3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

download 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

of 55

Transcript of 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    1/55

    ReassessingInternationalAgriculturalResearch

    forFoodandAgriculture

    PhilipG.Pardey

    PrabhuL.Pingali

    March2010

    ReportpreparedfortheGlobalConferenceonAgriculturalResearchforDevelopment(GCARD),

    Montpellier,France,2831March2010. PhilipPardeyisaprofessorintheDepartmentofApplied

    EconomicsattheUniversityofMinnesota,andDirectoroftheUniversitysInternationalScienceand

    TechnologyPracticeandPolicy(InSTePP)Center. PrabhuPingaliisaDeputyDirectoroftheGlobal

    DevelopmentProgramoftheBillandMelindaGatesFoundation. PartsofthispaperdrawfromPardey

    andAlston(2010),althoughthatpaperhasanexplicitU.S.focuswhereasthepresentpaperisoriented

    tointernationalresearch. TheauthorsthankConnieChanKang,JasonBeddow,JenniJames,andStan

    Woodforespeciallyvaluableinputintothepreparationofthispaper. Fundingtosupportthe

    preparationofthispaperwasprovidedbytheGlobalForumonAgriculturalResearch,drawingon

    researchfundedbytheBillandMelindaGatesFoundationbywayoftheHarvestChoiceproject(see

    www.HarvestChoice.org),andtheUniversityofMinnesota.

    Copyright(c)(2010)byPhilipG.PardeyandPrabhuPingali

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    2/55

    ABSTRACT

    The20thCenturybeganwitharapidrampingupofnationalinvestmentsinandinstitutions

    engagedwithresearchforfoodandagriculture. Privatephilanthropicorganizationslaunched

    agriculturalR&Dinitiativesaroundthemiddleofthecenturytospurtechnicalchangeinpoor

    countryagriculture. Thisbroadenedtoincludejointlyconceivedpublicandprivateeffortstofund

    internationalagriculturalR&Dinthe1970s. Asthe21stcenturyunfolds,theglobalscienceand

    agriculturaldevelopmentlandscapesarechanginginsubstantiveways,withimportant

    implicationsforthefunding,conductandinstitutionalarrangementsaffectinginternationally

    conceivedandconductedresearchforfoodandagriculture. Whilethereisageneralconsensus

    thatthepresentandprospectivefutureoftheagriculturalsciencelandscapebearslittle

    resemblancetothesituationsthatprevailedintheformativeyearsoftodaysfoodand

    agriculturalresearchpoliciesandinstitutions,manyofthesechangesarepoorlyunderstoodor

    onlybeginningtoplayout. Inthispaperwereportonnewandemergingempiricalevidenceto

    calibratetheprivateandpublicchoicesbeingmadethataffectfoodandagriculturalR&D

    worldwide. Weinvestigatetheresearchlag,benefitappropriability,andR&Dspilloverrealities

    facinginnovativeeffortintheseareas. WealsodiscusstheeconomiesofsizeandscopeofR&D,

    andbroadentheresearchperspectivebeyondinnovationtoencompasstechnology

    development,uptakeandregulation. Seeminglyseismicshiftsintheglobalagricultural

    productivitylandscapesarealsoquantitativelyexamined,alongwithnewinformationonthe

    trendsininvestmentinR&Dthathaveconsequencesforfoodandagriculture.

    Keywords:spillovers,public,private,lags,technologyregulation,productivity,spatial

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    3/55

    CONTENTS

    1. Introduction

    2.

    Policy

    and

    Practical

    Realities

    of

    Food

    and

    Agriculture

    R&D

    2.1 ResearchLagLengths

    2.2 TheShiftingLocationofAgriculturalProduction

    2.3 Appropriability

    2.4 R&DSpillovers

    Spatial

    Disciplinary

    2.5

    Economiesof

    Size,

    Scale

    and

    Scope

    2.6 ResearchTechnologyRegulation3.R&DandProductivity

    3.1 GlobalProductivityPatterns

    3.2 R&DPatterns

    4. TheWayForwardLinkingGlobalR&DtoNaonalNeeds

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    4/55

    ReassessingInternationalAgriculturalResearch

    forFoodandAgriculture

    1. IntroductionInthepasthalfcentury,agriculturalscienceachievedagreatdeal. Since1960,theworldspopulation

    hasmorethandoubled,from3.1billionto6.7billion,andrealpercapitaincomehasnearlytripled.

    Overthesameperiod,totalproductionofcerealsgrewfasterthanpopulation,from877millionmetric

    tonsin1961toover2,351millionmetrictonsin2007,andthisincreasewaslargelyowingto

    unprecedentedincreasesincropyields.1 ThefactthattheMalthusiannightmarehasnotbeenrealized

    overthepast50yearsisattributableinlargeparttoimprovementsinagriculturalproductivityachieved

    throughtechnologicalchangeenabledbyinvestmentsinagriculturalR&D.

    Butthereismuchlefttodo. Recentandsubstantialrunupsinglobalcommoditypriceshad

    directanddetrimentalimpactsonthenumberofhungrypeopleworldwide,andraisedoldconcerns

    abouttheabilityofsustainingincreasesinagriculturalsupplytomeetthefuturefood,feed,fiber,and

    fueldemandsplacedonagriculture.2 ThisinturnraisestensionsandpossibletradeoffsintargetingR&D

    investmentstoaddressglobalfoodsupplyandsecurityconcernsversusR&Ddesignedtomoredirectly

    addressincomedistributionandpovertyconcerns. Compoundingtheseconcernsarethestilllargely

    unchartedanduncertainimplicationsofglobalclimatechangesforworldagriculture.

    Theimmediacyandimportanceoftheseissues,andtheirimplicationsforfoodandagricultural

    R&D,bespeaktheirhistories. Publiclyfundedandconductedresearchforfoodandagricultureonlytook

    holdinthemid tolate1800s,butthenpickeduppaceintheearlydecadesofthe20thCenturyasthe

    scientificunderpinningsofsoilchemistry,Mendeliangenetics,thepurelinetheoryofJohannson,the

    1ObtainedfromUnitedNationsFAO,FAOSTATonlinedatabase,foundathttp://faostat.fao.org. AccessedSeptember

    2009.

    2InSeptember2008theUnitedNationsFoodandAgricultureOrganizationreleasedaprovisionalsetofestimates(FAO

    2008c)indicatingthatthenumberofundernourishedpeoplein2007increasedby75millionoverandaboveFAOs

    estimateof848millionundernourishedin200305,withmuchofthisincreaseattributedtohighfoodprices. Thisbrings

    thenumberofundernourishedpeopleworldwideto923millionin2007,ofwhich907million[are]inthedeveloping

    world. Morerecently,FAO(2009a)estimatedthatanadditional100millionpeoplearenowundernourished,increasing

    thetotaltooveronebillion.

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    5/55

    mutationtheoryofdeVries,andPasteursgermtheoryofdiseasebegantorealizetheirpotentials.3

    Private,philanthropicorganizationssuchastheRockefellerFoundationlaunchedagriculturalR&D

    initiativesaroundthemiddleofthecenturytospurtechnicalchangeinpoorcountryagriculture. This

    broadenedtoincludejointlyconceivedpublicandprivateeffortstofundinternationalagriculturalR&D

    conductedinpurposebuilt,independentcentersofresearchinthe1960s,whicheventuallygaveriseto

    aninstitutionalinnovationtocollectivelyfundthisresearch,namelytheConsultativeGroupon

    InternationalAgriculturalResearch(CGIAR)formedin1971.4

    Asthe21stCenturyunfolds,theglobalscienceandagriculturaldevelopmentlandscapesare

    changinginsubstantiveways,withimportantimplicationsforthefunding,conductandinstitutional

    arrangementsaffectinginternationallyconceivedandconductedresearchforfoodandagriculture.

    Manyofthesechangesarepoorlyunderstoodandsomeareonlybeginningtoplayout,sothe

    magnitudeandevendirectionofthedeparturesfromorthecontinuingpaceofpasttrendsisnot

    known. Nonetheless,theserealitieshaveimportantbearingsontheprivateandpublicchoices

    presentlybeingmaderegardingresearchthataffectsfoodandagriculture. Assemblingwhatweknow

    aboutthesestrategicdevelopmentsandunderstandingtheirlikelyimplicationsarekeytomakingmore

    informedand,hopefully,moreefficientuseofscarceresearchresources,andwhereappropriate

    mobilizingandprioritizingadditionalresearcheffort.

