1

Combined findings and recommendations from SSSAs led by:

FAO & Government of Ethiopia

CRS & partners

January2017

Seed System Security Assessment (SSSA)

Ethiopia 2016

Synthesis report

2

Background

Twolarge-scaleSeedSystemSecurityAssessments(SSSAs)wereconductedinEthiopiabetweenSeptemberandDecember2016.TheFoodandAgricultureOrganizationof theUnitedNations(FAO) Ethiopia CountryOffice andGovernment of Ethiopia led one assessment,with CatholicReliefServicescoordinatingtheotheralongwithitspartners1. Asbothassessmentsexaminedfarmers’ seed security in detail overmultiple seasons and used closely related tools, there isconsiderableoverlapintheirfindings.Thissummarysynthesizesfindingsandrecommendationsfrombothassessments in theareaswhereoverlap isapparent. It focuseson the impacts forsmallholder farmersofacutestressduring2016belg2andmeher3seasons,andprojections forthe2017belg season, so as to guide immediate implementationoptions for 2017. However,somecommentsaroundlonger-termissuesarealsoincludedattheendofthisreport.

The individual SSSA reports provide more detail than can be covered in this brief synthesis.Also,eachassessmentemphasizesslightlydifferentaspectsofseedsystems.Individualreportsareavailableat:

• FAO & Government of Ethiopia: http://www.fao.org/publications/en/(contact:Alemu.Manni@fao.org)

• CRS&Partners:http://seedsystem.org/wp-content/uploads/2017/01/Ethiopia-SSSA-Final-Report-Dec-2016.pdf(contact:matt.davis@crs.org)

DISCLAIMERTheauthors’viewsexpressedinthispublicationdonotnecessarilyreflecttheviewsoftheUnitedStatesAgencyforInternationalDevelopmentortheUnitedStatesGovernment.

1PartnersincludedISSDEthiopia,theReliefSocietyofTigray(REST),theOrganizationforReliefandDevelopmentin2ShortrainyseasongenerallybetweenFebruaryandMay.3Longrainy seasongenerallybetweenJuneandSeptember.

3

TheSeedSystemSecurityAssessments(SSSAs) TherationaleforconductingSSSAsatthistimewasthreefold:

• In 2015, Ethiopia farmers and systems experienced one of the worst droughts in 50years (comparable to the 1983 - 1985 drought) --- in large part due to El Niño(HumanitarianRequirementsDocument,2015).

• TheGovernmentofEthiopia(GoE)andotherhumanitarianpartnersrespondedswiftlytothecrisis,distributing32000tonnesofassortedcropseedduringthe2016belgandmeherseasonswhileplansfor2017arestillunderdevelopment.TheSSSAaimedtoassistmanagersandfieldstaffassesswhetherimmediateseedsysteminterventionswereontrack,aswellastobuildtheirseedsystemsecurityassessmentcapacity.

• Seedsecurityissuesarelinkedtofoodsecurityissues,butalsohavequitedistinctfeatures.Theseassessmentsweredesignedtoprovidehonedtechnicalinsightandtoshapetargetedinterventionsintheshort,mediumandlongterms.

FAOandtheGovernmentofEthiopia’sSSSA(henceforthSSSA-FAO)ranfrom31October2016to3December2016. SSSA-FAO covered two zones in eachof fivedifferentRegions4: Tigray,Oromiya, Amhara, Southern Nations, Nationalities and Peoples Region (SNNPR), and Afar, aswellasasinglezoneinbothSomaliandGambellaRegions.Intotal,SSSA-FAOinterviewed1270farming households, visited 17 markets, and undertook 45 focus group discussions. A fewhigherpotentialareas(e.g.EastGojjamZone,Ada’aWoredanearBishoftu)wereincludedintheSSSA-FAOsample,asitextendedbeyondareasofNGOactivity. CRSandpartnerscarriedoutaSSSA inEthiopiafrom28September2016to14October2016.Thisassessment(henceforthSSSA-CRS)tookplaceinfourRegions:Tigray,Oromiya,AmharaandSNNPR. In each region twoworedas were selected to represent a range of agro-ecologies,meherandbelgseasons,andtolinkwithpartners’areasofoperation.Intotal,486householdinterviews were conducted along with 46 seed trader/agro-dealer interviews, focus groupdiscussionsandcommunitymeetingsineachselectedregion.Thus,thecombinedSSSAsinterviewed1756farminghouseholds,thelargestnationalsampleofany SSSA performed to date. In both SSSAs,woredas and kebeles were selected to highlightdiverse factors suchas agro-ecology, tradeor stress, andhouseholdswere randomly selectedforinterviews.Bothtriangulatedfindingsusingmultiplemethodsandestablishedquantitativeand qualitative approaches, including conducting key informant interviews with many actorsinvolved in seed production, farmer assistance and value chains were utilized. These SSSAsreviewedthefunctioningoftheseedsystems,whichfarmersuse,bothformalandinformal,andassessedwhetherfarmerscouldaccessseedofadequatequantityandqualityintheshortandmedium term. Specifically, the work reviewed the actual seed sources farmers used andquantities of seed for their key crops for the 2016 belg and meher seasons, and farmers’projected seed sourcing for 2017 (belg only for SSSA-CRS, both seasons for SSSA-FAO). BothSSSAsusedsimilarsurveyinstrumentsforquestionsregardingseedsources,inputuse,seedaidexperience, and access to new varieties – though SSSA-CRS explored two crops per season,

