The case of cassava in Africa, A

18 Pages·2017·1.08 MB·English

RESEARCHARTICLE
Identifying crop research priorities based on
potential economic and poverty reduction
impacts: The case of cassava in Africa, Asia,
and Latin America
AregaD.Alene1*,TahirouAbdoulaye2,JosephRusike3,RicardoLabarta4,
BernardoCreamer5,MarthadelR´ıo4,HernanCeballos4,LuisAugustoBecerra4
1 InternationalInstituteofTropicalAgriculture(IITA),Lilongwe,Malawi,2 InternationalInstituteofTropical
Agriculture(IITA),Ibadan,Nigeria,3 AllianceforaGreenRevolutioninAfrica(AGRA),Nairobi,Kenya,
a1111111111 4 InternationalCentreforTropicalAgriculture(CIAT),Cali,Colombia,5 UniversidaddeLasAmericas
a1111111111 (UDLA),Quito,Ecuador
a1111111111
a1111111111 *[email protected]
a1111111111
Abstract
Itiswidelyrecognizedthatincreasingagriculturalproductiontothelevelsneededtofeedan
OPENACCESS expandingworldpopulationrequiressharplyincreasedpublicinvestmentinresearchand
Citation:AleneAD,AbdoulayeT,RusikeJ,Labarta developmentandwidespreadadoptionofnewtechnologies,butfundingfornationaland
R,CreamerB,delR´ıoM,etal.(2018)Identifying internationalagriculturalresearchhasratherdeclinedinrecentyears.Inthissituation,prior-
cropresearchprioritiesbasedonpotential
itysettinghasbecomeincreasinglyimportantforallocatingscarceresearchresources
economicandpovertyreductionimpacts:Thecase
amongcompetingneedstoachievegreaterimpacts.Usingpartialequilibriumeconomicsur-
ofcassavainAfrica,Asia,andLatinAmerica.PLoS
ONE13(8):e0201803.https://doi.org/10.1371/ plusmodelsandpovertyimpactsimulations,thispaperassessescassavaresearchpriori-
journal.pone.0201803 tiesinAfrica,LatinAmericaandCaribbean,andAsiabasedonthepotentialeconomicand
Editor:PaulC.Struik,WageningenUniversity, povertyreductionimpactsofalternativeresearchandtechnologyoptions.Theresults
NETHERLANDS showedthatefficientplantingmaterialproductionanddistributionsystemsandsustainable
Received:November29,2017 cropandsoilfertilitymanagementpracticeshavethegreatestexpectedeconomicandpov-
ertyreductionimpactsinthethreeregions.Lackofcleanplantingmaterialsisamajorcon-
Accepted:July23,2018
strainttoadoptionanditisenvisagedthatefficientproductionanddistributionsystemsfor
Published:August8,2018
plantingmaterialcanacceleratetechnologyadoptionbyfarmers.Similarly,sustainablecrop
Copyright:©2018Aleneetal.Thisisanopen
andsoilfertilitymanagementpracticesplayakeyroleinclosingtheobservedyieldgaps,
accessarticledistributedunderthetermsofthe
especiallyinAfrica.Thepaperdiscussestheresultsofthepriorityassessmentforkeycas-
CreativeCommonsAttributionLicense,which
permitsunrestricteduse,distribution,and savaresearchoptionsandconcludeswiththeimplicationsforcassavaresearchpriorities.
reproductioninanymedium,providedtheoriginal
authorandsourcearecredited.
DataAvailabilityStatement:Mostoftherelevant
dataarewithinthepaperandtherestwillbe
Introduction
includedintheSupportingInformationfiles.
Cassavaisthethirdmostimportantfoodcropinthetropicsafterriceandmaizeandisthesec-
Funding:ThisworkwassupportedbyCGIAR
ResearchProgramonRoots,Tubers,andBananas ondmostimportantfoodstapleinAfricaaftermaizeaccountingformorethanhalfofthedie-
(RTB)http://www.rtb.cgiar.org/. tarycalorierequirementsofover200millionpeople[1].HalfabillionpeopleinAfricaeat
cassavaeveryday,andthishigh-starchrootisalsoanimportantstapleinLatinAmerica
Competinginterests:Theauthorshavedeclared
thatnocompetinginterestsexist. andtheCaribbean.InAsia,cassavaservesasasourceoffoodandlivestockfeedwhilealso
PLOSONE|https://doi.org/10.1371/journal.pone.0201803 August8,2018 1/18
CassavaresearchprioritiesinAfrica,Asia,andLatinAmerica
providingrawmaterialforthemanufacturingofpharmaceuticals,industrialstarch,biofuels,
andotherproducts[2].Assuch,cassavaisimportantnotonlyforruralhouseholdsbutfor
nationaleconomies.Despitemajorbioticandabioticthreatstocassavaproductionandpro-
ductivity,cassavaproductionhasexpandedespeciallyinAfricaandthisislargelyattributedto
nationalandinternationalcassavaimprovementresearchefforts[1].