    In

    this

    paper

    we

    focus

    on

    some

    of

    the

    more

    important

    developments

    affecting

    (or

    being

    affected

    by)agriculturalR&Dworldwide. Wereexaminewhatweknowaboutresearchlags,andthe

    appropriabilityofthebenefitsfromresearch,whichhaveadirectbearingonpublicandprivate

    incentivesforinnovationinfoodandagriculture. Apivotaldimensionofinternationallyconceivedand

    conductedresearchforfoodandagricultureisthecrossborderpotentialforresearchdoneinonelocale

    toaffectproductivitygrowthinanother. Tothatendweprovideentirelynewinformationtobetter

    understandtheseresearchspilloverpotentials. WealsodiscussthechangingeconomicsoftheR&D

    3VonLiebigsbookOrganicChemistryinItsApplicationtoAgricultureandPhysiologypublishedin1840inbothGermany

    andGreatBritaintriggeredwidespreadinterestintheapplicationofsciencetoagriculture. Likeothers,Ruttan(1982)

    viewedvonLiebigsbookasthecriticaldividinglineintheevolutionofmodernagriculturalscience.

    4TheCGIARhasexpandeditssubjectmatterscopewellbeyonditsinitialfocusonfoodstaplesanditsinstitutionalscope

    wellbeyondthatofafinancinginstrumentbyassuminggovernance,representational,andserviceprovisionfunctions.

    SeeAlston,DehmerandPardey(2006),PingaliandKelly(2007)andthereferencesthereinfordescriptionsand

    interpretationsofthishistory.

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    6/55

    processitself,notablytheeconomiesofsizeandscopeofR&D,andbroadentheresearchperspective

    beyondinnovationtoencompasstechnologydevelopment,uptakeandregulation.

    Atroot,concernsaboutfoodandagriculturalsuppliesareconcernsaboutthepace,nature,

    directionandconsequencesofagriculturalproductivitygrowth,andsowereviewwhatispresently

    knownabouttheseproductivitypatterns. SomeoftheimportantR&Dinvestmenttrendsthat

    circumscribeproductivitypotentialsarealsointroducedandbrieflydiscussed.

    2. PolicyandPracticalRealitiesofFoodandAgricultureR&DInnovationinagriculturehasmanyfeaturesincommonwithinnovationmoregenerally,butalso

    someimportantdifferences. Inmanywaysthestudyofinnovationisastudyofmarketfailureandthe

    individualandcollectiveactionsnotablyinvestinginagriculturalR&Dtakentodealwithit. Like

    otherpartsoftheeconomy,agricultureischaracterizedbymarketfailuresassociatedwithincomplete

    propertyrightsoverinventions. Theatomisticstructureofmuchofagriculturemeansthatthe

    attenuationofincentivestoinnovateismorepronounced(andparticularlysoinmanyofthepoorest

    partsoftheworldwheretheaveragefarmsizeatleastasdenominatedbyfarmedareaissmall,

    andgettingsmaller)thaninotherindustriesthataremoreconcentratedintheirindustrialstructure.

    Ontheotherhand,unlikemostinnovationsinmanufacturing,foodprocessing,ortransportation,

    technologiesusedinproductionagriculturehavedegreesofsitespecificity. Thebiologicalnatureof

    agriculturalproductionmeansthattheappropriatetechnologyoftenvarieswith(local,sometimeson

    farm)variationinclimate,soiltypes,topography,latitude,altitude,anddistancefrommarkets. The

    sitespecificaspectcircumscribes,butbynomeansremoves,thepotentialforknowledgespillovers

    andtheassociatedmarketfailuresthatareexacerbatedbythesmallscale,atomisticindustrial

    structureofagriculture.

    Theseandsomeotherimportantrealitiesofresearchforfoodandagriculturearepoorly

    understoodbythosenotfamiliarwiththefacts. Otherrealitieshavechangedinwaysthatwehave

    failedtoproperlymeasureoradequatelyinvestigate. Inaddition,importanttechnical,marketand

    climatecircumstancesintheyearsaheadmaybedifferentinimportantrespectstoourmeasured

    past. Itistosomeofthepastandemergingrealitiesthatmaywellhavestrategicpolicyandpractical

    implicationsregardinginternationalresearchforfoodandagriculturethatwenowturn.

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    7/55

    2.1 ResearchLagLengths

    Theprojectbasedfundingthatcomesmainlyfromaidagencies(ratherthanscienceoreven

    agriculturalministries)andisnowtheprevalentformofsupportforCGIARresearchofteninvolvesup

    tothreeor,insomelimitedcases,fiveyearfundingcycles,onlysomeofwhicharerenewedand

    sustainedforlongerperiods.5 Impatientorpoliticallyconstrainedfundersalsofrequentlypressurefor

    demonstrableevidenceofdevelopmentimpactsattheterminationoftheseprojects,orasa

    preconditionforfurtherfunding. Unfortunately,manyoftheseagriculturalresearchinvestment

    initiativesaremyopic;fundamentallymisconstruingthenatureandlengthofthelagsbetweenR&D

    investmentsandtheeconomicandsocialreturnsrealizedfromthatinvestment. Infact,theselagsare

    generallylong,oftenspanningdecades,notmonthsoryears.

    Thedynamicstructurelinkingresearchspendingandproductivityinvolvesaconfluenceof

    processesincludingthecreationanddestructionofknowledgestocksandtheadoptionand

    disadoptionofinnovationsoverspaceandtimeeachofwhichhasitsowncomplexdynamics. The

    scienceinvolvedisacumulativeprocess,throughwhichtodaysnewideasarederivedfromthe

    accumulatedstockofpastideas. Thisfeatureofscienceinfluencesthenatureoftheresearch

    productivityrelationshipaswell,makingthecreationofknowledgeunlikeotherproduction

    processes. Theevidenceforlongresearchproductivitylagsiscompelling. Oneformofevidence

    stemsfromstatisticaleffortstoestablishtherelationshipbetweencurrentandpastR&Dspending

    andagriculturalproductivity. Thedozensofstudiesdonetodateindicatethattheproductivity

    consequencesofpublicagriculturalR&Daredistributedovermanydecades,withalagof1525years

    beforepeakimpactsarereachedandwithcontinuingeffectsfordecadesafterwards.6

    ThestatisticalevidencelinkingoverallinvestmentsinaggregateagriculturalR&Dtoagricultural

    productivitygrowtharereinforcedbytheotherevidenceaboutresearchandadoptionlagprocessesfor

    particulartechnologies,especiallycropvarietiesaboutwhichwehavealotofspecificinformation. The

    developmentanduptakeofvarietaltechnologiesworldwidehasbeenmuchstudied(see,forexample,

    5Likewise,counterpartfundingformuchpublicresearchconductedbynationalagencieshasbecomeincreasingly

    contestableandprojectbased.

    6Alstonetal.(2010seealsofootnote2)reviewedthepriorliterature. Theyalsodevelopedtheirownestimatesusing

    newlyconstructedU.S.statelevelproductivityover19492002andU.S.federalandstatespendingonagriculturalR&D

    andextensionover18902002. Theirpreferredmodelhadapeaklaggedresearchimpactatyear24andatotallaglength

    of50years.

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    8/55

    EvensonandGollin2003),butarguablythemostcomprehensiveevidenceonthesetechnicalchanges

    overthepastcenturyormorehasbeenassembledfortheUnitedStatesandisillustrativeofthemore

    generalpicture.

    Figure1providesnewdataonthreewavesofvarietaltechnologiesintheUnitedStates

    beginningintheearly1900s. Hybridcorntechnology,whichtookoffinU.S.farmersfieldsinthe1930s,

    haditsscientificrootsinfocusedresearchthatbeganin1918(andarguablybeforethen,atleasttothe

    early1890s). ThustheR&Dorinnovationlagwasatleast10yearsandmayhavebeen2030years. The

    timepathoftheadoptionprocessesextendsthelaglengthsevenfurther. Iowahad10percentofits

    cornacreageplantedtohybridsin1936(with90percentofitscornacreagesoplantedjustfouryears

    later),whileittookuntil1948beforeAlabamaastatewithdistinctiveagroecologicalattributes

    comparedwiththeprincipalCornBeltstateshad10percentofitscornacreageunderhybrids. By

    1950,80percentandby1960,almostallofthecorngrownintheUnitedStateswashybridcorn.

    Lookingacrossallthestates,thetechnologydiffusionprocesswasspreadoverabout30years,reflecting

    theenvelopeofadoptionprocessesthatweremuchmorerapidinanyindividualstate. Takingthe

    entireresearch,development,andadoptionprocessforhybridcornashavingbegunaslateas1918,the

    totalprocessthathadbeenaccomplishedby1960tookplaceoveraperiodofatleast40yearsand

    possiblydecadeslonger.

