performing regular growthmeasurement, using supplementary feed, use of theMTS

formilking and transportation. The analysis revealed that adoption rates bymale and female

headedhouseholdswerenotstatisticallysignificant,exceptinthecaseofaccessingvaccination

(somewhathigherforMHHthanFHH).

• The figures for milk production are 1323 litres/year for Female Headed Households and 959

litres/year for Male Headed Households. However, the difference was not found to be

statisticallysignificantasaresultofthelargevarianceandthesmallsamplesize.

• NetincomefiguresforFHHandMHHare6486Birr/yearand6304Birr/yearrespectively.Aswith

milkproductionthedifferencewasnotfoundtobestatisticallysignificant.

• In themajorityofcases, the revenue fromthesaleofmilkanddairyproducts iscontrolledby

menandwomenjointly(56%),followedbywomenonly(34%)andmenonly(10%).

Overalltheresultsabovepaintamixedbutgenerallypositivepicture.Whilethereremainsomegapsin

women's membership of the DFEGs, female headed households appear to have benefitted equally

fromtrainingsandthereceiptofinputs.Onthewhole,FHHwerenotfoundtobelesslikelythanMHH

toadoptkeypracticespromotedbytheprojectandthisisreflectedinsimilarlevelsofmilkproduction

and net income for both MHHs and FHHs. Perhaps most encouragingly, in the majority of cases,

householdsreportedthattherewasjointdecision-making/controloverincomeearnedfromthesaleof

milkanddairyproductsatthehouseholdlevel.

However,whenaskedabouthowtimespentondairyactivitieshadchangedoverthelastfouryears,

42%offemaleheadedhouseholdsand49%ofmaleheadedhouseholdsreportedanincrease.Looking

intotheextentofthisincrease,ahigherproportionofwomenheadedhouseholdsreportedthattime

increasewasextremecomparedtomaleheadedhouseholds(28%vs.15%).MHHweremorelikelyto

saythattheincreasewasmoderate(50%vs.37%),whileMHHandFHHwereequallylikelytosaythat

EVALUATIONREPORT

120

the time increase was ‘slight’ (35% vs. 35%). The average time spent on dairy activities is 4.1h for

female headed households and 4.0h for male headed households (difference is not statistically

significant,p>0.1).

EVALUATIONREPORT

121

FindingsfromthequalitativeassessmentTable43Qualitativefindingsonwomen'sempowerment


Establishingwomenandyouthgroups/enterprises

DA:womengroupsof30membersestablished. They received training, churnerandrefrigeratorfromSNVinAletaWondo

DEP+DFEGMembers inallworedas:DEPandDFEGmembersvoicedthatchange forwomen participation in dairy has improved but is very visible (Dangila+Wuchale,Lemu) as they aremainly involvedathome, in feeding,milking, and transportation.Change isduetotraining,posters, increasedgovernmentattentiontothetopic,andalsotoSNVprioritisingthis.NewemploymentopportunitiesforwomenhavearisenatDPUs (Lemu Bilbilo). Dairy product incomemainlymanaged bywomen (because ofSNVtraining-LemuBilbilo).

AccordingtotheDAinLemu-Bilbio,agroupofsevenyoungwomenbuyskimmedmilkfromtheDPUeachdayandthenprocessesitintocheese.Thesegirlshavebuyersandgetbetterincome(thechangewasnotattributedtotheproject).

Repeatedlyitwasvoicedthatdairysectorshasstartedtobecomeincreasinglyownedby youth, especially in cooperatives, but also at household level across severalworedas.Thisisdrivenbytheprojectandfamilyexperiences.

InDangila,womenwerereportedtobe involved intheproductionofureamolassesmineralblock,whichtheyaresellinglocallytodairyfarmers.Anincreaseinwomen’sparticipation in dairy activities was also reported. DFEGmembers also reported anincreased role of youth in dairy related activities at the household level. Efforts toencouragewomen’sparticipationintrainingactivitieswerealsonotedpositively.

InMachakel,respondentsreportedanincreasedrepresentationofwomenintheMilkCollection Committees and in the dairy cooperatives more generally. Youthinvolvementinforageandmolassesproductionenterpriseswasalsoreportedtohaveincreased

InWuchale,anoverall trendof increasedparticipationofwomenandyouth indairyactivities - notably trainings and cooperatives-was reported to have increased as aresultofbothprojectactivitiesandthecurrentdrivebythegovernment.