Internationalcassavaimprovementresearchwasinitiatedintheearly1970sattheInterna-
tionalInstituteofTropicalAgriculture(IITA)andtheInternationalCenterforTropicalAgri-
culture(CIAT)withafocusondevelopinghigh-yieldingvarietieswithresistancetomajor
pestsanddiseases[3,4].Inadditiontobreedingforhighyieldandresistancetomajorpests
anddiseases,cassavaresearchinvolveddevelopingbiologicalcontrolandintegratedpestman-
agementoptionstoreducelossesduetoinsectpests.InSub-SaharanAfrica(SSA),thework
resultedinanumberofseveralelitegenotypesthathadresistancetocassavamosaicdisease
(CMD)andcassavabacterialblight(CBB)aswellashighandstableyieldsandgoodconsumer
acceptability.Thedevelopmentofimprovedvarietiesandtheirdeliverytonationalprograms
fortestingunderspecificlocalconditionsduringthelate1970sand1980shasledtothesuc-
cessfulreleaseofhighyieldinganddiseaseresistantvarietiesforadoptionbyfarmers.Thenew
varietiescombineenhancedCMDtolerancewithpreferredpostharvestcharacteristics,wider
agroecologicaladaptation,and50–100%higheryieldsevenwithouttheuseoffertilizer[1,3].
Despitemajorresearchsuccessesinthepast,farmlevelcassavayieldsremainlowespecially
inAfricaduetoanumberofemergingthreatssuchaspestsanddiseases.Realizationofhigher
potentialyieldsinfarmers’fieldsrequirescontinuedinvestmentingeneticimprovementand
betteragronomyaswellaspestanddiseasemanagement.Tohelpcounterthethreatofpests
anddiseases,scientistsshouldidentifyandusebiotechnologytoolstodevelopmolecularmark-
ersfortraitssuchaswhiteflyresistance,quantitativetraitloci(QTLs)inpopulationsderived
fromheterozygousparentmaterials,andprotocolsforrapidmultiplicationofdisease-free
plantingmaterialsthroughtissueculture.
Itiswidelyrecognizedthatincreasingagriculturalproductiontothelevelsneededtofeed
anincreasingworldpopulationrequiressharplyincreasedpublicinvestmentsinresearchand
developmentandwidespreadadoptionofnewtechnologies,butfundingfornationaland
internationalagriculturalresearchhasratherdeclinedinrecentyears.Inthissituation,priority
settinghasbecomeincreasinglyimportantforallocatingscarceresearchresourcesamong
competingneedstoachievegreaterimpacts[5].Systematicpriorityassessmenthasbeencon-
ductedsincerecentlybycombiningscientists’viewsonthepotentialforaddressingparticular
constraintsthroughresearchandtechnologyoptionswithaneconomicassessmentoftheben-
efitsthatcouldarisefromadoptionofthosetechnologies[6–14].Followingitsofficiallaunch
in2012,theCGIARResearchProgramonRoots,TubersandBananas(RTB)embarkedona
strategicassessmentofresearchprioritiesforbanana,cassava,potato,sweetpotato,andyams.
Usingpartialequilibriumeconomicsurplusmodelsandpovertyimpactsimulations,this
paperassessestheexpectedeconomicandpovertyreductionimpactsofcassavaresearchand
technologyoptionswithaviewtoinformingstrategicprioritysettingofcassavaresearchin
Africa,LatinAmericaandCaribbean,andAsia.Whilealotofpastpriorityassessmentwork
focusedonstrategiccommoditypriorities,thisstudyundertakescrop-specifictechnologypri-
orityassessment.Thiskindofprioritysettingisbecomingincreasinglyimportantforanum-
berofCGIARResearchPrograms(CRPs)supportingasetofprioritycommoditiesthatneed
tofocusonhigh-impactlinesofresearch.Thepaperpresentsanddiscussestheprocedures
andresultsofthepriorityassessmentforkeycassavaresearchoptionsanddiscussestheimpli-
cationsforcassavaresearchpriorities.
Therestofthepaperisorganizedasfollows.Thenextsectionprovidesanoverviewofthe
methodologyused,whereassection3providesdetailsofthedatasources.Section4presents
PLOSONE|https://doi.org/10.1371/journal.pone.0201803 August8,2018 2/18
CassavaresearchprioritiesinAfrica,Asia,andLatinAmerica
anddiscussestheex-anteimpactassessmentresultsandthelastsectiondrawsconclusionsand
implications.
Methods
Economicsurplusmodelandcost-benefitanalysis
Severalimpactstudiesofagriculturaltechnologieshaveestimatedaggregateeconomicbenefits
throughextrapolationoffarm-levelyieldorincomegainsusingpartialequilibriumsimulation
modelssuchastheeconomicsurplusmodel[5].Theeconomicsurplusmethodisthemost
widelyusedprocedureforeconomicevaluationofbenefitsandcostsofatechnologicalchange.
Technologicalchangeduetoresearchinagricultureincreasestheyield,reducesyieldlosses,or
reducesthecostofproduction[5].Ifthenewtechnologyisyieldincreasing,theproducersells
moreofthegoodinthemarketandifdemandisdownward-slopingthepricedecreasesas
well.Technologyadoptionreducestheper-unitcostofproductionandhenceshiftsthesupply
functionofthecommoditydownandtotheright.Ifthemarketforthecommodityisperfectly
competitive,thiswillleadtoanincreaseinthequantityexchanged(Q toQ )andafallin
0 1
pricefromP toP (Fig1).Asaresult,consumersbenefitfromthepricereductionandpro-
0 1
ducersmaybenefitfromsellingmoreoftheproduct[5].