    The

    semi

    dwarf

    wheat

    and

    rice

    varietal

    technologies

    that

    lay

    at

    the

    heart

    of

    the

    Green

    RevolutionalsofoundtheirwayintoU.S.agricultureviaadaptiveresearch. Thefirstcommercially

    significantuseofsemidwarfwheatsintheUnitedStatesoccurredin1961. Theearly(andmostrapid)

    uptakeofthistechnologywasinCalifornia,withagroecologiesmuchlikethoseinNorthernMexico

    whereNormanBorlaugbredmostoftheearly,shortstaturedCIMMYTvarieties. Thelargewheatbelt

    statesoftheDakotasandMinnesotahaddistinctiverustandotherdiseaseproblemsthatdelayedthe

    entryofsemidwarfnessintotheselocalesuntilresistancetothesebioticconstraintswascrossbred

    intoshort

    statured

    wheats.

    Thus

    ittook

    30

    years

    before

    80

    percent

    of

    the

    U.S.

    wheat

    acreage

    was

    plantedtosemidwarfvarieties.

    Withitsemphasisonvarietalquality,thespreadofhigheryielding(butinitiallyatleast,less

    appealingtoeat)semidwarfricevarietiesintheUnitedStateslaggedconsiderablybehindthe

    irrigatedareasinAsiawherethepriorityinthe1960sand1970swastoraisecropyieldsandincrease

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    9/55

    riceproduction(HerdtandCapule1983). Beginningin1979,semidwarfricevarietiesgained

    acceptanceintheUnitedStates,butby2005only68percenttotheU.S.riceacreagewasplantedto

    varietieswiththischaracteristic.

    Hasmodern(bio)technologymateriallyspedupthisresearchinnovationadoptionprocess,as

    iscommonlysuggested? Geneticallyengineered(GE)cornwasfirstplantedonU.S.farmersfieldsin

    themid1990s. TheadoptioncumdiffusionprocessforGEcropsisnotyetcomplete,thetechnology

    itselfiscontinuingtoevolve,andthemaximumadoptionratehasnotyetbeenachieved;by2008,80

    percentofU.S.cornacreagewasplantedtoGEvarieties. Likehybridcorn,biotechcornhasbeen

    adoptedatdifferentratesindifferentstates,butperhapsfordifferentreasons. This,asyet

    incomplete,processoverlessthan15yearsrepresentsonlypartoftherelevanttimelag. Tothatwe

    mustaddthetimespentconductingrelativelybasicandappliedresearchtodevelopandevaluatethe

    technology,andthetime(andmoney)spentafterthetechnologyhadbeendevelopedtomeetthe

    requirementsforregulatoryapprovalbyarangeofgovernmentagencies.

    ComparedwiththeadoptioncumdiffusionprocessforhybridcornwithintheUnitedStates,

    theprocessforbiotechcornappearstohavebeenalittlefaster. Themaindifferencemaybethatall

    statesbegantoadopttogether,withouttheslowerspatialdiffusionamongstatesthatcharacterized

    hybridcorn,possiblybecauseofimprovedcommunicationsandfarmereducation,perhapsassistedby

    public

    extension

    services.

    Thus

    biotech

    corn

    achieved

    80

    percent

    adoption

    within

    13

    years

    compared

    with19yearsforhybridcornor30yearsforsemidwarfwheat. However,otherelementsofthe

    processmaybegettinglonger. Forinstance,theprocessofregulatoryapprovalmayhaveaddeda

    further510yearstotheR&Dlag(andthisregulatoryapprovallagforbiotechcropsappearstobe

    gettinglonger). Givenarangeof10to20yearsspentonR&Dtodevelopthetechnologiesthat

    enabledthecreationofbiotechcrops,andthenthetimespenttodeveloptheinitialvarietiesand

    improvethem,theoverallprocessofinnovationinthecaseofbiotechcornmayhavetaken20to30

    yearsso

    far.

    Insum,theseU.S.examplesspanaspectrumofresearchrealities:researchtodevelop

    fundamentallynew(bioengineered)traitsforspecificcrops,researchtoadaptandfacilitatespillinsof

    semidwarftechnologiesoriginatingelsewhereintheworld,andtheaggregateproductivity

    promotingconsequencesofoverallspendingonagriculturalR&D. Thesecaseshelpanchorour

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    10/55

    expectationsabouttheconsiderablelagsinvolvedinrealizingsocialandeconomicvaluefrom

    investmentsinR&D,eveninacountrysuchastheUnitedStatesthatisnotundulyconstrainedby

    limitedruralinfrastructure,poorcommunications,institutionalinstabilities,andrestrictiveseed

    release(andrelated)commercializationpoliciesandpractices. Bythismeasurealone,investmentsin

    agriculturalR&Darebestseenasanespeciallyeffectivemeansofachievinglongruneconomic

    growthanddevelopmentobjectivesspanningmanydecades,ratherthananinterventioninstrument

    toachievenearterm,incomedistributionoreconomicdevelopmentobjectives.

    2.2 TheShiftingLocationofAgriculturalProduction

    PolicymyopiaisoneproblemconfrontingagriculturalR&D. Another,andrelated,problemisa

    seeminglywidespreadlackofappreciationofthespatialmobilityofagriculture. Contrarytocommon

    perceptions,agriculturemoves,sometimesmarkedly,overthelandscape. Hence,thepresentlocation

    ofproductionofaparticularcropmaynotbeagoodindicationofwhereintheworldthatcropwillbe

    growndecadesfromnow. Thisideahasimportantconsequencesforfoodandagriculturalresearch.

    Theproductivityperformanceofmanyagriculturaltechnologiesissensitivetolocalagroecological

    factors(includingclimate,soils,landslopeandelevation,wind,anddaylength),andsotargetingand

    optimizingtechnologiesfortheseagroecologicalrealitiesisadistinctiveaspectofinnovationinfoodand

    agriculture. Coupledwiththeinherentlylonglagsfrominitiatingresearchtorealizingimpacts,itmay

    wellbefollytoprioritizeinvestmentsonR&Dtacklingaparticularprobleminaparticularcrop(or

    livestock)commodityassumingthepresentspatialpatternofproductionwillprevail.

    Thefactorsaffectingthelocationofproductionarecomplexandchanging. Moreover,

    technologiesthemselvesmayshifttheoptimallocationofagriculturalproduction. Pressuresoutside

    agricultureandbeyondconsiderationsofagroecologiesarealsoimportant. Climatechange,for

    instance,mayhaveabigbearingontheoptimallocationofproduction,orthetechnicalstrategiesbest

    suitedtoadaptingtothesechangesinagivenlocale. Investmentsinruraltransport,coldchain,and

    communicationinfrastructurealongwiththechangingspatialpatternsof(ruralvsurban)population

    densitiescandemonstrablyaffecttheagriculturallandscape. Thusasmarketaccessimproves,local

    productionincentivescanbeskewedtowardhighervalued,perishableproduction(suchasfreshfruits

    andvegetables,meatanddairyproducts)andawayfromstapleormoretraditionalfoodcrops.

    Likewise,investmentsinirrigation,terracingandotheragriculturallandimprovementscanalterthe

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    11/55

    incentivestoproducecertainagriculturalproductsincertainlocations,withsubstantivefollowon

    consequencesforR&Dpriorities.7

    CroplandMovements

    Sowhatevidencedowehaveoftheextentandnatureofthespatialmovementofagricultural

    production? Unfortunately,thisaspecthasbeenlittlestudied,butthereisasmallandgradually

    growingbodyofevidence,someofwhichisbrieflyintroducedhere. Agriculturetakesupalotofspace:

    anestimated40percentoftheworldslandareaispresentlycommittedtocropandlivestock

    production(withalmost13percentofthelandbeingincrops). Butthatwasnotalwaysso. Beginning

    in1700,agriculturalcroplandoccupiedjust3.5percentoftheworldstotallandarea,withmostofthat

    croplandlocatedinAsia(accountingfor48.5percentoftheworldscroppedareaatthattime),Europe

    (28.5percent),andAfrica(19.6percent). Notably,thesparselysettledNewWorldsofAustralia,New

    Zealand,andtheAmericascollectivelyaccountedforjust3.2percentofthelandworldwideunder

    permanentcropsin1700. By2000,theNewWorldsharehadgrownto27.1percentofthetotal

    croppedarea.

    DrawingonsimulatedSAGEdatadevelopedbyRamankuttyandFoley(1999)andRamnakutty

    etal.(2008),Beddowetal.(2010)illustratechangesinthespatialpatternofproductionoverthelong

    run. Figure2,Panelsaandbprovidemappedsnapshotsofthelocationofcroppedareain1700and

    2000

    respectively.