4Ethiopiaiscomprisedofnineregionalstates:Tigray,Afar,Amhara,Oromiya,Somali,Benishangul-Gumuz,SNNP,andHarari,alongwithtwocharteredcities,AddisAbabaandDireDawa.

4

whileSSSA-FAOaskedaboutthreecropsperseason.Exceptwhereotherwisestated,allfiguresinthissynthesispoolnationwideresultsfrombothsurveys.SelectSSSAfindingsarepresentedbelow,acrosssites,andgroupedintoshort-termfindingsbyseason(Acuteseedsecurityfindings,BelgandMeher2016,andBelg2017).

5

I. ACUTE SEED SECURITY FINDINGS: SSSA Despite the initial shock (or shocks), diverse indicators suggest the seed security of EthiopianfarmersinthesevenseparateRegionshasbeenstablefor2016belgandmeherseasons—andisprojectedassignificantlyrecoveringforthebelg2017season.BelgSeason20161. Forbelg 2016, farmers sourced the lion’s share of the seed they planted (77 percent) by

themselves. The twomajor seed source channelswere informal sectorones: home-savedstocks (40 percent) and seed purchased from localmarkets (35 percent). This is striking,giventheimportantemergencyseedaidactivitiesinallbutafewoftheworedassampled;seedaidprovided20percentoftheseedplantedduringbelg2016.

2. SSSA-CRS compared quantities sowed for each crop against normal amounts (number ofcomparisons=523); forbelg2016,quantitiessownwere in therangeofnormal (anoveralldipof6.0percent).Lessthanhalfofrespondents(40percentofallcases)didsowlessofagivencropinbelg2016:farminghouseholdswereaskedtheirmainreasonforsowinglessthannormalforeachcropcase.Amongthisgroupofpotentiallyvulnerablefarmers,threereasonsweregivenasparamountforreducingsowingamounts.Nomoneytobuyseed(15percentof responses), insufficientaccess to land for theseason (20percentof responses)and simply poor weather (37 percent) - which deterred farmers from risking sowing fullamountsofthecrop.Apositivedevelopmentwasthefourthmajorfactorcited:useoflessseeddue tobetteragronomicpracticesof rowplanting. All four regionsprovided similarreasons for reductions,witha lackofoxenadditionallybeinghighlighted in SNNPR. Poorseedavailability(inmarkets,shops,withneighbors)wasan insignificantreasonforsowingless(lessthan1percentofresponsescitedthisreason).Similarly,SSSA-FAOalsofoundnoevidence of constraints to seed availability. When asked if each seed source provided“enough seed”, only 12percent responded “no” (out of 662 responses). Absolute seedunavailabilityislikelylowerthanthis,asfarmersusemultipleseedsources.

Note:Directdistributionsofseedaidnormallyaddressanabsenceofseed(seednotavailable).Evidence suggests thiswas not farmers’main constraint during that season. While free seedmight mitigate financial constraints, such direct seed aid would not address the two drivingproblemsfordeclinedseeduse--poorweatherandinsufficientland/fieldaccess

3. Combining both assessments, the quality of seed by farmers was assessed as good or

average(anoverwhelming97percentofresponses,fromasampleof1583).Thoughitwasastressedperiod, farmers ratedcropsyieldsasgood (34percentofcases)oraverage (27percent of cases). Therefore, over 61percent of farmer crop cases received acceptableyieldsfortheircrops-evenfollowingthedrought.Thereportedcasesofpooryieldsvariedby cropand region. For instance, a relativelyhighproportionof farmers ratedwheat andbarley production as poor that season, while most respondents in the Somali Regionreported poor belg harvests in general. Interestingly, there was no association betweenpooryieldsandseedsources:farmersreportedsimilarproportionsofpooryieldfromseedsourcedfroma)seedaidb)homestocksorc)localmarkets.