DPUMachakel:nottoomanywomenparticipating.

MOLF Lemu: Awareness creation but no changes visible with regard towomen/youthparticipation

Difficulttoincludeyouthastheyarelandless(DEPLemuBilbilo)

WhenDEPcallfortraining,sometimesfamilymemberswhoarenotcarryingoutdairy-relatedworkcomefortraining.Forinstancewomencarryoutmoreof the home related dairywork but onlymales appear during the trainings(AletaWondo)

Various respondents across all woredas reported that the involvement ofwomenandyouthinthedairyvaluechainremainedlimited.Inmostworedas,respondents noted that the majority of beneficiaries were male and thatinsufficientattentionhadbeengiventocreatingopportunitiesforwomenandyouthengagementindairyrelatedactivities.

EVALUATIONREPORT

122

AssessmentofRelevance

Strengthening the position ofwomen and youth in the dairy value chain through the promotion of

womenandyouth-ledenterprisesandinitiatives,holdsthepotentialtoaddresssomeofthekeyissues

facedbybothgroups.However,thisevaluationdidnotconductaspecificassessmentofthesewomen-

ledgroups,withtheresultthatlimitedconclusionscanbedrawnregardingtheseefforts.

Whileanenterprise-ledapproachholdssignificantpotentialforempoweringwomeninpublicspaces,

evidence fromtheendlinehouseholdsurveyandfromthequalitativesurveysuggest thathousehold

level gender dynamics and norms are a key area that require further attention. Most findings

highlightedthatwomencarryoutthemajorshareofdairyrelatedactivities(particularlylookingafter

the cows,milking themand producingmilk products) at the household level, adding significantly to

theirexistingdomesticwork.Atthesametime,womenwerefoundtoparticipatelessintrainingsthan

menandtohavelessofaroleindecision-makingrelatedtoeconomicmatterssuchasthepurchaseof

inputsandthemarketingofmilkproducts.

AssessmentofEffectiveness

Anin-depthassessmentoftheeffectivenessofthewomenandyouthenterprisedevelopmentwasnot

carriedoutaspartofthisevaluation.Analysisiscomplicatedbytheabsenceofacleargenderstrategy

duringtheinitialstagesoftheinterventionandthedelaysinimplementationofthestrategy.


An in-depth assessment of the sustainability of the women and youth engagement in enterprise

developmentwasnotcarriedoutaspartofthisevaluation.

Keyissuesforconsiderationinasubsequentphase

Anyfuturedairy-relatedinterventionneedstointegratethegendercomponentmuchmorecentrallyin

thedesignofactivitiesfallingunderStrategicObjectives1and2.

Attentionshouldgobeyondengagingwomeninvaluechainactivitiestoaddressinggenderdynamics

at the household level, particularly given women’s central role in dairy-related activities at the

householdlevelandwell-establishedevidenceregardingthe‘doubleburden’.

The inherent trade-offs between productive and reproductive work of women in these households

shouldbeinvestigatedinhouseholdswherewomenplayamoreprominentrole.Thisclearlyalsohas

implicationsforthenutritionallevelofchildrenandtheamountofmilkthatisconsumedbyhousehold

membersincludingchildren.

Findingwaystoaddresswomen'sburdenoflabourrelatedtodairyischallenging.Whileengagement

ofmen,includingmaleyouth,offersonepossibleavenue,thisalsorisksunderminingwomen'scontrol

overincome.Thepossibilitiesofahouseholdleveldairyfarmthatemploysmorepeople(andthereby

resultsinareducedlabourburdenforeachindividualandgeneratesincomeforemployees)maynot

beviableforthegreatmajorityof individualhouseholds.However,theestablishmentofprofessional

servicesrelatedtodairying–forexamplemilking,milkcollection,feeding,etc.-couldprovideameans

EVALUATIONREPORT

123

toreducethisburden.SuchservicescouldbelinkedtotheCooperativeDPUsorcouldseektoengage

traditional/informal traders. In somecountriesmodelsofdairyingbasedoncontract farming (where

dairycowsarekeptinacentrallocationandprofessionallymanaged)arealsobeingtested.Thecruxis

thatsolongasthelabourentailedindairyingcannotbereduced,outsourcingofspecificactivities(e.g.

toprofessionalserviceproviders)maybetheonlyviableoption.