    The

    net

    effect

    of

    the

    movement

    of

    land

    in

    and

    out

    of

    cropped

    agriculture

    means

    thatagricultureisgeographicallymobile,asillustratedinFigure2,Panelc,whichusestheSAGEseries

    toestimatechangesincroppedareaoverthefourdecadesspanning1960to2000. Itindicatesthe

    localizedmovementofacreageinandoutofagriculturesince1960,or,morespecifically,thechange

    intheareasharededicatedtocropproductionforeachofthe259,200mappedpixelsforexample,a

    valueofminus50percentindicatesthathalftheacreageinthatpixelshiftedoutofcropping

    agriculturesince1960. Thedarkertheredshading,thegreaterthepercentdeclineincroppedarea

    7Whilegroundingtechnologytargetstolocalproductionconstraintsiscriticallyimportant,thesespatialdynamics

    complicatedecisionsaboutwhoseparticularproductionconstraintshavebearing. Isittodaysfarmersgiventodays

    productionproblems,ortomorrowsfarmersandtheirprospectiveproblemsthataremostrelevant? Clearlyinsome

    casesthetwosetsofproblemsareinessencethesame,butthiswillnotalwaysbeso. Moreover,farmersarenotalways

    fullyinformedaboutscientificpotentials,andtotheextentthatlocationandproductionchoicesareendogenously

    determinedbytechnicaloptions,scientificopinionisimportantaswell. Forexample,onemightspeculatethattherewas

    littleifanyfarmerdemandforsemidwarfnessinwheatorricetechnologies,yettheproductivityboostofthesevarietal

    innovationsweregloballytransformative.

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    12/55

    perpixel;thedarkerthegreenshading,thegreaterthepercentincreaseincroppedareaperpixel.

    ThecollapseoftheformerSovietUnionisevidentintermsofsubstantialdeclinesincroppedarea

    throughoutEasternEurope. TheSAGEdataalsoindicatedeclinesincroppedareainpartsofWestern

    Europe,northeastern,southern,andsoutheasternUnitedStates,andsignificantpartsofChina.8

    TherewasasubstantialincreaseincroppedareasthroughouttheIndochinaPeninsula,Indonesia,

    WestAfrica,Mexico,andBrazil. Theoverallpictureisoneofcontractingareaundercropsin

    temperateregionsandincreasingcroppedareaintropicalpartsoftheworldduringthelastfour

    decadesofthetwentiethcentury.

    Figure2,Paneld providesanindicationofthedistanceanddirectionofthespatialrelocation

    ofagriculturegloballyoverthelongrunbyplottingthemovementinthecentroidsorcentersof

    gravityofproductionbyregionfortheperiodbeginningin1700(wheneachregionscentroidis

    centeredonazerolatitudelongitudegridcoordinate)throughto2000. Eachcentroidisanestimate

    ofthegeographiccenter(centerofmass)ofthecroppedareainthecorrespondingregion. The

    locationofthecentroiditselfisnotparticularlyenlightening,anditcouldeasilybethecasethata

    centroidisinalocationthatdoesnotproduceanycropsatall,orisotherwisenotrepresentativeof

    thegeneralagriculturalsituationinacountry. However,movementsinthecentroidarerevealingas

    anindicationoftheinfluencesofchangingpatternsofsettlement,infrastructure,andtechnologieson

    the

    location

    of

    agriculture.

    Accordingtothesedata,NorthAmericaandAfricahaveseenthelargestmovementsintheir

    productioncentroids,bothshiftingabout1,300kilometersoverthe300yearperiod. Aswasthecase

    withtheothercontinents,mostofthismovementoccurredafter1900. However,theyear2000

    centroidsforotherregionsmoreorlessrepresentacontinuationofthetrendfrom1950to1992;the

    onlyanomalyseemstobeinAfrica,wherealmostallofthemeasuredmovementinitscentroid

    occurredbetween1992and2000.9 TheAsiancentroidmovedtheleast,changingbyonly15

    kilometersto

    the

    east

    and

    137

    kilometers

    to

    the

    south.

    8Wood,Sebastian,andScherr(2000,p.28)documentthereductionincultivatedlandinChinaduringthefirsthalfofthe

    1990s,largelyattributingthistoexpandedindustrialandurbanusesofland.Zhangetal.(2007)implythatthistrend

    continuedintoatleasttheearlypartofthetwentyfirstcentury.Forexample,theauthorsestimatethat 260,000haof

    Chinesecultivatedlandwasconvertedtononagriculturalusesbetween1991and2001.

    9Itseemsmorelikelythattheyear1992and2000datasetswerenotfullyconformablethanthatamassivestructuralshift

    inAfricanproductionoccurredduringthisperiod.However,thenorthwardmovementofagricultureinsubSaharanAfrica

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    13/55

    ExceptinAfricaandAsia,thegeneraltrendfavoredmovementinlongituderatherthan

    latitude.ThepronouncednorthwardmovementinAfricawasalmostmatchedbyanequivalentmove

    westward,and,whiletheAsiancentroidshowedmuchmoreabsolutemovementalongtheeastwest

    axis,thenetmovementovertheperiodwasalmostduesouth. Averagingacrossalloftheregions,the

    netlongitudinalmovementwas4.6timesaslargeasthenetlatitudinalmovement.

    MovementofCrops

    Sometechnologytargeting,andtheirimpliedagriculturalR&Dinvestmentchoices,areusefully

    informedbybroadbrushperspectivesonthespatialmobilityofaggregateagriculturalorcropped

    land. ButmanyR&Dinvestmentchoiceshingeonmorerefined,cropspecificsensesofthepresent

    andchanginglocationofproduction.

    Diggingbeneaththeaggregatecropareasjustdiscussed,whatdoweknowaboutchangesin

    thelocationofproductionofindividualcrops,especiallyatagloballevelandforthelengthyperiods

    requiredfortechnical(andother)changestohaverealizedtheirfullproductionandproductivity

    consequences? AppendixTable1revealsanewlycompiledglobalseriesusedbyBeddow(2010)to

    examinecountry andcropspecificproductiontrendsstretchingbacktothe1880s. Thetablereports

    thelongrunhistoryofcountryspecificproductionbyperiodandthecorrespondingshareof

    production(inbrackets)formaize,wheatandrice. Thetabulationincludesthetopfiveproducersfor

    these

    three

    crops

    and

    how

    those

    producers

    evolved

    over

    time

    relative

    to

    other

    countries.

    If

    a

    country

    appearedasatopproducerforacropinanyparticulartimeperiod,thatproducer'srankand

    percentageshareareshownforallreportedtimeperiods. Thuswefind,forexample,thatJapanwas

    oncethesecondrankedproducerofrice,buthasfalleninrankandshareoverthelongrunandisno

    longeramongthetopfive(aviewthatcannotbeobtainedfrommediumrundatalikethoseavailable

    fromFAOSTAT).

    Thereareseveralstrikingfeaturesinthesedata. First,measuredglobalproductionhasbeen

    spatiallyconcentrated,

    especially

    for

    maize

    and

    rice.

    Since

    the

    beginning

    of

    the

    20

    thCentury,

    the

    top

    twoproducingcountrieshavealwaysaccountedformorethanhalftheglobalproductionofmaize

    andrice,andoften7080percentoftheworldproductionoccurredinjustfivecountries. Wheat

    isconsistentwiththefindingofLiebenberg,Pardey,andKahn(2010)thatthefarmedareainSouthAfricanagriculture

    peakedat91.8millionhectaresin1960,thendeclinedsteadilyto82.2millionhectaresby1996,whereithassincebeen

    moreorlessstable.

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    14/55

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    15/55

    reducedthedegreedayrequirementsofmaize,allowingfarmerstocompleteaseasonfurthernorth

    thanwasotherwisepossible.

    In1879,thefirstyearforwhichbothoutputandareadataareavailableatthecountylevelfor

    theUnitedStates,therewasaconcentrationofmaizeproductioninthelowermidwest. However,

    maizewasproducedfairlyhomogenouslythroughouttheeasternportionsofthecountrysothatin

    total,26percentofthecountry'smaizewasproducedintheSoutheastand42percentwasproduced

    intheNortheast. By2006,theSoutheasthadlostitsstatusasamajormaizeproducer,whilethe

    locusofproductionshiftedtotheNorthCentralregion. Bothtechnologicalandnontechnological

    factorsspurredthisrelocation. Someoftheshiftwasmadepossiblebyimprovedtransportation

    systems,whichallowedSoutheasterngrowerstoproducemorenonnutritivecropssuchascotton

    andtobacco,andlaterbytherapiduptakeofhybridtechnologyintheNorthCentralportionofthe

    country.

    Thesamespatialprocessesarenodoubtatplayinotherpartsoftheworld,althoughthepace

    andspecificsofthelocationalchangeshaveyettobecarefullyassessed. Arguably,therapidpaceof

    urbanizationandthepotentialtoradicallyaffecteconomicaccesstomarketsviaimprovementsin

    transportationandcommunicationinfrastructurepointstothepossibilityofmajormovementin

    Africanagricultureinthedecadesahead. Compoundingthesepressuresforspatialchangein

    agriculture

    are

    the

    prospects

    of

    localized

    changes

    in

    production

    potential

    attributable

    to

    changes

    in

    climate.