Theeconomicsurplusmodelwasthereforeusedtoderivesummarymeasuresofthepoten-
tialimpactsofcassavaresearchoptionsforaperiodof25yearsstartingfrom2014.Thebene-
fitsweremeasuredbasedonaparalleldownwardshiftinthe(linear)supplycurve.Theannual
flowsofgrosseconomicbenefitsfromcassavatechnologieswereestimatedforeachofthe
countriesandaggregated,withtheaggregatebenefitsandcostsfinallydiscountedtoderivethe
presentvalue(in2014)oftotalnetbenefitsfromtheinterventions.Thekeyparametersthat
determinethemagnitudeoftheeconomicbenefitsare:(1)theexpectedtechnologyadoption
intermsofareaunderimprovedtechnologies;(2)expectedyieldgains(oravoidedlosses)fol-
lowingadoption;and(3)pre-researchlevelsofproductionandprices.Giventhelimitedinter-
nationaltradeoptionsforcassavainmostoftheproducingcountries,theeconomicsurplus
modelfortheclosedeconomyshowninFig1wasusedtocalculatetheeconomicbenefitsfor
Fig1.Effectsoftechnologicalchangeonproducerandconsumersurplus[5].
https://doi.org/10.1371/journal.pone.0201803.g001
PLOSONE|https://doi.org/10.1371/journal.pone.0201803 August8,2018 3/18
CassavaresearchprioritiesinAfrica,Asia,andLatinAmerica
eachcountryfromadownwardshiftinthesupplycurve.Thedemandforthecommodityis
denotedbyD,whereasthepre-researchsupplycurveisS andthepost-researchsupplycurve
0
followingtechnologicalchangeisS .Theinitialequilibriumisdenotedas(P ,Q ),whilethe
1 0 0
post-researchequilibriumis(P ,Q ).Thatis,theinitialequilibriumpriceandquantityareP
1 1 0
andQ ,whereasafterthesupplyshifttheyareP andQ .Thetotalbenefitfromtheresearch-
0 1 1
inducedsupplyshiftisequaltotheareabeneaththedemandcurveandbetweenthetwosupply
curves(ΔTS=areaabce).Thetotalbenefitcomprisesthesumofbenefitstoconsumers
(ΔCS=areaP bcP )andthebenefitstoproducersintheformofthechangesinproducersur-
0 1
plus(ΔPS=areaP ceminusareaP ba).Undertheassumptionofaparallelshift(sothatthe
1 0
verticaldifferencebetweenthetwocurvesisconstant)areaI deequalsareaP ba.
0 0
Inaclosedeconomy,economicsurplusmeasurescanbederivedusingformulaspresented
inAlstonetal.(1995):(1)ChangeineconomicSurplus(ΔES)=P Q K(1+0.5Zη);(2)Con-
0 0 t t
sumersurplus(ΔCS)=P Q Z(1+0.5Zη);andProducerSurplus(ΔPS)=(K−Z)P Q (1
0 0 t t t t 0 0
+0.5Zη),whereK isthesupplyshiftrepresentingtheproductofcostreductionpertonofout-
t
putasaproportionofproductprice(K)andtechnologyadoptionattimet(A);P represents
t 0
pre-researchpricefor2010─2012(US$/ton);Q ispre-researchlevelofproductionfor
0
2010─2012;ηisthepriceelasticityofdemand;andZ istherelativereductioninpriceattime
t
t,whichiscalculatedasZ =Kε/(ε+η),whereεisthepriceelasticityofsupply.Theresearch-
t t
inducedsupplyshiftparameter,K,isthesinglemostimportantparameterinfluencingtotal
economicsurplusresultsfromunitcostreductionsandwasderivedasK =[((ΔY/Y)/ε–(ΔC/
t
C))/(1+(ΔY/Y))]×A whereΔY/Yistheaverageproportionalyieldincreaseperhectare;εis
t
theelasticityofsupplythatisusedtoconvertthegrossproductioneffectofresearch-induced
yieldchangestoagrossunitproductioncosteffect,ΔC/Cistheaverageproportionalchange
inthevariablecostsperhectarerequiredtoachievetheyieldincrease,andA istherateof
t
adoptionoftheimprovedtechnologyattimet—theproportionoftotalcroppedareaunder
theimprovedvarietiesandpractices.Annualsupplyshiftswerethenprojectedbasedonpro-
jectedadoptionprofileforimprovedtechnologies(A)fortheperiodfrom2014to2039.Adop-
t
tion(A)isassumedtofollowthelogisticdiffusioncurve.