Climatechange

Asacross-cuttingissuewereportclimatechangeundersection‘Climatesmartpractices’above.

EVALUATIONREPORT

124

5. Inclusive, sustainable dairy value chain development

inEthiopia

Value Chains consist of a series of transactions and actors necessary to bring a product or set of

products from the input stage to the finalmarket. Every part and interaction in that chain involves

addingvalue,andcollaborationamongactorsandtheir stakeholders.Thekeygoal is to findways in

whichthesesustainablevaluechainscanhavesignificantimpactsonthefoodsystem(i.e.availability

ofmilkandmilkproducts) throughenhancingeconomicstabilitybycreatingvalue (higher incomes,

more diversified livelihoods, on and off -farm income streams, better nutrition, women and youth

economic empowerment), social sustainability by facilitatingmore equitable distribution of addedvalue among stakeholders (farmers, woreda livestock officers, DPUs, DFEGs, processors, AgIPs, AI

technicians, EDGET,women, youth) anddriving environmental sustainability by reducing ecologicalfootprintsthroughoutthevaluechain.

Whilsttheearlierpartofthisevaluationfocusedonthecomponentsofthesubsystemundereachof

the strategic objectives, this section looks at the interconnections and interdependencies necessary

betweenthedifferentcomponentsofthesystemtoenablethesystemtofunctionasawholeandbe

sustainable.

Household level outcomes may be sustainable either (1) because the project provided material or

technicalsupportthatcontinuestobeusedandproducebenefitsintheabsenceoftheproject,or(2)

because theproject contributed to thedevelopmentof functioningand sustainable subsystems that

continue to deliver critical services and inputs to farmers beyond the life of the project. A critical

aspectof this secondaspectof sustainabilityalso relates to theeconomicviabilityofdairyingat the

householdlevel.Whiletheresultsintermsofnetincomeatthehouseholdlevelareencouraging,itis

difficulttodeterminetheextenttowhichthisisbolsteredbytheprovisionofinputsuchasforageseed

thathaseffectivelycreatedmonetarysavings for farmers. It is importanttonotethatcalculationsof

net income carried out using the household survey data do not factor in the opportunity cost of

reallocatingagriculturallandtoforagecultivationfromfoodcrops.

There are four main aspects identified in this section where improvements will benefit the system

workingasawholeandcanguidefutureinterventions.Theseare:

• Thekeyactorsdrivingbettermilkproductionworkingininterconnectedways;

• Abetterunderstandingofhowmilkmarketsareworking inEthiopia,especially in theworedas in

question;

• Ensuringthatthereareclearandcontextualisedplansforengagementinthesemilkmarkets;and

• InvestinginanM&Esystemthatdrivesevidencebaseddecisionmakingandgeneratesinsightsthat

arenimbleandhelpfulinguidingimplementation.

EVALUATIONREPORT

125

5.1KeyActorsDrivingBetterMilkProductionfromProducers Theextensionand input relatedactivitiesexaminedduring thisevaluation foundthatactivitieswere

well-received and led to tangible benefits for actors such as government extension officers, feed

suppliers,AgIDsandthedairyfarmersthemselves.Whilefindingsfromthehouseholdsurveyindicate

that the EDGET project has created considerable success in the adoption of improved practices,

particularly with respect to the adoption of forage cultivation practices, the sustainability and

replicabilityoftheextensionservicedeliverymodel(s)isstillunclear.

In looking at various actors, the performance of DFEGs was mixed, DAs were found to be quite

overstretched and dairy was often not a high priority within the overall woreda extension office

(understaffed,under-budgetedandoften competingwithmoredominant crop-productionactivities)

andthereappearedarelianceontheEDGETprojectDEPs.

SNVhashadsignificantpresence,withaDEPposted ineachworeda.Whilst theseDEPshaveplayed

extremely valuable and important roles, complementing DAs and even compensating for their

shortcomings, their presence or persistence as part of the model raises questions. Is the model

dependent on having such a large, dedicated and highly skilled cadre of officers in theworeda? To

whatextenthavetheskills,resourcesandcapacitiesrequiredbeentransferredtotheDAsandWLOs

andwhatlightertouchsupportmodalitiescouldfillthegap?