6

Note:Lastminuteweatherpatternsandaccesstoland/fields(whichcanfluctuatebetweenseasons)seemedtoaffecttheamountofseedfarmerssowed,toamarkeddegree.Theseseemedtodrivefarmers’sowingamounts,ratherthanwhetherseedwasonhand.Suchdynamismwillbeimportantforunderstandingfarmers’seeddemand.Meher Season 2016 Farmers’pointofview:demandandseeduseissuesThemeherseason2016wasprojectedtobeamorestressedperiodthanthe2016belgseason,andthemajorityofseedaidwasdeliveredforthistime.Thequantitativefindingsaroundseedsecurity were similar for this season as for the belg season (points 1-3 above), although areexplored ingreaterdepth formeher, asbothSSSAsoccurredduring this season,with farmingdecisionsandfieldassessmentshappeninginrealtime.5 CombiningbothSSSAs,farmersontheirownsourcedtwo-thirds(67percent)ofalltheseed

theyplanted(focusingoneachfarmer’smostimportantcrops).Home-savedseedprovided41percentofallseedplanted,withlocalmarkets24percent. It is importanttonotethatthesetwochannelsprovidedalargerpercentageofseedthanseedaidprovided(21percentfromcombinedgovernment,FAOandNGOaid).

In stress periods, local markets tend to be the source for purchasing seen to overcome seedinsecurityandforobtainingtherangeofcrops/varietiestobolsterresilience.Thiscouldpartlybeattributedtoeffortsbylocalmarkettradersinsourcingseedfromlessaffectedfarmersand/orregions.Localmarketsupportmightthereforemeritgreaterattention. 6 SSSA-CRSagaincomparedactualquantitiessowedwithnormalamounts(fromatotalof905

specificcropcomparisons):Meher2016sowingshovereddirectlyaroundnormal(anoverallaveragedecreaseofjust1.30percent).Foragivencrop,farmerssowedlessthannormalin42 percent of all cases. Among this group, themain reasons they sowed less seed thannormalwerethesameasforbelg2016:poorweather(mainreason26percentofthetime);financial constraints (23 percent); land constraints (16 percent); and improved sowingtechniques(15percent).Barelyany(1.3percent)farmerscitedalackofavailableseedasareason for sowing less. Again, the SSSA-FAO findings support the conclusion that seedavailability was not amajor constraint: when asked, crop-by-crop, if therewas “enoughseed”, only 17percent (of 2 372 responses) said “no”. As before, ‘seednot available’ islikelytobelessofanissuethanimpliedbythisfigure,asfarmershavemultiplechannelsforobtainingseed,and“notenough”couldreflectaccessratherthanavailabilityconstraints.

7 CombiningbothSSSAs, farmersagain judgednearlyall their seedas ‘good’or ‘average’ intermsof cleanliness andgermination (98percentof 4 305 farmer assessmentsof specificcropsandseedsources).Yieldobtainedortobeobtainedfromthemeher2016seasonwasdeemed promising or average for 74 percent of cases with seed from community-basedgroups and government/FAO/NGO aid receiving particularly high scores. Farmers raisedspecificqualityconcernsinafewcases,suchassomecasesofmixedwheatorteffvarieties,or regarding brokenmaize seed supplied from local government or cooperative sources.However,thesewerelocalizedcases.

7

Note:Theneedforfarmerstohaveflexibilityinwhattheysow,particularlyduringstressperiods.Farmersmayaltercropsandvarietiesusedaccordingtotheimmediateweatherpatterns,fieldsavailable,orprevailingmarketseedprices.Factoringinfarmerchoiceandabilitytostrategizecouldimprovetheresultsofaidresponses.Accesstoseed–purchase8. Both SSSAs showed that farmers bought much of their seed in 2016. In belg 2016,

48percent of all seed planted was purchased; inmeher 2016, 58 percent of seed waspurchased. Nearly all purchases were with cash, with less than two percent of seedaccessedviacredit.Seedpurchasewascommonacrossallsitesandcrops,andwasthemostimportantsinglemeansofaccess,aboveself-supply fromone’sownstocks,andfaraboveexchanges with other farmers or free gifts from seed aid. That farmers, even after astressed year, continued to get seed via cash purchase is striking. It highlights farmers’willingness to invest in this important input, shows thatmarkets are supplying significantvolumes to farmers, and suggests future possibilities for market-led approaches tostrengthenfarmers’longer-termseedsecurity.