    2.3 Appropriability

    Thepartialpublicgoodnatureofmuchoftheknowledgeproducedbyresearchmeansthat

    researchbenefitsarenotfullyprivatelyappropriable. Indeed,themainreasonforprivatesector

    underinvestmentinagriculturalR&Disinappropriabilityofsomeresearchbenefits:thefirm

    responsiblefordevelopingatechnologymaynotbeabletocapture(i.e.,appropriate)allofthe

    benefitsaccruingtotheinnovation,oftenbecausefullyeffectivepatentingorsecrecyisnotpossible

    orbecausesomeresearchbenefits(orcosts)accruetopeopleotherthanthosewhousetheresults.

    Forcertaintypesofagriculturalresearch,therightstotheresultsarefullyandeffectivelyprotectedby

    patentsorotherformsofintellectualpropertyprotection,suchthattheinventorcancapturethe

    benefitsbyusingtheresultsfromtheresearchorsellingtherightstousethem;forinstance,the

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    16/55

    benefitsfrommostmechanicalinventionsanddevelopingnewhybridplantvarieties,suchashybrid

    corn,areappropriable. Often,however,thosewhoinvestinR&Dcannotcaptureallofthebenefits

    otherscanfreerideonaninvestmentinresearch,usingtheresultsandsharinginthebenefits

    withoutsharinginthecosts.11 Insuchcases,privatebenefitstoaninvestor(orgroupofinvestors)are

    lessthanthesocialbenefitsoftheinvestmentandsomesociallyprofitableinvestmentopportunities

    remainunexploited. Theupshotisthat,intheabsenceofgovernmentintervention,investmentin

    agriculturalresearchislikelytobetoolittle.

    Thetypesoftechnologyoftensuitedtolessdevelopedcountryagriculturehavehithertobeen

    ofthesortforwhichappropriabilityproblemsaremorepronouncedtypesthathavebeen

    comparativelyneglectedbytheprivatesectorevenintherichestcountries. Inparticular,until

    recently,privateresearchhastendedtoemphasizemechanicalandchemicaltechnologies,whichare

    comparativelywellprotectedbypatents,tradesecrecy,andotherintellectualpropertyrights;andthe

    privatesectorhasgenerallyneglectedvarietaltechnologiesexceptwherethereturnsare

    appropriable,asforhybridseed. Inlessdevelopedcountries,theemphasisininnovationhasoften

    beenonselfpollinatingcropvarietiesanddisembodiedfarmmanagementpractices,whicharethe

    leastappropriableofall. Therecentinnovationsinrichcountryinstitutionsmeanthatprivatefirms

    arenowfindingitmoreprofitabletoinvestinplantvarieties;thesamemaybetrueinsomeless

    developed

    countries,

    but

    not

    all

    countries

    have

    made

    comparable

    institutional

    changes.

    2.4 R&DSpillovers

    Whilethemostimmediateandtangibleeffectofthenewtechnologiesandideasstemmingfrom

    researchdoneinonecountryistofosterproductivitygrowthinthatcountry,newtechnologiesand

    ideasoftenspilloverandspursizableproductivitygainselsewhereintheworld. Inthepast,low and

    middleincomecountriesbenefitedconsiderablyfromtechnologicalspilloversfromhighincome

    countries,inpartbecausethebulkoftheworldsagriculturalscienceandinnovationoccurredinrich

    countries. AsPardeyandAlston(2010)observed,increasingly,spilloversfromrichcountriesmaynot

    11Forinstance,anagronomistorfarmerwhodevelopedanimprovedwheatvarietywouldhavedifficultyappropriating

    thebenefitsbecauseopenpollinatedcropslikewheatreproducethemselves,unlikehybridcrops,whichdonot. The

    inventorcouldnotrealizeallofthepotentialsocialbenefitssimplybyusingthenewvarietyhimself;butifhesoldthe

    (fertile)seedinoneyearthebuyerscouldkeepsomeofthegrainproducedfromthatseedforsubsequentuseasseed.

    Hencetheinventorisnotabletoreapthereturnstohisinnovation.

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    17/55

    beavailabletodevelopingcountriesinthesamewaysortothesameextentforseveralreasons. First,

    richcountryR&Dagendashavebeenreorientedawayfromproductivitygainsinfoodstaplestoward

    otheraspectsofagriculturalproduction,suchasenvironmentaleffects,foodquality,andthemedical,

    energy,andindustrialusesofagriculturalcommodities. Thisgrowingdivergencebetweendeveloped

    countryresearchagendasandtheprioritiesofdevelopingcountriesimpliesthatfewerapplicable

    technologieswillbecandidatesforadaptationtodevelopingcountries. Second,technologiesthatare

    applicablemaynotbeasreadilyaccessiblebecauseofincreasingintellectualpropertyprotectionof

    privatelyownedtechnologiesand,perhapsmoreimportantly,theexpandingscopeandenforcement

    ofbiosafetyregulations. Differentapproachesmayhavetobedevisedtomakeitpossiblefor

    countriestoachieveequivalentaccesstotechnologicalpotentialgeneratedbyothercountries. Third,

    thosetechnologiesthatareapplicableandavailablearelikelytorequiremoresubstantiallocal

    developmentandadaptation,callingformoresophisticatedandmoreextensiveformsofscientific

    R&Dthaninthepast. Therequirementforlocaladaptiveresearchisalsolikelytobeexacerbatedas

    changesinglobalandlocalclimatepatternsaddfurthertotheneedforadaptiveresponsesto

    changingagriculturalproductionenvironments. Notwithstandingthesedevelopments,itisimperative

    thatbothnationalandinternational(spillover)potentialsbeoptimizedinthedecadesaheadifthe

    necessaryglobalproductivitygainsaretoberealized.

    Spatial

    Spillovers

    Analysesofagriculturalproductivitygainshaveshownthatspatialspillinsareamajorsourceof

    productivitygains,accountingforuptohalfof localproductivity increases. Thepotentialforspatial

    spilloversgoestotheheartoftheconceptionofandraisondtreforinternationalagriculturalR&D,

    whetherthatbeconducted inthepublicarena(suchasbyCGIARfundedcenters)orbyregionalor

    multilateralprivatefirms. Absentthesespatialspilloversthemarketfailurerationale(and,relatedly,

    the size and scope rationale discussed below) for internationally conceived or conducted R&D is

    severelycurtailed.

    What

    do

    we

    know

    about

    the

    likelihood

    for

    research

    or

    technologies

    to

    spill

    from

    onecountrytoanother?

    Because agricultural production is especially dependent on natural inputs such as soil and

    climate conditions which affect the performance of particular crops or production practices, the

    degreeofagroecologicalsimilarityaffects thedegree towhichspillinscanbeexploited. Countries

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    18/55

    that share agroecological characteristics are likely to have high potential for spilloversi.e.,

    technologies or crop varieties developed in one country may be readily adopted in the other.

    Similarly,spillinsalsotendtoflowmorereadilyamongcountriesthatproducesimilarcropmixes. On

    thecontrary,technologicalspilloverswillbelimitedamongcountriesthataretechnologicallydistant,

    ordissimilarintheiragroecologicalcharacteristicsorproductionpatterns.

    James, Pardey and Wood (2010) develop and report a range of metrics of the technological

    distancebetweencountries. Theirdistancemetricrangesbetweenzeroandoneoneindicatingthat

    countriesare technologicalclose (and so thepotential for technologyspilloversarehigh),andzero

    indicatingtheyaretechnologicaldistant(with lowornospilloverpotential). InFigure4,distance is

    establishedbyassessing thedegreeofconcordance in thecropmixamongcountries. Panela, for

    example, shows the concordance in crop area shares for each country relative to a richcountry

    averageoftheareasharesplantedtoeachof20crops. Thus,iftheshareofcroppedacreageplanted

    toeachof20cropsforaparticularcountrywereidenticaltothecorrespondingareasharesaveraged

    among the highincome countries, then the distance metric would take the value 1.0: that is, the

    countryinquestionistechnologicalclosetothehighincomecountriesasagroupwhenviewedfrom

    theperspectiveofitscroporientation. Byextension,onewouldexpectacountrywhosecropmixis

    similar instructuretothemixofcropsproduced inthehighincomecountries,onaverage,tohave

    greater

    potential

    to

    capture

    technological

    spillins

    from

    the

    research

    done

    in

    those

    rich

    countries.