Inthecontextofextensionsupport,theDFEGsemergeasvulnerableelementsinthesystem,linkedto

variable leadership, limited capacity, lack of role clarity within the groups and limited incentives to

continue functioning beyond the life of the project. However, the DFEGs have proven integral for

extendingthereachofdairyextensionservicestoalargerpoolofdairyfarmersthanwouldotherwise

be possible and their attributes include proximal enduring relationshipswith dairy farmers. Further

analysisiswarrantedtounderstandexactlywhatdrivesthesuccessofDFEGsindifferentcontexts,the

natureof incentivesthatcanbeintroduced(forexample,throughpublicrecognitionofDFEGleaders

with high group performance), how the effectiveness of these can be assessed and how less well

performingDFEGscanbeidentifiedandstrengthenedintimelyandappropriatemanners(forexample,

byrotatingtheleadershiporaddressingconflictsorunproductivegroupdynamics).

EffortstostrengtheninginputsystemsthroughAgIDsprovedtobeverysuccessfulinenablingAgIDstoestablish viablebusinessmodels and topursuegrowth. The linkagesestablishedwith feed suppliers

wereseentobemutuallybeneficialforallparties,andverywellsupportedbythelinkagesestablished

betweenAgIDs and farmers.While some challengeswere found related to pricing, quality, quantity

and variety of feed, the results from the household survey - for example pertaining to farmers

willingness to continue purchasing calf feed without project support - are very encouraging and

potentially illustrate elements of valuable potential sustainability. The challenges for these actors is

how toenable their growth, increaseprofitability,haveaccess to reliable finance,diversifyproducts

andservices,andincreasetheirreachtounderpinsustainabledeliveryofinputsandrelevantservices.

Theuptakeofforageproductionbyfarmershasproventobesuccessful,withasignificantnumberof

farmersgrowingadditional foragecrops.Farmer-ledseedproductionoffersanavenuefor increasing

the supplyof forageandpotentially creatinganadditional livelihoodactivity that isprofitable. Their

EVALUATIONREPORT

126

challenges relate to shortages in the supply of forage seed, littlemarket information on the forage

market and unclear demand for forage seed as a product in itself. These aspects continue to

underminethesustainabilityofthesystemasawhole.Goingforward,astrongemphasiswillneedto

beplacedonexpandingthesupplyofseedpotentiallythroughadiversityoflargeractorsandaccessto

reliablefinanceandmaybeworkinginamoreformalarrangementwithfarmerssupplyinghighquality

forageseed(contractfarming)anddemonstratingtheviabilityofamarketforforageseedatscale.

Throughout the EDGET project, SNV maintained very strong and constructive relations with the

Ethiopiannationalandregionalgovernments.Indeed,theoverallmodalityofprojectdeliverywasto

work through the government linedepartments - in this case through theMinistry of Livestock and

FisheriesatandtheregionalBureausofLivestockandFisheries.Thisledtoverypositiveatmosphereof

collaboration,withgovernmentrepresentativesatthenationalandregionallevelsclearlyfeelingthat

theprojecthad contributed toenhancing their capacityandaddressing specific knowledge, skill and

materialgaps.At thesametime, thegovernmentrepresentativesalsorecognisethat it is theroleof

thegovernmenttodrivethework inthedairysectorandtoscaleupthe interventions.Theybelieve

thatthesupportfromEDGEThasbeenalignedwithgrowingimportancetolivestockanddairy inthe

Ethiopian Agricultural Growth Plan and has expanded their ability to do this - and that the EDGET

project is a good approach. The key constraints identified at the government level are related to

prioritisationof livestockanddairyactivitieswithinthebroaderAGP,humanresources(quantityand

capacity),budgetallocationfordairyactivitiesandtheabsenceofawell-developedsectoralregulatory

andpolicyframeworkfordairydevelopment.

Thecoreservicesprovidedbythegovernmentincludesoverallplanning,coordinationandoversightof

dairyrelatedactivitiesandprovidingAIandanimalhealthservices,withthemajorityofEDGETproject

supportgoing into technicalandmaterial support for theprovisionofAIservices.Evidence indicates

that this support has been very relevant (filling in key gaps) and at least moderately effective

(increased service delivery capacity). It is also likely to be sustainable. However, key gaps in the

quantityandqualityofinputsupply(forbothnitrogenandsemen)constitutebottlenecksthatneedto

beaddressedforviabledairydevelopmentactivitiestobecarriedoutatscale.Morebroadly,overall

tracking of herd size, genetic composition, etc., are important factors that need to be tackled for

effective,modernmanagementofdairydevelopment.