t
Foreachcountryi(i=1...N),thechangesineconomicsurplus(ΔES)andtheresearchand
extensioncosts(C)arediscountedatarealdiscountrate,r,of10%perannumtoderivethe
t
netpresentvalues(NPV)asfollows:
(cid:18) (cid:19) (cid:18) (cid:19)
X25 XN DES X25 C
NPV¼ i;t (cid:0) t
ð1þrÞt ð1þrÞt
t¼1 i¼1 t¼1
Theaggregateinternalrateofreturn(IRR)wasalsocalculatedasthediscountratethatequates
theaggregatenetpresentvalue(NPV)tozeroasfollows:
(cid:18) (cid:19) (cid:18) (cid:19)
X25 XN DES X25 C
i;t (cid:0) t ¼0
ð1þIRRÞt ð1þIRRÞt
t¼1 i¼1 t¼1
Estimationofpovertyimpacts
Extendingtheresultsoftheconventionaleconomicsurplusandcost-benefitanalysis,the
impactofeachofthecassavaresearchoptionsonruralpovertyreductionwasestimatedfol-
lowingAleneetal.[15].Itweighstheeconomicsurplusresultsaccordingtothepovertylevels
ineachofthecountries,theshareofagricultureintotalGDP,andtheagriculturalgrowthelas-
ticityofpoverty.Theimpactofeachresearchoptiononruralpovertyreductionwasestimated
byfirstestimatingthemarginalimpactonpovertyreductionofanincreaseinthevalueofagri-
culturalproductionusingpovertyreductionelasticitiesofagriculturalproductivitygrowth.
PLOSONE|https://doi.org/10.1371/journal.pone.0201803 August8,2018 4/18
CassavaresearchprioritiesinAfrica,Asia,andLatinAmerica
Thereductioninthetotalnumberofpoorwasthencalculatedbyconsideringtheestimated
economicbenefitsastheadditionalincreaseinagriculturalproductionvalue.Thirtleetal.[16]
foundthata1%growthinagriculturalproductivityreducesthetotalnumberofruralpoorby
0.72%inAfrica,0.48%inAsia,and0.15%inLatinAmericaandtheCaribbean(LAC).Under
theassumptionofconstantreturnstoscale,a1%growthintotalfactorproductivityleadstoa
1%growthinagriculturalproduction.Foreachcountry,thenumberofpoorliftedabovethe
$1-a-daypovertylinewasthusderivedasfollows:
(cid:16) (cid:17)
(cid:18) DES (cid:19) @ln NNp
DN ¼ (cid:2)100% (cid:2) (cid:2)N
p Agriculturevalueadded @lnðYÞ p
|fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} |fflfflfflffl{zfflfflfflffl}
GainsfromR&Eas%of agriculturalproduction Povertyelasticity
|fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}
Povertyreductionas%of thepoor
|fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}
Numberof poorescapingpoverty
whereΔN isthenumberofpoorliftedabovethepovertyline,N isthetotalnumberofpoor,
p p
Nisthetotalpopulation,Yisagriculturalproductivity,andΔESisthechangeineconomicsur-
plus.Thepovertyelasticityisinterpretedasthemarginalimpactofa1%increaseinagricul-
turalproductivityintermsofthenumberofpoorreducedasapercentageofthetotalpoor
(N ),andnotofthetotalpopulation.
p
Estimationofthenumberofpotentialbeneficiaries
Dataonaveragecropareaperhouseholdandaveragehouseholdsizewereusedtoestimatethe
numbersofbeneficiaries,followingaprocedureanddatasetdevelopedtoestimatetotalnum-
berofRTBpoorbeneficiaries[17].Dataforindividualcountrieswereobtainedmostlyfrom
FAOstatisticaldatabase,publishedsourcesofinformation,orexpertopinionwhenneeded.
Estimatedareaundertwoadoptionscenarios(highandlowadoption)wasdividedbytheaver-
ageareaperhouseholdtoestimatethenumberofadoptinghouseholds,andthenmultiplied
byhouseholdsizetoestimatetotalnumberofbeneficiaries.
Datasources
Constraintsanalysisandidentificationofresearchoptions
Expertsurveysandconsultationswereconductedbetween2011and2013toguidethecon-
straintsanalysisandtheidentificationandrankingofresearchoptions.Recognizingthe
importanceoffarmers’voiceinprioritysettingofagriculturalresearch,aliteraturereviewwas
firstundertakentotakestockofavailableevidenceandsecondarydataonproductionand
marketconstraints,technologypreferences,yieldgaps,andfarmlevelimpactsfrombaseline
andadoptionstudiesinvolvingfarmersaswellasfromon-farmfarmerparticipatoryresearch
work.Theoutcomeofthereviewservedasaguidenotonlyfordesigningthequestionnaires
usedfortheexpertsurveysbutalsoforfacilitatingtheconsultationsduringworkshopsthat
wereorganizedtoelicitandvalidateindividualexpertopinionsandestimatesaboutthemajor
constraints,yieldgaps,andtheprospectsofarangeofpromisingresearchandtechnology
options.Thesurveysengagedstakeholdersfromabroadrangeofdisciplinesandbackgrounds.
Thecassavaexpertcommunitywasinvolvedintheidentificationoftheproductionandmarket
constraintsandintheselectionofresearchandtechnologyoptionsthatcanaddresstheidenti-
fiedconstraints.Consultingabroadrangeofexpertswithdifferentfieldsofexpertiseenabled
ustocapturekeyconstraintsirrespectiveofinstitutionalprioritiesandcapacity.Overall,the
PLOSONE|https://doi.org/10.1371/journal.pone.0201803 August8,2018 5/18
CassavaresearchprioritiesinAfrica,Asia,andLatinAmerica
expertsurveysenabledtheidentificationofthemajorconstraintsandassociatedresearch
optionstobeincludedintheex-anteimpactassessmentinthesubsequentstepsofthepriority
assessmentexercise.