9. SSSA-CRS calculated seed purchase expenses for an average farmer duringmeher 2016,basedonthreemajorcropspersite,averageamountsofseedactuallypurchasedforthosecrops, and current prices in local markets. These figures are indicative only, as actualamounts for each householdwill depend on land area and cropmixture, but they give asenseofthescaleofspending.InSire-andDodotaWoredas(ArsiZone,Oromiya,intheRiftValley) theaveragespendwas2550Birr forseedof teff,wheatandbarley (equivalenttoapproximately USD 116). This reflected large farm sizes (nearly half the farmers theresowed over 2 ha), and the predominance ofwheat and barley,which farmers across thedrierpartsofEthiopiatendtosowathighseedrates(150kgperhectareormorewasnotuncommoninseveralsites).Incontrast,DessieZuriaandTehuledereWoredas(SouthWolloZone,Amhara),averagespendingtobuyteff,wheatandfieldpeawas170Birr(equivalentto approximately USD 8), in part due to smaller land sizes in this highland area (three-quarters of households there farmed less than1 ha). Theother siteswere intermediate:average spending inOfla andAlamata (Tigray)was 416Birr (approximatelyUS19) and inWolaytaandKambata(SNNPR)itwas290Birr(roughlyUSD13).

While most farmers buy some seed, these figures show how that the amounts spent varyconsiderablybylocationandthatanycalculationsforseedamountand/orcash/vouchersupportwould also need to be tailored.Ofamore general note is that costs could be a challenge forsome,particularlythosepurchasinglargequantities. Inaddition,fertilizerhastobefactoredintothetallyofinputcosts.Onthesupplyside-seed/graintraders:Canthemarketsdeliver?Wherearefarmersbuyingseed?Agro-dealernetworksareonlyjustemerginginEthiopia,witha few shops in larger towns andhigher-potential areas (e.g. Adama) selling hybridmaize andvegetableseed. Farmers ina fewareas (e.g.WestGojjam)canalsochooseamongarangeofmaizevarietiesinpurchasingseed,assomemaizeseedproducershaveagreementswithagents

8

or cooperatives and market directly to farmers. However, this currently only happens withmaize inhigh-potentialareas. Elsewhere,farmerscanbuyseedviatheBureausofAgricultureandNaturalResourcesaspartofaninputpackage,oftenviacredit.Packageprogramstendtofocusonasmallnumberofcrops-nearly80percentofformalseedproduction5byvolumein2016 - 2017 is for wheat and maize – and centralized management of seed production andsupplymeanfarmersmaynothavemuchscopetochooseamongvarieties.Formostcrops,farmersbuymuchmoreoftheirseedfromseed/grainmarketsthananywhereelse.Inpartthisreflectschoice:awiderangeofcropsandvarietiesareonoffer.Butaccessisalso an important factor, as these markets offer lower prices or terms of trade than formaloutlets.Hence,theSSSAsfocusedattentionontheseseed/grainmarkets,wherefarmersscoutoutgrainthatissuitableforplanting.Forsowingmaterial,farmersseekadaptedvarietiesandlookforgrainofgoodquality(matureandnotbroken,withpebbles,dustandtwigssorted,andnopestdamage).Notallgraininsuchmarketsissuitableforsowing;thatwhichcouldbesowedisreferredtoas‘potentialseed’.Thesemarketstakeavarietyofforms,fromopen-airtraderstothosewithpermanenttradingandstoragepremises,andoperateatdifferentscales.Localseed/grainmarketswerenotedasparticularlyimportantforthelegumes(commonbeans,chickpea, cowpea, faba bean, and field pea), for some cereals (especially barley, teff andsorghum),andforlocalvegetables.Inthe2016meherseason,localmarketsprovidedbetween25to50percentoftheseedsownforthecropscitedabove.10. Themajorityofseed/graintraders(totaling116separatesupplyassessments)assessed

meher2016suppliesas‘normal’or‘moreabundantthannormal’,acrossarangeofcrops(wheat,barley,teff,commonbeans,sorghumandmaize).

11. Noteverygraintradersellspotentialseed,butthosethatdovaryfromsmall-scaleopenmarketsellerstolarger-scaletraderswithstorageandtransportfacilities.Thesevendorsmaytakestepstoenhancequality(e.g.sourcingpotentialseedfromreputedgrowersorregions),andsellpotentialseedatapremium(10to18percent)abovegrainprices.

12. Linkedtotheabove,‘potentialseedflows’fromoneregiontoanotherprovedextensiveandshortagesinanyoneareawereoffsetbyincomingsuppliesfromanother.Forinstance,atraderinKorem(TigrayRegion)wasdirectlycommissionedbythegovernmenttosourcebarley(‘potentialseed’)fromthesurroundingareasofKombolchaandCh’erch’erandchickpeafromasfaroffasAdama.Forthemeher2016alone,thetraderinquestionsold200MT(2000qt)totheKoremOfficeofAgricultureandRuralDevelopment.Specificadaptedvarietieswerepurchasedwithqualityscreeningmonitoredbygovernmentstaffduringthetransaction.

Any seed security zonal or woreda level plans might practically project for such for inter-zonal/woreda flows, acknowledging that supply dips in one region can be compensated byinflowsfromanother.Seedsecurityplansshouldalsorecognizethekeyimportanceof‘potentialseed’ traders. Such traders need to be identified and supported in their quest to gather orproduce quality seed. Activities might be targeted to this group to enable them rapidly andefficiently move needed and appropriate seed among areas. Again, they provide important

5FormalproductionismainlyviatheEthiopianSeedEnterprise,andRegionalSeedEnterprises.