Beyond this, the government has a role to play in developing the infrastructure (e.g. roads and

electricity)thatunderpinsandde-riskeffectivedairymarkets,forensuringthatnaturalresources(e.g.

waterand land)areallocated inamannerthatsupportstheoveralldevelopmentof thedairysector

and that appropriate regulatorymechanismsare inplace toensure thatquality standards in service

provisionaremaintained.Thisrequirescross-ministerial/bureaucoordinationandalignmentonaclear

policyfordairyvaluechaindevelopmentinthecountry.

5.2IntegrationandInterdependenceofcomponents Whenconsideringtheoverallsustainabilityofthedairyvaluechain,itisnecessarytolookholisticallyat

(a)thesustainabilityoftheindividualsubsystems(extension,forageseed,concentratefeed,AI,etc.),

(b)thewaysinwhichthesesubsystemsintegrate,interrelate,reinforceorsustainthedairyvaluechain

EVALUATIONREPORT

127

inadynamicwayand(c)theextenttowhichthereisanenablingenvironmentforthedevelopmentof

the dairy value chain including functioning output markets and supporting policies. The following

pointsillustratetheseinterdependencies:

• Withoutcross-breedcows(andsoanefficientandeffectivesystemforbreedimprovement),milk

production cannot reach the critical threshold required to create a viable and vibrant dairy

market.

• Withouttheavailabilityofquality,affordableforage,cowandcalfhealthmilkproductionisatrisk.

• Withoutsufficientdemandforqualityaffordableforage,privateproviderswillhavelittleincentive

toproduceforageorforageseed.

• Withoutaccesstoremunerativemilkmarkets,thedemandforcrossbreedcowsandforforagewill

belimited.

• Withoutapredominanceofmarkets/buyersthateffectivelydiscriminateandpayforthequalityof

milk supplied, farmers incentives to keep crossbreed cows and invest in them will be limited,

therebyfurtherlimitingthesupplyofqualitymilk.

Figure30below,presents the key challenges reportedby farmers in theendline survey, inorderof

decreasingfrequency.Theresultsshowthatthepatternofchallengesreportedisverysimilarinmost

cases in both treatment and comparisonworedas,with issues related to the availability and cost of

forageandconcentrate feedranking in thetop3challenges.Statisticallysignificantdifferenceswere

only found for the high cost of forage feed (more frequently cited in intervention woredas),

unavailability of AI services (more frequently cited in comparison woredas), lack of information on

dairy cow management (more frequently cited in comparison woredas), a low success rate of AI

services(morefrequentlycitedininterventionworedas),alackofaccesstoveterinaryservices(more

frequently cited in intervention woredas) and a shortage of labour (more frequently cited in

interventionworedas).While the responses reflectdairy farmers'perceptionsand theseperceptions

are likely to have been influenced by the intervention, the results reveal that the challenges faced

relatetodifferentsub-systemsofthedairyvaluechain.Challenges inanygivenarearisk limitingthe

farmers' ability to improve their dairy related practices, increase milk production and earn more

income from dairy activities. This points to the importance of addressing the interdependencies in

termsofhowtheyaffectdairyfarmers'optionsandchoices.

EVALUATIONREPORT

128

Figure 30 Proportion of respondents by comparison and intervention group that mentioned dairyproductionrelatedchallenges(*p<0.1,**p<0.05)

It is clear that the EDGET project did consider such interdependencies. The complementary mix of

interventions and activities targeting different subsystems of the dairy value chain reflects this.

However,thechallengefortheEDGETprojectistotranslatethecomplementarityofinterventionsinto

functionaloroperational integrationamongst thekeyactors in thevaluechainacrossdifferentsub-systems.Inparticular,thismeansensuringthatthelinksbetweeninputsuppliers,extensionworkers,

AgIDs, dairy farmers, and cooperatives with DPUs are properly established and functioning

synergistically.Anybottlenecksthatmayariseinonesubsystem(e.g.duetoashortageofinputs,weak

leadership, lack of electricity or equipment for dairy processing, etc.) will cascade into the others,

limiting the extent to which the whole can function as intended. Thus, for example, delays in the

supply of equipment to DPUs limited the extent to which the dairy cooperatives could play their

envisagedrole.