    Figure4,Panelsb, candd report the same cropbaseddistancemetricsusing the croparea

    averages forLatinAmerica&Caribbean,AsiaandPacific,andsubSaharaAfrica respectivelyas the

    pointof reference. Table1 reports theaveragevaluesof thecropdistancemetric forcountries in

    eachregionoftheworldrelativetothesefour,baseregionaverages. Bythismeasure,countries in

    subSaharan Africa have comparatively low potential to capture technological spillins from crop

    researchdoneintherichcountries(seethedistancemetricvalueof0.40). Onaveragethecropping

    patternsinLatin

    America

    are

    closest

    to

    those

    insub

    Saharan

    Africa,

    although

    the

    concordance

    of

    crop

    mixesisstillquitelowbyinternationalstandards(seethedistancemetricvalueof0.54).

    Similarityincropproductionmixisbutonedimensionoftechnologicalcloseness. Eveniftwo

    countrieshadsimilarcroppingshares,itmaybethattheagroecologicalconditionsfacingcrop

    productioninonecountryaredissimilartothoseinanothercountry,meaningdifferentcropvarieties,

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    19/55

    cropmanagementpracticesorinputmixesarerequired. Theseagroecologicaldissimilaritieswould

    acttounderminethepotentialforresearchspillovers(or,alternatively,raisethecostsoftheadaptive

    researchrequiredtoporttechnologydevelopedinonecountrytoanagroecologicallydissimilarother

    country). Figure5parallelsFigure4initsconstruction,butthistimetheagriculturalareasineach

    countrywereparsedinto26differentagroecologicalclassesandtheconcordanceamong

    agroecologieswasassessed. Mostevidently,countriesthroughoutsubSaharanAfricaaremuchmore

    distantfromtherichcountriesonaverageintermsoftheiragroecologiesthantheircropmixes(see

    thegenerallylightershadingthatislowerdistancemetricvaluesforsubSaharanAfricainPanelaof

    Figure5comparedwithPanelaofFigure4). Infact,theagriculturalareasincountriesthroughout

    subSaharanAfricaareagroecologicallyclosesttotheagriculturalareasinLatinAmerica(and,

    specifically,Brazil). TheyarealsoreasonablyclosetoareasthroughoutSouthandEastAsia,notably

    IndiaandpartsoftheIndoChinapeninsular(seethedarkershadedcountiesinFigure5,Paneld,

    wheretheaveragecropecologythroughoutsubSaharanAfricawastakenasthepointofreference

    forcalculatingdistancemetrics).

    Figure6goesonestepfurthertojointlyevaluatetechnologicaldistanceintermsofthe

    agroecologicaldifferencesamongcountieswithinspecificcroppingareas. Herethereference

    regionistheagroecologiesfoundinthetopfiveproducingcountriesforeachofthefourincluded

    crops;

    wheat,

    rice,

    maize

    and

    soybean.

    Thus,

    for

    example,

    countries

    throughout

    sub

    Saharan

    Africa

    generallyhavequitedissimilaragroecologiescomparedwiththeagroecologiesfoundinthewheat

    growingareasoftheworldsleadingwheatproducers(Figure6,Panela). Incontrast,partsofwest,

    northcentralandeasternAfricaareagroecologicallyclosertotheworldsprincipalriceproducers

    (Figure6,panelc),suggestingthatricetechnologies(e.g.,newvarietiesorcropmanagement

    techniques)emanatingfromtheseimportantriceproducingcountrieshavegreaterpotentialtospillin

    topartsofAfrica(orrequirelessadaptiveresearchtorealizetheirspillinpotentials).

    Carefulanalysis

    of

    these

    types

    of

    technological

    distance

    metrics

    could

    substantially

    fine

    tune

    our

    strategicsenseoftechnologicalspillovers,withsignificantimplicationsforinternationalresearch

    collaborationsandtechnologytargetinginvolvingpublicorprivateagencies. Ofcourseotherfactors

    canhelporhindertherealizationoftheseresearchspilloverpotentials,suchasopennesstotrade(in

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    20/55

    technologies)includingphytosanitaryandbiosafetypolicies,intellectualpropertyrights,andarange

    ofmarketrealities.

    Disciplinary,AgencyandSectoralSpillovers

    LookingforwardforfoodandagriculturalR&D,asharpersenseofspatialspilloverpotentialswill

    becriticalformakingmoreinformedstrategicinvestmentandinstitutionalchoices. Butthereare

    otherdimensionsofspilloversthatarecriticalaswell. Onedimensionisthe(twoway)spillover

    betweenpubliclyandprivatelyperformedR&D. Thisspeakstotheappropriabilityaspectstouchedon

    brieflyabove,andareaffectedbythenatureandpracticeofintellectualpropertyrightsandthe

    industrialstructureoftheseinnovationmarkets. Another,oftenunderappreciated,aspectof

    spilloversinvolvesthetransferofideas,knowhow,andtangibleinnovationsamongdifferentsectors

    oftheeconomyanddifferentdisciplinesorfieldsofscientificinquirymoregenerallyconstrued. For

    example,innovationsinbiometrics,remotesensing,informatics,imagingtechnologies,plusthebasic

    biochemistry,molecularbiology,genomicsandproteomicsciencesareallpivotaltotechnicalprogress

    infoodandagriculture,butrarelyconstruedasfoodandagriculturalR&D.12 Forthisreason,

    Section3.2belowplacesfoodandagriculturalR&Dinvestmentsinthecontextofglobalpublicand

    privatespendingonallthesciences.

    2.5 EconomiesofScaleandScopeManytypesofresearchexhibitsignificanteconomiesofscaleorscope,sothatitmakessense

    toorganizerelativelylargeresearchinstitutions;butmuchagriculturaltechnologyischaracterizedby

    sitespecificity,relatedtoagroecologicalconditions,whichdefinesthesizeoftherelevantmarketina

    waythatismuchlesscommoninotherindustrialR&D(AlstonandPardey1999).13 Onewaytothink

    ofthisisintermsoftheunitcostsofmakinglocalresearchresultsapplicabletootherlocations(say,

    byadaptiveresearch),whichmustbeaddedtothelocalresearchcosts. Suchcostsgrowwiththesize

    12

    Foranevenmoreconcreteexample,considersignificantpartsofthesciencesupportingadvancesinprecisionagriculture(see,forexample,GebbersandAdamchuck2010)ortheagriboticsresearchunderwayattheDistributed

    RoboticsLaboratoryoftheMassachusettsInstituteofTechnology(Economist2009). Theamalgamofvisionsystems,laser

    sensors,satellitepositioningandinstrumentationtechnologiesbeingbroughttobearonautomatingcropharvestingand

    greenhouseproductionsystemswouldrarelyifeverbecountedasfoodandagriculturalR&D. Likewise,thereis(andhas

    longbeen)asignificantinterplaybetweenthehealthandagriculturalsciencesinamyriadofareasincludingepidemiology,

    basicmolecularbiology,nutritionsciences,andsoon.

    13ForadiscussionofthesescopeandscaleideasinthecontextofagriculturalR&DseePardey,RoseboomandAnderson

    (1991),ByerleeandTraxler(2001),andJin,Rozelle,Alston,andHuang(2005).

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    21/55

    ofthemarket. Consequently,whileeconomiesofscaleandscopeinresearchmeanthatunitcostsfall

    withsizeoftheR&Denterprise,theseeconomiesmustbetradedoffagainstthediseconomiesof

    distanceandadaptingsitespecificresults(thecostsoftransportingtheresearchresultsto

    economicallymoredistantlocations). Thus,asthesizeoftheresearchenterpriseincreases,unit

    costsarelikelytodeclineatfirst(becauseeconomiesofsizearerelativelyimportant)butwill

    eventuallyrise(asthecostsofeconomicdistancebecomeevermoreimportant).

    Inevaluatingtheneedforandinstitutionalarrangementsconcerninginternationallyconceived

    and,possibly,conductedagriculturalR&Ditisimportanttoconsidertheeconomiesofscaleandscope

    inknowledgeaccumulationanddissemination. Forinstance,iftechnologicalspilloverscontinuetobe

    fairlyavailableandaccessible,astheyhavebeeninthepast,itmightnotmakesenseforsmall,poor,

    agrariannationstospendtheirscarceintellectualandothercapitalresourcesinagriculturalscience.

    Howeverifspillinsfromdevelopedcountriesdecrease,developingcountrieswillneedtoconduct

    moreoftheirownresearch,butmanynationsmaybetoosmalltoachieveanefficientscaleinmany,

    ifany,oftheirR&Dpriorityareas. Forexample,40percentoftheagriculturalresearchagenciesin

    subSaharanAfricaemployedfewerthanfivefulltimeequivalentresearchersin2000;93percentof

    theregionsagriculturalR&Dagenciesemployedfewerthan50researchers. Creativeinstitutional

    innovationstocollectivefundandefficientlyconducttheresearchinwaysthatrealizethesescaleand

    scope

    economies

    will

    be

    crucial.