9

amounts of seed and particularly of the legumes and minor cereals, which are poorlyrepresentedintheformalcommercialchannels.

13. Trader analysis of peak prices per quintal also showed relatively stability, comparing the

currentmeher2016withthemeher2015saleprices.Overallpricerisesweremeasuredat+18 percent,with variable changes by crops (some, like teff or sorghum, increasedmorethan20percent,otherssuchaswheatorfababeanwereactuallycheaperin2016thanin2015).

Allinall,seed/graintradersateachsiteviewedpotentialseedsuppliesasavailable,withpriceschangesnotunusuallyhigh,atleastaccordingtothetraders’pointofview.

Otherkeyissuesacrossseasonsrelatedtoseeduse:belg2016andmeher201615. BothSSSAsamplesshowimpressivelevelsofaccesstonewvarietiesofseed.Seventy-three

percentofhouseholdsreportedtheyhadgainedaccesstoanewvarietywithinthelastfiveyears. However 79percent of these new accessions have been of maize, wheat or teff.There has been negligible access to new varieties of the legumes, which are key fornutrition.

16. NewvarietieswerealsooverwhelminglyaccessedviagovernmentorFAO/NGOchannels(74

percent of cases),mostly coming in the form of aid, rather than through demand-drivencommercialoutletsthatmightservefarmersonamorecontinuingandsustainablebasis.

17. Across both SSSA samples, inorganic (chemical fertilizer) was employed by 74percent of

farmersforthemeher2016season.Usagerateswerelowerforthebelg(41percent,datafromSSSA-CRS).Inrainfall-stressedareas,farmersnotedthatitcanberiskytousefertilizeras it ‘burnsthesoil if there isa lackofrain’. Fertilizer ismostlyappliedonmaizeandteffduringthebelgseasonandtowheatandteffduringmeher.

18. Seedaid,thatisfreedistributionofseedaspartofemergencyresponseanddevelopmentinitiatives,hasbeenconductedona largescale,with55percentof thesamplepopulationhavingreceivedsuchaidwithinthelastfiveyears.Inlowerpotentialareas,theproportionwhich received aidwas higher. Aidwas received in the general population 1.6 times onaveragewithin the last fiveyears,withahighof seven times.Mostof theaid caseswereimplementedthroughdirectseeddistribution(84percent),withafewcitingseedloans(15percentofcases).

Aidmethodswhichallowfarmersachoiceandtheabilitytostrategize,suchascash,vouchersorseedfairswerevirtuallynon-existentforthefullsample.

10

Projections for Belg 2017 Projections for seed sourcing for the belg 2017 were also obtained, crop-by-crop. As theupcomingseasonwasseveralmonthsawayatthetimeoftheassessment,suchfiguresarebutspeculative.Majorpointsforthebelg2017projectionsare:

• Farmersexpecttorelyoninformalchannelsforthebulk(69percent)oftheirseedfortheirmajorcrops.Theyprojectownstockstosupply33percent,whilelocalmarketsareestimatedtoprovide36percentoftheirseedneeds.

• Farmers project overall sowing rates to rise sharply compared with normal rates: an

averageincreaseof28percent(showingalevelofoptimism).

• Farmers have already factored in important government/FAO/NGO assistance for16percentofseedforthenextseason,focusedespeciallyonmaize,wheatandcommonbean. This may represent farmers’ preference or hope (seed aid is a way to receivecertified seed, and free of charge). However, itmay also signal dependency in somecases,especially in locationswhereseedaidhasbecomeroutine.For theothercrops,theyarecountingongeneralself-sufficiency.

Longer-term Issues Both SSSAs reviewedbreeding, seed production and seedmarketing systems in Ethiopia, andthefullreportsprovidemuchdetailoftheseaspects.Thisbriefsummaryhighlightsafewsalientpoints.19. Ethiopia’s National Agricultural Research System has released a large number of crop

varieties. As SSSA-FAO highlights, since 1991 there have been 669 varieties released inEthiopia (530ofthemsince2001), including:322cerealvarieties (halfof thosewheatandmaize); 196 pulse varieties (with haricot bean, field pea and faba bean predominant), 81varieties of root crop (led by Irish and sweet potatoes), and 71 varieties of oilseed crops(mainlygroundnut,sunflower,linseed,andsesame).