The complexity of the dairy value chain and the multiple interdependencies between its

subcomponentsmakesitdifficulttorepresentthesystemasawhole,particularlyintheformofamore

linearTheoryofChangeorresultsframework.Byfocusingontheactorsinthesystemandtheirroles

63%

61%

43%

50%

35%

37%

43%

36%

26%

22%

33%

18%

27%

18%

19%

14%

17%

13%

10%

66%

63%

49%

44%

43%

41%

36%

36%

32%

27%

26%

24%

24%

22%

21%

20%

20%

19%

13%

0% 10% 20% 30% 40% 50% 60% 70%

Unavailabilityofforagefeeds

Unavailabilityofconcentratefeeds

Highcostofconcentratefeeds

Unavailabilityofmedicinesandotherrelatedhealthservices

Highcostofforagefeeds**

Lackofspaceforlivestock

UnavailabilityofAIservices*

Lackofaccesstomarketstosellmilk

Highcostofmedicinesandotherrelatedhealthservices

HighcostofAIservices

Lackofinformationondairycowmanagement*

LowsuccessrateofAIservices*

Lackoffacilitiesformilkprocessing

Low/fluctuatingmilkprice

Other

Lackofaccesstoveterinaryservices*

Lackofskilledtechnicians

Shortageoflabour*

Toomany/illegal middlemen

Comparison Intervention

EVALUATIONREPORT

129

(aspresented in the actormaps in section4of the report), it becomespossible to arrive at amore

accuratepictureoftheinterdependenciesthatareatworkinmakingthevaluechainwork.However,

combining all the actor maps together again leads to a high level of complexity that is difficult to

represent.

Inordertoapproachtheoverallvaluechainasan integratedmap,amodelof thekeyvariablesthat

makeupthesystemcanbeconstructedthatspansmultiplesubsystems,reflects thedifferentactors

andshowthedirectionofthefeedbackloopsthatcharacterisetherelationshipsbetweenthedifferent

variables. This not only reveals the interdependencies and relationships within the system but also

providesanindicationofthekeyparametersthatneedtobeassessed,measuredandtrackedinorder

todesign,monitorandevaluateinterventionswithinthesysteminaneffectivemanner.

5.3AClearerUnderstandingofMilkMarketsandtheirdynamics Efforts to promote the processing and marketing of dairy products, specifically by supporting the

establishment and strengthening of cooperatives with DPUs and Milk Collection Centres, have

establishedtheearlygroundworkforasubsequentphase,buildingonthelessonslearned.

While this evaluation was not able to show the anticipated gains in terms of volume of milk

processed/marketed and income earned as might have been hoped for, the timeframe of the

evaluation with respect to implementation (many DPUs saw significant delays in receipt of dairy

processing equipment) is an important consideration in making overall judgements regarding this

component.

FindingsfromthequalitativecomponentoftheevaluationrevealthatDPUsandmilkcollectioncentres

have thepotential to function reasonablywell, provided there are strongoutputmarket linkages in

place, competent and accountablemanagement committees. This almost certainly requires external

support -aboveandbeyondwhatmostworedacooperativeagenciesarecapableofproviding -and

will most likely continue to require such support into the future. Appropriate mechanisms for

deliveringsuchbusinessandorganisationaldevelopmentsupportwillneedtobedevised.Muchofthis

willbenecessarywithakeeneyeontheenablingenvironmentandthematurityofoutputmarketsin

differentcontextsthatdeterminetheirabsorptivecapacitiesforbothmilkandprocessedproducts.

Considering the status of milk markets more broadly, evidence from this evaluation points to very

weak formal outputmarkets formilk. This presents a rather confusing picture, as one of themajor

assumptions underpinning the project is that stable demand for milk at sufficient scale will drive

mediumsizedbusinessesformilkaggregationandprocessing.

On the one hand, regional milk processors reported limited (albeit growing) consumer demand for

packaged milk products. On the other they reported limitations regarding a sustainable supply of

qualitymilk.Concurrently,DPUcooperativesfrequentlycomplainedaboutthelackofmarketsfortheir

milk.Themajorityof farmerswere foundtobeselling theirmilk to informal tradersand individuals,

whowerereportedtoprovidehigherpricesthancooperativesorprivatesectorbuyers.Whileinformal

tradersarewidelycriticisedfordrivingdownqualitystandardsformilkquality,theytypicallypaymore,

andpayfasterthanotherbuyers.Theircontributiontoincreasingthehouseholdincomeofsmallholder

EVALUATIONREPORT

130

dairy farmers is often under-acknowledged. Inherently, this provides strong incentives that drive

farmerbehaviour.