Theidentificationofcassavaresearchoptionsstartedwithanalysisofthedataobtained
fromtheglobalexpertsurveyinwhichasampleof343cassavaexpertsidentifiedthepriority
constraintstocassavaproduction,processing,andmarketing.Theopinionsofscientistswho
arecloselyinvolvedinresearchoncassavaproduction,processing,andmarketconstraints
servedasthemajorsourceofinformationforidentifyingresearchoptionstoaddressthose
constraints.Forthisobjective,aglobalsurveyinstrumentwasdesignedinconsultationwith
scientistsatCIATandIITAinSpanish,English,French,andPortuguese.Aglobalonlinesur-
veyofcassavaexpertswasconductedin2012usingtheonlineSurveyMonkeytooland60
questionnaireswerecompleted.Inaddition,questionnaireswereadministeredtocassava
expertswhoattendedinternationalevents.Atotalof282responseswereobtainedattheSec-
ondScientificConferenceoftheGlobalCassavaPartnershipforthe21stCentury,heldon18–
22June2012,inKampala,Uganda.Atthe16thTriennialSymposiumoftheInternational
SocietyforTropicalRootCropsheldon23–28September2012inAbeokuta,Nigeria,another
29questionnaireswerecompleted.Finally,cross-countrysurveysofthenationalcassavapro-
gramsandexpertconsultationswereconductedin2013inAfricaaswellasinLatinAmerica
andtheCaribbean(LAC)andAsia.Theresultsofthesurveybasedonthe343completedques-
tionnairesarepresentedinAleneetal.[18]
Potentialresearchoptionswereidentifiedbasedontheexpertsurveysandconsultationsfor
furtherformalevaluationusingtheeconomicsurplusmodel[5].Theseresearchoptions
includedthosethataddresstheconstraintsrelatingto:(1)rootyields;(2)productioncosts;(3)
postharvestprocessingandutilization;and(4)sustainableproduction.Theinitiallistof
researchoptionswaspresentedanddiscussedwiththescientistsfromIITAandCIAT,and
laterattheRTBpriorityassessmenttaskforceworkshopheldon12–16August2013,inLima,
Peru.TheseresearchoptionswerelaterlinkedwithCIATandIITAresearchoutputs.The
researchoptionswereselectedtomatchselectedresearchoptionsassociatedwithRTBflagship
projects,whichcontributetotherequiredattainmentofIntermediateDevelopmentOutcomes
(IDOs).Thefinalsetofresearchoptionswasthendevelopedandagreeduponatthefinal
workshopheldon12–14November2013inLima.Theseincluded:(1)High-yieldingvarieties
withresistancetomajordiseases(CMD/CBSD);(2)High-yieldingvarietieswithhighdrymat-
terandstarch;(3)High-yieldingvarietieswithlongershelflife;(4)High-yielding,drought-tol-
erantvarietiesandincreasedwater-useefficiency;(5)Sustainablecropandsoilfertility
managementpractices;(6)Integratedpestanddiseasemanagementpractices,includingresis-
tantvarieties;(7)Efficientandmassivehigh-qualityplantingmaterialproductionanddistribu-
tionsystems;(8)Processingtechnologiesforvalueaddition;(9)Strategiestoprevent
introductionofexoticpestsanddiseases;and(10)High-yieldingvarietiestoleranttocold
weatherandfrost.AdetaileddescriptionofthecassavaresearchoptionsisprovidedinAlene
etal.[18].
Expertestimatesofthevaluesofkeyparameters
Cassavaresearchandextensionexpertsservedasthemajorsourceofinformationfortheeco-
nomicsurplusanalysisofcassavaresearchoptions.Astructuredquestionnairewasdeveloped
toguideconsultationswithIITAandCIATscientistsaswellaswithNARSpartnersinAfrica,
LAC,andAsiawhoareworkingonparticularcassavaproductionandmarketconstraintsto
elicitkeyparameterestimatesfortheresearchoptionsaddressingthoseconstraints.Expert
consultationsatIITAinvolved12scientists:cassavabreeders(6),agronomists(3),virologists
PLOSONE|https://doi.org/10.1371/journal.pone.0201803 August8,2018 6/18
CassavaresearchprioritiesinAfrica,Asia,andLatinAmerica
(2),andprocessingandutilizationspecialists(1).Thecross-countrysurveyinAfricainvolved30
expertsfromNARSpartnersinAfrica:Benin(1),Cameroon(1),DRC(1),Ghana(4),Kenya(1),
Mozambique(3),Nigeria(2),Togo(3),Uganda(3),Tanzania(9),andZambia(2).InCIAT,a
groupof14scientists(breeders,agronomists,postharvestprocessingexperts,molecularbiolo-
gists,entomologists,plantphysiologists,andvirologists)workinginLACandAsiawascon-
sulted.Anonlinesurveywasalsodesignedandimplementedand46responseswereobtained.
Foreachresearchoptionidentified,scientistswereaskedtoestimatethevaluesofthefol-
lowingkeyparameters:maximumadoptionrate,yearofbeginningofadoption,yearstomaxi-
mumadoptionrate,expectedyieldincrease(%),areaaffectedbytheconstraint(%),cost
changeduetoinputs(%),andprobabilityofresearchsuccess(%).Thevaluesofsomeparame-
terssuchasresearchcostswereassembledfromseveralsources,suchasRTBprogrampro-
posalandpastempiricalwork[15,16]aswellasfromFAO(http://faostat.fao.org/)andthe
WorldBank(http://data.worldbank.org/indicator).Thelimitationofexpertopinionsurveys
relatestothedegreeofsubjectivitywiththeestimationofthevaluesofkeyparametersthat
determinethesizeoftheexpectedbenefits.Whileitistruethatmanyofthejudgementsmade
intheprocessaresubjective,theuseofamoretransparent,participatoryanditerativeap-
proachfacilitatesgreaterdialogueandconsensusbuildingtoensuresomelevelofobjectivity.