20. Seedproductioniscarriedoutbybothpublicandprivateseedenterprises,withlimitedseedproduction from select Farmers’ Cooperatives and individual farmers. Aggregate nationalproduction isestimatedtobe130000tonnes,90percentof this forcereals. Othercropsreceivefarlessattentionfromseedmultiplicationefforts.Equally,notallreleasedvarietiesaremultiplied.TheOromiyaSeedEnterprise,forexample,producesarangeofwheat(23)andmaize(8)varieties,butonlyonetotwovarietiesformostothercrops.Whilefocusingona limitedproduct range isunderstandable for theseenterprises,atasystems level thismeansthatmanypotentiallyusefulcropvarietiesarenotabletoreachfarmers.

21. Certified seed, produced by from enterprises or cooperatives, is typically supplied viaWoredaAgriculturalDevelopmentOffices.Issuesnotedhereincludeshortageofsupplyofsome crop seed (e.g. barley and a number of pulses) or late delivery to some locations.About12percentofseedproductioniscarriedoverasunsoldstockintheformalsector;this

11

highlights the challenges faced in estimating andmeeting seeddemand in a complex andcentrallyorganizedsystem.

22. Both SSSAs documented the need for innovative ways of seed production and delivery -muchclosertozoneswherefarmersplant,andofferingfarmersthelargearrayofcropsandvarieties they require—not just for production, but also to meet their resilience andnutritionneeds.Somedecentralizedseedproductionisstartingtofillthisgap,suchasfromFarmerCooperativeswhichmultiplyseed.Fortheseeffortstobesustainable,andprovidefarmerswith cropsandvarieties choices, theyneed tobedemand-drivenwith farmersasthe clients (rather than institutional buyers, such NGOs or Bureaus of Agriculture andNaturalResources).

SALIENT POINTS AND REFLECTIONS: Belg 2016, Meher 2016, Belg 2017

1. FarmersdrawuponmultipleseedsourcesinEthiopia.Duringthebelg2016andmeher2016 seasons, farmers sourced seed for theirmajor crops largely from informal seedsources,withafocusontheirownstocksandlocalmarketchannels(77percentofseedsourcedfrominformalchannelsforthebelg;and67percentinformalforthemeher.)

2. Overall, thequantities of seed farmers sowed in 2016 showed little change from theamountstheynormallysowed.

3. Emergency seed aid during these belg 2016 andmeher seasons provided around 20percentof theseedsowed for theirmajorcrops ineachseason. It isnotpossible tocalculate the absolute degree to which this aid was crucial: farmers sometimespreferredtosowthenewvarietiesandcertifiedseedevenwhentheyhadseedintheirownhome-savedstocks.Certainly,theaidhadsomepositiveeffect,bysecuringaccesstocertifiedseedofnewvarieties,orbyallowingfarmerstousetheirmoneyforotherkeypurchases.

4. Focusing on the potentially vulnerable, that is, those sowing less of a given crop ineither thebelgormeher2016, threemain reasonsweregiven for the reduction. Nomoney tobuy seed, insufficientaccess to land/fields for the season, and simplypoorweather,whichkeptfarmersfromwantingtosowthefullamountofseedforthecrop.Lackofseedavailability(inmarkets,shops,orwithneighbors)featureinsignificantlyasarationaleforsowingless.Reasonsforreductionsweresimilaracrossallfourregions,withlackofoxenadditionallybeinghighlightedinSNNPR.

Note:Whileproviding free seedmighthelpwith financial constraints, suchdirect seedaidwouldnothavesolved the twodrivingproblems fordeclinedseeduse -poorweatherandinsufficientland/fieldaccess.

Intermsofsowinglessduetofinancialconstraints(whichwillbekeyforcalculationofcash/voucherneeds), this affected6.3percent of the total farminghouseholds duringthebelg2016and10percentofthetotalpopulationduringthemeher2016.

12

5. Seedgrain traders, thosecrucial for seed security in stressperiods,assessed suppliesforthemeher2016asnormalorabovenormalandconfirmedthatnormaltradingandsupply routeswere open. They calculated that peak prices for potential seed overallrose18percentfrommeher2015tomeher2016(althoughperhapsnotastrikingrise).Several cases were noted where seed/grain traders also became part of thegovernmentseedaidprocurementprocess.

6. For most legume and cereal crops (with the exception of wheat and maize), localmarkets provided 30 to 50percent of the seed farmers sowed during these twoseasons. Seed security plans might recognize the key importance of potential seedtraders.Suchtraderscouldusefullybeidentifiedandsupportedintheirquesttogather,transportoridentifyqualityseed.

7. Farmerspurchased48percentofallseedsowedinbelg2016,and58percentinmeher2016with the vastmajority of itwith cash. Varied sourceswere used, though localmarkets dominated. Purchase amounts and overall costs varied greatly by region,accordingto landsizesanddifferentcropprofiles.As farmersarealready investing inbuying seed, this suggest opportunities for using market-led approaches to offer awider choice and good quality seed to farmers – though business models need tounderstand,cultivate,andserveactuallocaldemand.Affordabilitymaybeanissueforsomefarmers.Understandingoflocalcontextandfarmingstrategieswillbeimportantforunderstandingfarmers’overallseedpurchasestrategies.