Giventhecontextualdiversityinmilkmarkets,futureworkonstrengtheningthedairyvaluechainmay

benefitfromamorenuancedsegmentationoftargetworedasbasedonthematurityofmilkmarkets

withinthemandmodelsofinterventionsthataretailoredtothecontextualstrengthsandweaknesses

withinthoseworedas.Thismayinclude,adoptingcontextspecificstrategiesfornascentmilkmarkets,

emergingmilkmarketsandmoredevelopedcommercial/formalmilkmarkets.

5.4PrinciplesforDPUandMilkCollectionCentres

Targeting and selection of woredas and kebeles for establishment of DPUs requires careful

considerationtoensurethatDPUsareestablishedinlocationswheretheyhavesufficientpotentialto

develop.Thismeansthatissuessuchasroadconnectivity,electricitysupplyaswellasaccesstoviable

marketsandasufficientactualvolumeofmilkproduction in thecatchmentareashouldbecarefully

assessed.FailuretoadequatelyconsiderallthesefactorslimitstheabilityofcooperativeswithDPUsto

succeed and risks contributing to negative perceptions about their relevance/utility amongst dairy

farmers.

Linkages with buyers: Cooperatives with DPUs can only serve their purpose if they have stablerelationshipswithbuyerswhocan(a)absorbthevolumeofmilkthatfarmersneedtosupplyand(b)

offeracompetitiveprice.Thecapacityofthecooperativestoestablishtheselinkagesthemselvescan

bequitelimited.Athoroughmarketassessmentandclearcommitmentsfrompotentialbuyersmaybe

aprerequisitetosuccessfuldevelopmentofthecooperativeswithDPUsnrespectivecontexts.

Organisationalcapacity:TheorganisationaldevelopmentofDPUsislikelytobealong-termchallenge.

Appropriate mechanisms for providing business, technical and management support to the

cooperatives is essential.However, the requirementsmaybebeyondwhat theWoredaCooperative

Agencies are capable of providing. As such, alternative support organisations may need to be

established that can play the required role across multiple woredas, creating alignment in the

approach and facilitating standards and best practices across these cooperatives. This is a role that

couldevenbeplayedby largerprivateplayerswhodecideto linkupwiththeDPU/cooperativesand

potentialpublic/privatepartnershipsmodelsalso.

SequencingofDPUdevelopmentactivities:ThesequencingofDPUdevelopmentactivitiesshouldbe

carefully managed. DPUs need to have simultaneous access to appropriate equipment (e.g. for

refrigeration)andlargerscalebuyersinordertoofferaviablemilkmarketfordairyfarmers.Failures

oneitherfrontcanunderminetheoverallfunctioningoftheDPU.

5.5PlanningandStrategy Anyfutureinterventionwouldbenefitfromhavingclearerstrategiesorasetofmodelsthateffectively

stratifyorcategorise interventionsitesbasedonarigorouscontextanalysis.Havingaclearmodelof

interventiontailoredtothecontextindifferentworedasmayhavevalueinhelpingbothtounderstand

EVALUATIONREPORT

131

actor capabilities,whichactorsarepresent,howtheywork togetherasa systemandenhanceactor

interactionsaswellestablishaclearerunderstandingofthematurityofthemarketsinthosecontexts.

Such plans and strategies should adopt a phased or tiered approach with different intervention

approaches for woredas that have different characteristics in terms of milk production, forage

production and market development potential. Other key issues to consider include infrastructure

availability,andaccesstocapital.Anexamplecategorisationisprovidedinthetablebelow:

Table44Characteristicsandapproachesfordifferentworedacategories

Category Characteristics Approach High

commercial

potential

Woredaswithhighmilkproduction,forageproductionandmarket

development.Theseworedaswillbelocatedclosetowell-developed

milkmarketswithestablishedmilkbuyersandprocessors.Theywill

haveahighlevelofintegrationbetweeninputsystemsandoutput

markets.Thedemandformilkwillserveasstrongincentivefordairy

farmerstoengageinandinvestindairyactivities.

Inclusivecommercialbusiness

modelssuchascontractdairy

farming;requiresintegrationwith

largerscaleprivate/cooperative

enterprises

Emerging

inclusive Woredas with high milk production potential, forage production

potential and an intermediate level of market development. These

woredas may be on the periphery of more dynamic and established

milkmarketsandwillhaveanumberofvaluechainactorsoperatingat

variousscalesandlevelsofperformance.