    2.6 ResearchTechnologyRegulationInmanypartsoftheworld,agriculturallyrelatedtechnologiesaresubjecttoanexpanding

    rangeofgovernmentregulation,withconsequencesforthenatureandamountofresearcheffortthat

    isnowrequiredtorespondtotheseregulatoryrequirements. Theseregulatoryregimescanhave

    substantialimplicationsonthepaceandnatureofinnovationandtechnologyreleaseanduptakein

    agriculture. Withoutdoubttheyaddtothecostofdevelopinganddeliveringtechnologiestofarmers.

    However,comparativelylittle(economic)attentionhasbeengiventostreamliningtheseregulatory

    regimes,strivingtomaximizethesocialpayoffstothecostsandcomplianceeffort(onthepartof

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    22/55

    technologydevelopers,suppliers,aswellasfarmers)thattheyincur.14 TheUnitedStateshasarguably

    donethemostinthisregard(althoughthetechnicalandadministrativecostsofcomplianceinthe

    UnitedStatesarehighandrisingandneedcontinuedvigilance),optingforsciencebasedapproval

    approachesthatfacilitateinnovationandtechnicalchangewhileseekingtoobjectivelyassessand

    managethehumanandenvironmentalrisksassociatedwiththosechanges. However,manypartsof

    thedevelopingworldstillhaveinefficientordysfunctionaltechnologyassessment,releaseand

    oversightsystems,whetherthatisinreferencetomodernbiotechnologiesorlesscontentious

    technologieslikeconventionallybredcropvarieties. Thereareamyriadofreasonsforthese

    institutionalfailures,butonekeyaspectisalackoflocaltechnicalexpertisetoconductorevaluate

    thenecessaryprereleasetrialsandstewardthetechnologiesoncetheyareinuse. AsPardeyand

    Alston(2010)pointedout,loweringthecostsofaccesstothenecessarytechnical(oftenresearch

    informed)informationwouldlikelyplayakeyroleinspurringlocalinnovationindevelopingcountries

    andfacilitatethetransferinandadaptationoftechnologiesdevelopedelsewhere.

    3. R&DandProductivity15Growthindemandforagriculturalcommoditieslargelystemsfromgrowthindemandforfood,whichis

    driven by growth in population and per capita incomes (especially the economic growth of the fast

    growingeconomiesofAsia),coupledwithnewdemands forbiofuels. Growth insupplyofagricultural

    commoditiesisprimarilydrivenbygrowthinproductivity,especiallyasgrowthintheavailabilityofland

    and water resources for agriculture has become more constrained. Productivity improvements in

    agriculture are strongly associated with lagged R&D spending, as revealed in a large compilation of

    countryspecific studies reported in Alston et al. (2000). Thus, the rate of growth of investments in

    agriculturalR&Dandtheusestowhichthoseresearchdollarsareputwillbeapivotaldeterminantoflong

    termgrowthinthesupply,availability,andpriceoffoodoverthecomingdecades.

    3.1

    Global

    Productivity

    Patterns

    14See,forexample,Frisvold,HurleyandMitchell(2009),andthearticlestheyintroduce,forasuiteof(economic)analyses

    ofanotableandregulatedbioengineeredtechnology;specificallyherbicideresistantcrops. SeealsoJust,Alstonand

    Zilberman(2006).15ThissectiondrawsonAlston,BeddowandPardey(2009)andAlston,PardeyandBeddow(2010),whoprovideadditional

    informationbeyondthehighlightsincludedhere.

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    23/55

    Conventionalmeasuresofproductivitymeasurethequantityofoutputrelativetothequantity

    ofinputs. Ifoutputgrowsatthesamepaceasinputs,thenproductivityisunchanged:iftherateof

    growthinoutputexceedstherateofgrowthintheuseofinputs,thenproductivitygrowthispositive.

    Partialfactorproductivitymeasuresexpressoutputrelativetoaparticularinput(likelandorlabor).16

    Multifactorproductivitymeasuresexpressoutputrelativetoamoreinclusivemetricofallmeasurable

    inputs(includingland,laborandcapital,aswellasenergy,chemicals,andotherpurchasedinputs).

    Measuresofagriculturalproductivitygrowthbetheycropyields,otherpartialfactorproductivity

    measures(forexample,measuresoflandandlaborproductivity),orindexesofmultifactor

    productivityshowgenerallyconsistentpatternsintermsofsecularshifts,includingindicationsofa

    recentslowdowningrowth.

    CropYields

    Thelongresearchlagsandinherentlyspatialnatureofagriculturalproductionmeansthereis

    valueintakinganexplicitlylongtermandgeospatialperspectiveoncropyields. Figure7plotsthe

    distributionofaveragenationalcropyieldsworldwideformaize,wheatandriceforselectedperiods

    beginninginthemid1800s. Thereareseveralstrikingfeaturesofthesecropyielddistributions. The

    rightwardmovementinthemodeofthedistribution(andimplicitlytheaverageaswell)isconsistent

    withanincreaseinaveragecropyieldsworldwide. However,thepaceandtimingofthatrightward

    shift

    occurred

    at

    different

    times

    and

    at

    different

    rates

    among

    the

    different

    crops,

    but

    notably,

    as

    the

    centerofgravityofeachdistributionshiftedtotherightthevariancearoundthatcenterofgravity

    alsoincreasedinallthreecases. Thusasglobalmeanyieldsgrewovertime,thevariationofyields

    amongcountriesalsobecamemorepronounced.

    Figure8givesamappedsense(atroughlya10kmby10kmpixelresolution)ofthespatial

    variationincropyieldsforthesethreecrops(plussoybeans)in2000. Thelightertheshadingthe

    lowerthecropyieldsrelativetothehighestyieldingpixels(indicatedbydarkblue). While37percent

    oftheworldsmaizeproductioncomesfromthe20percentofcroppedmaizeareareportingthe

    highestyields,only24percentoftheworldssoybeanproductioncomesfromthehighestyielding

    areasforthatcrop. However,therewassubstantiallylessspatialvariationinsoybeanyieldsthanin

    16Cropyieldsrepresentaparticularpartialproductivitymeasurewhereinthephysicaloutputforaparticularcropis

    expressedrelativetolandinput.

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    24/55

    cornyields. Theratioofaverageyieldsinthe20percentofareasowntosoybeansreportingthe

    highestyieldswas1.76timesgreaterthantheyieldsinthecorresponding20percentofarea

    reportingthelowestyields. Formaizetheaverageyieldratiobetweenthehighestandlowestyielding

    areaswas4.41.

    GlobalannualaverageratesofyieldgrowtharereportedinTable2,whichincludesseparate

    estimatesforhigh,middle,andlowincomecountriesandtheworldasawhole,fortwosubperiods:

    19611990and19902007. Thereisaslowdownevidentfortheglobalaverage,althoughbeginning

    fromcomparativelylowyields,lowincomecountrieshadincreasingratesofgrowthinwheatandrice

    yieldssince1990. Thuslowincomecountriesgainedsomegroundsince1990,howevertherebound

    inyieldgrowthinthispartoftheworldfailedtofullymakeupforthecomparativelylowgrowthrates

    theyexperiencedin19611990. Consequently,significantyieldgapspersists,andasAlston,Pardey

    andBeddow(2010)report,thelowincomecountryversusworldrelativitiesofaveragemaize,wheat,

    andriceyieldsin2007havefallenbelowthecorresponding1961relativities. Lowincomecountries

    hadaveragesoybeanyieldsthatwereabout50percentoftheworldaveragein1961,andthatsame

    gappersistedthroughto2007.

    Forallfourcommodities,inbothhigh andmiddleincomecountriescollectivelyaccounting

    forbetween78.8and99.4percentofglobalproductionofthesecropsin2007

    averageannualrates

    ofyieldgrowthwerelowerin19902007thanin19611990. Thegrowthofwheatyieldsslowedthe

    mostand,forthehighincomecountriesasagroup,wheatyieldsbarelychangedover19902007.

    Globalmaizeyieldsgrewatanaveragerateof1.77percentperyearduring19902007comparedwith

    2.20percentperyearfor19611990. Likewisericeyieldsgrewatlessthan1.0percentperyearduring

    19902007,lessthanhalftheiraveragegrowthratefor19601990. Moreover,theslowdownincrop

    yieldsisquitepervasive. Inmorethanhalfofthecountriesthatgrewthesecrops,yieldsforrice,

    wheat,maize,andsoybeansgrewmoreslowlyduring19902007thanduring19611990(Table3).

    Morecritically,theslowdownwasgenerallymorewidespreadthanamongthetoptenproducing

    countriesworldwide.

    Theslowdownisalsopervasiveandevenmorepronouncedwhencountriesareaggregatedin

    termsofharvestedarea. Lookingattheperiodafter1961,thegrowthinyieldsofwheat,rice,and

    soybeansslowedafter1990incountriesaccountingformorethan70percentoftheworlds

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    25/55

    harvestedarea;forcornaround65percentofharvestedareawasincountrieswithsloweryield

    growthafter1990. LatinAmericaistheonlycontinentwherecountriesaccountingformorethanhalf

    theharvestedareaforallfourcropshadyieldsgrowingatmorerapidratesafter1990thanbefore.