8. Farmers are obtaining impressive access to new varieties, but mainly through freedistributionswithafocusoncereals.Legumes,whicharekeyfornutrition,receivelessattention in emergency aid. Furthermore, good quality seed remains difficult forfarmerstoaccessonanon-goingbasis.

9. Emergencyseedaidiscommon,withoverhalfofallfarmersreceivingaid,onaverage1.6 times in the last fiveyears.Also,direct seeddistribution is thedominant formofthisaid. Thisapproachgivesfarmers littleornoabilitytostrategize instressperiods.Farmers routinelyalterwhat they sowaccording to the immediateweatherpatterns,fieldsavailableorprevailingmarketprices.Suchflexibilitymightbefactoredintofuturesupporttoincreaseaideffectiveness.

10. BothSSSAsinvestigatedfarmers’projectionsforseedsourcingforthebelg2017,cropby crop. As the upcoming season was several months away at the time of theseassessments,suchfiguresmustbeconsideredasspeculative. Farmersexpecttorelyoninformalchannelsforthebulkoftheirseedoftwomajorcrops(69percentofseedsown).Farmersprojectoverall sowingamounts tobehigher thannormal:anaverageincreaseof28percent.Thatsaid,farmersinthesamplealsoproject16percentoftheirseedwill come fromgovernment/FAO/NGOassistancenext season,whichmay signalthat theyhope toget aid seed (due to itsquality),or that theyexpect to (as aidhasbecomeroutine).

13

II. RECOMMENDATIONS: Short-term Below are key recommendationswhich are applicable across all sites. They emerge from ananalysisofthefieldevidenceandfocusonrecommendationsintheshort-term

1. Direct seed aid (distribution) for 2017 should be limited. There is little evidence ofseed unavailability in home stocks and markets, and farmers do not cite seedunavailabilityasareasonforplantingless.

1.1 Tominimizerisk,anydirectseeddistributionshouldfocusoncropsandvarietiesalreadyknownandusedbyfarmersinagivenregion;

1.2 Directseeddistributioninemergenciesshouldcriticallyreflectthesuitabilityofsupplying technology obliging repeated re-purchase (such as hybrid maize) issuitableforthepoor,seedinsecurerecipients,orforrisk-pronelocations.

2. Vulnerable farmers should be provided means to access seed in belg 2017 (cash,vouchers,possibly through fairs).Themajorseed-relatedreasonforfarmers’plantinglesshadtodowithmoney.Thiswastrueforallsitesandbothbelgandmeherseasons.

2.1 Theamountof any cash/voucher transfer shouldbest to tailoredby regionasseedcostsvarydramaticallyaccordingtolandsizeandcropprofile.

2.2 As vouchers, cash and fairs also aim to allow farmers to strategize duringperiodsof stress, specificefforts shouldbemade toensureawide rangeof cropsareonoffer(alsolegumesandminorcereals).

3. Vulnerablefarmersshouldalsobegivenmeanstoaccess/alleviateotherproductiveconstraints.

Vouchersforoxen,orforfarmlabor,couldbeexplored.Someanalysisofvouchersforfieldrentalwouldbeworthyofconsideration.

4. Support for local markets during emergencies and period of stress should be

considered. Localmarketsprovided30to50percentoftheseedsownforalllegumesand keyminor cereals. Select seed/grain traders are also already serving to provideemergencyseedstocksinkeyregions.

o Seedsecuritytradersmightbeusefullyidentifiedineachregion;

o Seedsecuritytradersmightreceivesupporttoensuringaqualityproduct;§ Trainingonseedsourcingandselection§ Possiblecreditforbetterstorage,orincentivesforaccessingstorageinputs,

suchaspallets.

14

RECOMMENDATIONS: Medium-Term More broad-based thinking is required regarding methods to improve the seed security ofsmallholderfarmersinEthiopia.Governmentdevelopmentefforts,aswellasrepeatedaid,arecurrently driving production and delivery mechanisms. This has had significant impact onincreasing the supply of certified seed ofmajor cereal crops. Expanding farmers’ options tochoose from among awide diversity of crops and varieties, and acquire good quality seed inways theycanaffordwill strengthenseedsecurity. Asmodestareas forwideraction, severalsuggestionsarelistedbelow:

5. Varietychoiceshouldbeenhanced.Governmentandpartnersshouldprovideat leasttwovarietieswithintheconventionalseeddistributionchannels inanysingledelivery,thus allowing farmers greater choice. In the medium term, seed enterprises andproducers,incollaborationwithextensionoffices,shouldsupportparticipatoryfarmer’sdemonstrations of new varieties at kebele levels for at least one to two seasons tocreate effective demand and avoid negative effect associated with “unknown new”varieties.Furthermore, there isaneed to increaseefforts indecentralized/direct seedmarketing,bothscalingupexistingefforts,aswellasexpandingeffortstonewregionsandtocropsbeyondmaize.Inthelongterm,moreparticipatoryplantbreedingeffortsincropssuchascowpea,mungbeansandgrasspeaareneeded.