Emphasis on fostering linkages with

key actors within the established

milksheds

Nascent

informal Woredas with high milk production potential, forage production

potential andanascent level ofmarket development. Theseworedas

may lie outside established milksheds yet have access to small and

medium sized townswith growingmilkmarkets. They are likely to be

servedbyweakinfrastructureandhavelessdevelopedvaluechains.

Emphasis on developing milk

production potential, household

levelconsumptionandinformalmilk

markets

The key insight here is that a systematic approach to differentiating what types of interventions,

impact pathways and indeed results are appropriate for different segments. Such an analysiswould

helptoensurethat interventionsaremorecloselyalignedwiththeactualopportunitiesandthatthe

emphasis of resource allocation can be placed on addressing the most critical bottlenecks in a

sustainablemanner.Thealternativetothis,whichtosomeextentreflectstheEDGETprojectapproach

todate, istohaveasingledairyvaluechainapproachwithsomedegreeofadaptationbasedonthe

context.While this has been valuable in terms of generating insights and learning, as documented

throughoutthisreport,scalingsustainablywillrequireamoresystematic,stratifiedstrategicapproach.

5.6Metricsandaversatileandeffectivelearningsystem A new strategy for engagement in the livestock sector should be underpinned by a performance

measurementandlearningsystemthatiscloselyalignedtothenatureofthesystemanditsdynamics.

While the tracking of activities and periodic data collection on key quantitative outcomes provide

usefulevidenceofprojectprogress,theydonotnecessarilyprovidesufficientinsightintothewaythat

thedairyvaluechainisfunctioningasasystemorwherethekeybottlenecks,risksandopportunities

are situated. Moreover, they tend to focus on relatively more linear pathways from activities to

outputs and outcomes rather than providing insights into the interdependencies between different

componentsofasystemandthequalityofthoseinterdependencies.

EVALUATIONREPORT

132

Amoresystemicapproachtomeasurementandlearning isadvocated,tiedtoastratifiedstrategyor

tailoredsetofmodelsforintervention.Suchasystemshouldprovideinsightsintohowthedairyvalue

chainasawhole–acrossdifferentsegments–isactuallyfunctioning;bothintermsoftheindividual

actors within the system and in terms of the extent to which there is effective integration across

differentactorsinthesystem.Asetofharmonisedindicatorsrelevantacrossthestratawouldprovide

practical insights intotherelativeperformanceofeachsegmentandguidedecision-making. Itwould

provideapictureacrosstheEthiopiandairyvaluechainthathelpsprovideabetterunderstandingof

theincentivesanddriversforamoreformal,inclusiveandsuccessfulvaluechain.

Underpinning this is the need for high quality data collected through accurate and rigorous context

analysisandwell-designedcredibleandmeaningfulbaselinesandperiodicevaluations.Itisimperative

that such studies are carriedout in a thoroughand systematicmanner thatprioritises accuracy and

quality.Thisdemandsthatsufficienttimebeallocatedtothedesignandroll-outoftheevaluationsand

also that the evaluability of the strategy is considered during its design. This would enhance the

credibilityandutilityofresultsandgearstudiestogeneraterobust insightsonthefunctioningofthe

systemandtheimpactsthatarebeingachieved.

Moving froma linear resultsmeasurementmodel toone that encompasses the feedback loops and

interdependenciesofasystemrequiresclearerrepresentationoftheexpectedpathwaystoimpactof

theinterdependentsystem.Thiscanbeachievedthrough:

1. Stratifiedmodelsof interventionwithcontextualisedTOCsandactormaps,assumptionsand

externalitiesdesignedtofitspecificcontexts

2. Aharmonizedstrategyofmeasurement tocomparemodels for thekeyattributes leading to

functionalsystemsindifferentcontextsandacrossthenationalsystem

3. A set of metrics and a measurement system that provide timely evidence on how

subcomponents of the systemareworking, individually and in combination, tomake course

correctionsandadjustmentsthatcanbedeliveredpromptly

4. Studies that provide clear directiononmarket potential andopportunities formilk andmilk

productsales

Documents

EDGET project final evaluation report FINAL1 - SNV · 2018-08-31 · The project also includes two cross-cutting objectives: a) To promote women and youth entrepreneurship and b)