Table1presentsthedescriptionofthekeyproject,technology,andmarket-relatedparameters
usedandthecorrespondingdatasources.
Dataonsocioeconomicparameters
Table2presentsthedataonthekeysocioeconomicparametersusedintheeconomicsurplus
analysisofcassavaresearchoptionsforindividualcountriesinAfrica,Asia,andLAC.Dataon
annualharvestedarea,production,andproducerpriceswereobtainedfromtheFAOSTAT
database(http://faostat.fao.org/).Weusedthree-yearnationalaveragesforeachcountryfor
theperiod2010–2012.IncaseswhereFAOdatawerenotavailableforparticularcountriesand
years(e.g.,producerprices),weuseddataobtainedfromtherespectiveministriesofagricul-
tureandofficesofstatistics.Dataontheincidenceofpoverty,thenumberofpoor,andagricul-
turalvalueaddedwereobtainedfromtheWorldBank’sWorldDevelopmentIndicators
database(http://data.worldbank.org/indicator).
Wealsousedpovertyelasticitiesof0.72,0.48,and0.15forAfrica,Asia,andLAC,respec-
tively[16].Thedataoncassavaareaperhouseholdandhouseholdsizethatwereusedforthe
estimationofthenumbersofbeneficiariesweretakenfromadatasetputtogetherfortheesti-
mationofthepotentialnumberofbeneficiariesoftheRTBprogram[17].
Dataontechnologydevelopment,dissemination,andadoptionparameters
Theeconomicsurplusmodelemployedfortheex-anteimpactanalysistypicallyusesmarket-
relateddataonsocioeconomicparametersandtechnology-relateddataontechnologydevel-
opment,dissemination,andadoptionparameters[5].Therefore,inadditiontothesocioeco-
nomicparameterssuchasproductionandprices,theeconomicsurplusmodelusesanumber
ofparametersthatrelatetotheresearchanddisseminationprocessandincludesthosethat
relatetotheexpectedeffectsofnewtechnologyadoptiononyieldgains(orreducedyield
losses)andproductioncosts.Inadditiontoparametersrelatedtoexpectedyieldgainsandpro-
ductioncostchangesfollowingtechnologyadoptionbyfarmers,othertechnology-related
parametersofimportanceinclude(1)theresearchlagdefinedasthenumberofyearsittakes
untilanadoptableinnovationwillbeavailabletofarmers;(2)adoptionceilingdefinedasthe
maximumadoptionrateasaproportionoftotalcroppedarea;(3)adoptionlagdefinedasthe
numberofyearsuntilmaximumadoptionisreached;(4)thecostsrequiredtoconductR&D
PLOSONE|https://doi.org/10.1371/journal.pone.0201803 August8,2018 7/18
CassavaresearchprioritiesinAfrica,Asia,andLatinAmerica
Table1. Assumptionsanddatasourcesforkeyparametersusedintheeconomicsurplusanalysis.
Parameter Assumption/Source
Timeperiod 25years(2014–2039);10yearsforresearchinvestment—
researchlag(maximumtimeperiodforRTB).Mostofthe
R&Dinvestmentswillrunfor10years,thoughother
researchoptionsmayeitherbelongerorshorter.
Elasticitiesofsupplyanddemand Elasticitiesofsupplyanddemandwereassumedtobe1
and0.5respectivelyacrosstechnologiesandforall
countriesduetolimitedavailabilityofinformation.
Productivityeffects Expertestimatesforaparticulartechnologysupportedby
fieldtrialdata.
Inputcostchanges Expertestimatesforaparticulartechnologysupportedby
farm-levelsurveys;changesincostsforparticularinputs
estimatedintermsofrelativeshareinoverallproduction
costs.
Probabilityofresearchsuccess Maximumvalueof80%forquickwinswasassumedand
(theprobabilityofsuccessfullycompletingtheresearch lowervaluesifuncertaintyofresearchsuccessishigher(or
anddevelopingtheintendedtechnologywiththe implementationuncertain—e.g.,GMcrops).Success
desiredcharacteristicssuchashigheryielding,early probabilitiesshouldbedifferentacrosstechnologies,
maturing/bulking,greaterresistancetodiseases, allowingfordifferencesatleastacrossregionsforthesame
greatertolerancetodrought,etc.) technology.Country-levelsuccessprobabilitieswerenot
available,butthesecouldbeincludedinsubsequent
assessments.
Depreciationrate 1%acrossalltechnologiesandcountries
Discountrate 10%
Production Nationalaverageannualproductionfor2010–2012from
FAOSTAT(http://faostat.fao.org/).Wheredatawere
missing,weuseddatafrompreviousyears.
Prices Nationalaverageannualproductionandpricesfor2009–
2011fromFAOSTAT(2013).Wheredataweremissing,
weuseddatafrompreviousyears.