6. Criticallyreviewcapacityinseedqualitymanagement.Regionalseedlaboratoriesneedsufficientequipmentandhumancapacitiestobeabletocarryouttheirroles. Iftheselaboratoriesaretosupport,andeventuallycertify,moredecentralizedseedproductiongroups,theywillneedtobesufficientlyresourced.

7. Support a diversified seed sector to increase the availability and diversity of goodqualityseedforfarmers.Thiscouldbesupportedinanumberofways.Public/privatepartnershipswillbeimportant,aswilladequatesupplyofbasicseedfromawiderangeof crops and varieties to all seed sector actors. Improved credit services or othertargetedsupportcouldhelpattractmoreprivateparticipantsintheseedsector.

8. Decentralizedseedproductionneedstobecomeamorestrategicandeffectiveforceinserving farmersas the formalseedsectorwillneverbeable tohandlea) therangeofcropsneededforstresszones;norb)therangeofvarieties.Effortsneedtobemadetosupport cooperatives and organized groups to multiply crops which attract littleattentionfromtheseedenterprises,suchasharicotbeans,fababeans,chickpea,fieldpea as well as some cereals including sorghum and barley. Decentralized seedproduction and deliverywill prove particularly important for the legumes and for thevegetatively propagated crops. At this point, the decentralized seed multiplicationinitiativesseemtobehavingmodestgains.Thosevisitedhadlimitedcropportfoliosandtheir expansion was hampered by the full-fledged requirement of Certificate ofCompetency (CoC). As a general recommendation, sustainable decentralized seedproductionmodelsneedtobeconfirmedinmanyregionsofEthiopia.

9. Deliverymechanismsforgivingallfarmersregularaccesstonewvarietiesneedtobeintensified. Diversifying sale outlets beyond Farmer Cooperatives, unions and agro-dealerscouldhelpexpandfarmers’choicetoawidersetofcropsandvarieties.Saleof

15

diverseseedthroughabroaderrangeofoutlets,suchasviaruralshopsonconsignment,may help expand access. The experience of some seed producers whomarket theirseeddirectly to farmers, via agents (as inWestGojjam) shouldbe studied for lessonsandpossiblereplicationelsewhere.Saleinsmallerpacksizes(1kg,2kgand5kg)mayalsoprovideopportunitiesforpoorerfarmerstoaccessnewvarietiesandqualityseed.

10. Given that localmarkets (and their traders) are important for farmers’ seed supply,more attention shouldbe given to encouraging that theseopen seed/grainmarketssupplythekindsofpotentialseedfarmerswantandneedonamoreconsistentbasis,andnot just inemergency.Asonepointofdeparture,seedandgraintraderscouldbepowerfulpartners inhelping tomovenewmodernvarietieswidely,withinandamongfarming communities. Traders could also receive basic information on grain and seedqualityaspectsandmanagementthroughtrainings,leafletsand/orradiomessages.

11. Economicdevelopmenteffortsshouldbelinked.Strengtheninglocalfinancialservicesand, particularly around credit provision, will be useful, especially in Oromiya, Afar,GambellaandSomaliRegions.Linkedtothis,valueadditionshouldbesupportedwherepossible, as better market opportunities enhance farmers’ purchasing power, andunderpinsustainedseeddemand.

12. Research for development (R4D): Seed issues are interrelated with productionchallenges, and therefore addressing some of the crop production challenges willdirectly or indirectly affect seed security. Some of the challenges that need dueattentionandactioninclude;

• Developmentofstrigaresistant/tolerantvarietyofseed.Thiswillminimizetheimpactoftheweedonproductivityandsubsequentlyone’sownsavedseed.

• IntegratedPestManagement.Forcontrolofmajorpests,thereisaneedtopromoteuseofintegratedpestmanagement(IPM)practices,whichmayincluderegularmonitoringandscoutingoffieldpests,useofappropriateculturalandproductionpracticesaswellappropriateapplicationselectivepesticides.

13. Finally, the focus on quality seed for increased production could be broadened to

includethegoalsof ‘enhancedresilience’and ‘enhancednutrition’.Aprimefocusoncereals alone (the currentde facto strategy)maynot be sufficient to help strengthenfarmingsystemsinthesetimesofrepeatedclimatestressandfoodinsecurity.

16

Withfinancialsupportby