Adoptionprofile Logisticadoptioncurve;adoptionceilingbasedonexpert
estimates(asshareoftotalareainpotentialadoption
domain);timetoreachadoptionceiling(inyears);
adoptionrateinfirstyearofadoptionis1%ofadoption
ceilingforalltechnologies;yearoffirstadoptionandyear
ofdisadoptionbasedontimeframeandexpertassessment.
Twoadoptionscenarios:(1)adoptionscenariobasedon
expertassessmentand(2)conservativeadoptionscenario:
50%ofexpertassessment.
R&Danddisseminationcosts Researchcostsestimatedasthesumof:(1)RTBbudgetsas
presentedintheprogramproposalbythematicarea(some
themesactuallymatchingtheresearchoptionsidentified);
(2)bilateralprojectsatIITAandCIAT(assumedtobe
equaltoRTBbudgets);and(3)NARScosts,whichare
assumedtobeequaltoIITAandCIATbudgetscombined.
Disseminationcostsfornewvarietyis(US$50/ha)and
(US$80/ha)forotherknowledge-intensivetechnologies,
suchascropmanagementinterventions.
Poverty Povertyincidence(%livingonlessthanUS$1.25/day),the
numberofpoorpeople,andagriculturalvalueaddedfrom
WorldBank’sWorldDevelopmentIndicatorsdatabase
(http://data.worldbank.org/indicator).
Agriculturalvalueadded WorldBank’sWorldDevelopmentIndicatorsdatabase
(http://data.worldbank.org/indicator).
Numberofbeneficiaries Country-specificestimatespreparedforRTBproposal:
cropareaperHHforspecificcropandnumberofpersons
perHH.
https://doi.org/10.1371/journal.pone.0201803.t001
PLOSONE|https://doi.org/10.1371/journal.pone.0201803 August8,2018 8/18
CassavaresearchprioritiesinAfrica,Asia,andLatinAmerica
Table2. Dataonthesocioeconomicparametersusedintheeconomicsurplusanalysis.
Country Price(US Quantity(’000 Areaharvested Householdsize Areaper Poverty Numberofpoor AgriculturalValue
$/ton) tons) (’000ha) (persons) farm incidence (million) Added
(ha) (%) (US$billion)
Angola 350 13,673 936 6 0.50 56 10.7 10.6
Benin 470 3,611 251 5 0.50 45 4.0 2.5
BurkinaFaso 268 4 3 5 0.50 45 7.4 3.5
Burundi 374 564 65 5 0.50 81 6.8 0.9
Cameroon 357 3,744 263 5 0.50 9 1.8 4.9
Chad 698 230 22 5 0.50 45 5.0 1.5
Congo 330 1,177 135 5 0.50 53 2.2 0.5
Coted’Ivoire 243 2,309 347 5 0.50 24 4.7 6.2
DRC 330 15,224 1,960 5 0.50 86 56.8 8.1
Ghana 163 13,325 883 4 0.50 25 6.0 9.2
Guinea 354 1,065 129 6 0.50 42 4.2 1.5
Kenya 130 608 64 4 0.50 41 16.4 11.0
Liberia 295 494 62 6 0.50 83 3.3 0.9
Madagascar 171 3,173 473 5 0.50 78 16.2 2.9
Malawi 333 4,028 194 4 0.50 67 10.0 1.3
Mozambique 201 8,501 1,267 5 0.50 60 13.9 4.4
Nigeria 259 43,920 3,449 4 0.50 68 107.2 85.9
Rwanda 299 2,325 196 4 0.50 67 7.1 2.3
Senegal 328 164 26 9 0.50 25 3.1 2.1
SierraLeone 295 446 84 6 0.50 45 2.6 2.2
Togo 174 934 148 5 0.50 39 2.3 1.2
Uganda 120 5,073 417 5 0.50 43 14.3 4.7
Tanzania 210 5,037 898 5 0.50 67 41.5 7.8
Zambia 240 1,193 200 5 0.50 66 8.6 4.0
Argentina 116 182 18 4 0.40 1 0.4 49.1
Bolivia 299 249 29 4 0.50 16 1.6 3.1
Brazil 125 24,907 1,761 5 0.75 6 12.1 123.8
Cambodia 263 4,038 189 4 0.50 19 2.7 4.7
China 127 4,528 277 4 0.25 12 158.6 732.2
Colombia 310 2,166 204 5 0.40 8 3.8 23.5
CostaRica 238 500 34 5 1.00 3 0.1 2.5
Cuba 62 402 71 5 1.00 2 0.2 3.0
Ecuador 245 57 19 5 1.00 5 0.7 7.8
Haiti 160 573 140 5 0.20 62 6.2 1.9
India 160 8,586 245 5 0.60 33 399.1 337.1
Indonesia 198 23,322 1,180 12 0.50 16 39.5 127.0
Jamaica 449 18 1 5 0.75 0.21 0.01 1.0
Laos 160 465 20 5 0.50 34 2.2 2.6
Malaysia 231 48 3 5 0.50 1 0.2 34.6
Paraguay 63 2,563 180 4 0.45 7 0.5 5.5
Peru 165 1,174 100 4 0.40 5 1.5 10.6
Philippines 132 2,118 218 4 0.50 18 17.5 29.2
Thailand 60 24,669 1,210 4 0.50 0.38 0.3 41.5
Venezuela 922 498 36 4 0.50 7 2.0 19.0
Vietnam 112 9,008 521 4 0.50 17 14.8 27.2
Source:FAOSTAT(http://faostat.fao.org/andWorldBank(http://data.worldbank.org/indicator).
https://doi.org/10.1371/journal.pone.0201803.t002
PLOSONE|https://doi.org/10.1371/journal.pone.0201803 August8,2018 9/18
CassavaresearchprioritiesinAfrica,Asia,andLatinAmerica
(i.e.,R&Dcosts);(5)thedisseminationcostsforeachtechnology(eitherUS$80orUS$50for
everynewhectareofadoptiondependingonthetypeoftechnology);and(6)theprobabilityof
researchsuccess.
Sincetheoutcomesofresearchinvestmentscannotberealizedformanyyears,ex-antetech-
nologygenerationandadoptionparameterscanonlybebasedontheopinionsofR&Dexperts
whodrawonawealthofexperienceandknowledgeinmakinginformedpredictions.Mostof
thedatarelatingtocassavatechnologydevelopment,dissemination,andadoptionwere
obtainedprimarilythroughexpertsurveysandconsultations.Expertestimationofthevalues
ofsomeoftheseparametersinvolvedanumberofstepsdesignedtofacilitatetheelicitation
process.Forexample,estimationoftheadoptionceilinginvolvedestimationofthearea
affectedbytheunderlyingconstraintasaproportionofthetotalcroppedareaandtheexpected
adoptionrateasaproportionoftheaffectedarea.ForAfrica,theaffectedareawasthusused
onlytofacilitatetheestimationoftheultimatevalueofadoptionasaproportionofthetotal
croppedarea.Thatis,adoptionasaproportionoftotalcassavaareaisestimatedastheproduct
ofadoptingaproportionoftheaffectedareaandtheaffectedareaasaproportionoftotalarea.
Foralmostallresearchoptions,however,cassavaexpertsworkingespeciallyinAfricaargue
thatmuchofthecassavaareahasbeen(orisexpectedtobe)affectedbytheunderlyingcon-
straints,suchaslowyieldpotential,poorresistancetopestsanddiseases,shortershelflife,and
lackofcleanplantingmaterialmultiplicationanddistributionsystem.Consequently,the
expertsarguethatimprovedseedsystemsandimprovedvarietieswithhigh-yieldattributes
wouldbeappropriateforalmostallrecommendationdomains.However,varietieswithresis-
tancetopestsanddiseasesshouldbedevelopednotonlyforthoseareasthatarecurrently
affectedbythediseasesbutalsoforallareasthatwillbeaffectedinthemanyyearstocome
(includingpre-emptivemeasures).Indeed,usingcurrentlyaffectedareaasarecommendation
domainforadoptionwouldunderstatepotentialadoptionofthosetechnologies.Lookingat
thenatureofmostofourresearchoptionsthatmakeexplicitmentionof“highyield,”theyalso
saythatmuchofthecassavaareashouldbearelevantadoptiondomain,especiallybecause
widergeographicadaptationisalsooneofthekeycriteriaofvarietalrelease.
Ontheotherhand,R&Dcostswereestimatedasthesumof(1)CRP-RTBinvestmentsin
cassavaresearchdisaggregatedbyresearchtheme[17];(2)bilateralprojectfundingforIITA
(mainlyforAfrica)andCIAT(mainlyforAsiaandLAC),whichwasestimatedtobeapproxi-
matelyequaltotheCRP-RTBfunding;and(3)NARSpartnercosts,whichwereassumedtobe
equaltothetotalofCRP-RTBandbilateralfundingthroughIITAandCIAT.Aggregatingthe
costsacrosscountriesforeachresearchoptiongivestheglobalR&Dcostsneededforcalculat-
ingtheglobalNPVsandIRRs.TheCRP-RTBcostswereestimatedbasedontheallocationsin
theRTBprogramproposal.Theannualcassavabudgetwasallocatedacrosstheresearch
options.Forsomeoptionssuchas“plantingmaterials,”theRTBproposalhaddetailsofthe
allocationalreadymadeandonlyrequiredlittleadjustmenttoreallocatetheoverheadsand
CRPmanagementcosts.DisseminationcostswereestimatedtobeUS$50perhectareof
adoptedareafornewvarietiesandUS$80perhectareofadoptedareaforotherknowledge-
intensivetechnologies,suchascropmanagementinterventions.
Table3providesanoverviewoftheparametersrelatedtocassavaresearchandtechnology
disseminationprocess.CassavaresearchinAfricadatesbackto1936,whenscientistsstarted
doingresearchtoaddressmajorproductionconstraintssuchasCMD.However,effortsto
addressCBSDbydevelopingvarietieswithdualresistancetobothCMD(includingthenew
Ugandavariant)andCBSDstartedrecently.Ascanbejudgedfromtheyearwhenresearch
startedtoaddressparticularconstraints,someresearchandtechnologyoptionshavebeenpur-
suedforanumberofyearswhereasotherlinesofresearchstartedonlyrecentlybeforethey
werebothintegratedintothenewRTBprogram(2012–14).Inthisassessment,wetreatall
PLOSONE|https://doi.org/10.1371/journal.pone.0201803 August8,2018 10/18