    Notably,countriesaccountingformorethan90percentoftheharvestedareaamongthehighincome

    countriessawthepaceofgrowthofmaizeandriceyieldsslowafter1990,whileallofthehighincome

    countrieshadwheatandsoybeanyieldsgrowingataslowerrateinthemorerecentperiod.

    LandandLaborProductivity

    Movingbeyondcropyieldstomorebroadlyconstruedproductivitymeasures,global

    productivitytrendsshowa2.4foldincreaseinaggregateoutputperharvestedareasince1961,

    equivalenttoannualaveragegrowthof2.0percentperyear. Accompanyingthisincreaseinland

    productivitywasa1.7foldincrease,or1.2percentperyeargrowth,inaggregateoutputper

    agriculturalworker(Table4). Theseproductivitydevelopmentsreflectglobalagriculturaloutput

    growingrelativelyquicklycomparedwiththegrowthintheuseofagriculturallandandlabor0.3

    percentand1.1percentperyear,respectively.

    Inparallelwiththeglobalcropyieldevidencepresentedabove,thelongerrungrowthinland

    andlaborproductivitymasksawidespreadalbeitnotuniversalslowdownintherateofgrowthof

    bothproductivitymeasuresduring19902005comparedwiththepreviousthreedecades. Chinaand

    Latin

    America

    are

    significant

    exceptions,

    both

    having

    considerably

    higher

    growth

    rates

    of

    land

    and

    laborproductivitysince1990. Amongthetop20producingcountriesaccordingtotheir2005valueof

    agriculturaloutput,landandlaborproductivitygrowthwassubstantiallyslowerin19902005thanin

    19611990oncethelarge,andinmanyrespectsexceptional,caseofChinaissettooneside. After

    settingasidethetop20producingcountries,onaverageacrosstherestoftheworld,theslowdownis

    evenmorepronounced:forthisgroupofcountries;landproductivitygrewby1.83percentperyear

    duringtheperiod19611990,butbyonly0.88percentperyearthereafter;laborproductivitygrewby

    1.08percent

    per

    year

    prior

    to

    1990,

    but

    barely

    budged

    during

    the

    period

    1990

    2005.

    After1990,theglobalgrowthrateoflandproductivityslowedfrom2.03percentperyearto

    1.82percentperyear,whereasthegrowthrateoflaborproductivityincreasedfrom1.12percentper

    yearfor19611990to1.36percentperyearfor19902005. Onceagaintheseworldtotalsare

    distortedbythesignificantandexceptionalcaseofChina. NettingoutChina,globallandandlabor

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    26/55

    productivitygrowthhasbeenslowersince1990thanduringthepriorthreedecades. Thesame

    periodrelativitiesprevailiftheformerSovietUnion(FSU)isalsonettedout,althoughthemagnitude

    oftheglobalproductivityslowdownnetofChinaandtheFSUislesspronouncedbecausebothpartial

    productivitymeasuresfortheFSUactuallyshrankafter1990.

    Insummarizingtheexistingevidenceonpartialandmultifactorproductivitytrendsin

    agricultureworldwide,Alston,BabcockandPardey(2010)concludethat.eventhoughwehave

    manyreasonsforbeingcautiousinthisarea,wefinditdifficulttoreachanyconclusionotherthan

    thatweareseeingevidenceofaslowdowninglobalagriculturalproductivitygrowth,especiallyinthe

    worldsrichestcountries. Comingtoaconsensusonthestructureandextentofaproductivity

    slowdownisdifficult,buthelpful. Drawingpolicyimplicationsfromthisevidenceisdoublydifficult.

    Alston,BabcockandPardeywentontoobservethattheAustralian[productivity]slowdownhas

    beenobservedduringthemostsevereandextendeddroughtinthatcountryshistory. Other

    countries,too,mayhavebeenaffectedbyarunofunusuallyfavorableorunfavorableseasons. Andit

    ishardalsototellthedifferencebetweensustainedchangesingrowthandthemultiyeareffectsofa

    changethatisreallyepisodicinnature(e.g.,themassiveinstitutionalreformsinChinaandtheformer

    SovietUnion). Notwithstandingtheproblemsofproductivitymeasurementandinterpretation,the

    apparentandapparentlypervasiveslowdowndoesraisequestionsastowhetherthecurrentglobal

    investment

    in

    agricultural

    R&D

    will

    be

    sufficient

    to

    enable

    the

    development

    of

    innovations

    and

    productivitysuchthatagriculturalsupplywillgrowfastenoughtokeeppacewiththeinevitable

    growthindemand. ItistotheR&Dinvestmentevidencethatwenowturn.

    3.2 R&DPatterns17

    In2000,globalinvestmentinfoodandagriculturalR&Dtotalled$36.2billion(2005prices).18

    Around67percentoftheresearchwasperformedbypublicagencies,andtheremaining33percent

    byfirmsinthefood(processing,transport,andstorage),beverage,chemical,andmachinerysectors

    17TheresearchanddevelopmentestimatesreportedheredrawinpartfromestimatesmadebyDehmerandPardey

    (2010)andPardeyandChanKang(2010)thatarestillconsideredpreliminary. TheyexcludetheFormerSovietUnionand

    EasternEuropeancountriesduetolackofdata.

    18 Year2000isthelastyearforwhichinternationallycomparabledataonagriculturalR&Dinvestmentsarepresently

    available. Thesedatawereconvertedtointernationaldollarsusingpurchasingpowerparity(PPP)indexes. UsingPPPsto

    convertlocalcurrenciestoanumerairecurrencyresultsinsignificantlylargersharesoftheglobalresearchtotalbeing

    attributedtolowerincomecountriesthanifmarketexchangerateswereusedforthecurrencyconversion.

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    27/55

    servicingfoodandagriculture. Figure9,Panelabreaksdownthepublicplusprivatefoodand

    agriculturalR&Dspendingaccordingtothehighincomeandlow andmiddleincomecountrieswhere

    thisresearchisperformed. Almost70percentofthatpublicandprivateresearchtookplaceinhigh

    incomecountries,andaroundhalftherichcountryresearchwasconductedbyprivatefirms. In

    contrast,foodandagriculturalresearchconductedinlowandmiddleincomecountrieswas

    overwhelminglycarriedoutbypublicagencies(privatefirmsaccountedforjustover6percentofthe

    estimated$10.8billionspentonfoodandagriculturalR&Dinthesecountries).

    PublicspendingonagriculturalR&Dishighlyconcentrated,withthetopfivepercentofcountries

    inthedataset(i.e.,6countriesinatotalof129)accountingforapproximatelyhalfofthespending. The

    UnitedStatesaloneconstitutedaround16percentofglobalspendingonpubliclypreformedagricultural

    research. TheAsiaandPacificregionhascontinuedtogainground,accountingforaneverlargershare

    oftheworldanddevelopingcountrytotalsince1981(20.3percentoftheworldtotalin2000,upfrom

    12.5percentin1981). In2000,justtwocountriesfromthisregion,ChinaandIndia,accountedfor29.1

    percentofallexpenditureonpublicagriculturalR&Dbydevelopingcountries(andmorethan14

    percentofpublicagriculturalR&Dglobally),asubstantialincreasefromtheir15.6percentcombined

    sharein1981. Instarkcontrast,subSaharanAfricacontinuedtolosegrounditssharefellfrom17.9

    percentofthetotalinvestmentinpublicagriculturalR&Dbydevelopingcountriesin1981to12.2

    percent

    in

    2000.

    Private

    spending

    is

    also

    geographically

    concentrated

    with

    around

    72

    percent

    of

    the

    worldsprivatefoodandagriculturalR&Dconductedinjust5countries.

    ThesignificantinterdisciplinaryandcrosssectoralspilloversbetweenfoodandagriculturalR&D

    andresearchdonebyothersciencesandinothersectorsindicatesthatameaningfulappreciationofthe

    sourcesofinnovationinfoodandagriculturemustbecognizantofthemagnitudeandchangingnature

    oftotalinvestmentsinR&D. Figure9,Panelb,showsthatin2000,foodandagriculturallyorientedR&D

    accountedforonly5percentoftheestimated$782.7billioninvestedinallformsofR&Dworldwide

    (increasingto

    $970.6

    billion

    in2006).

    Collectively,

    the

    high

    income

    countries

    (whose

    average

    per

    capita

    incomesexceeded$11,906)accountedfor85percentoftheworldsR&Dspendingin2000(80percent

    in2006). Thedevelopingcountryshareoftheworldtotalhasgrownovertimefrom5percentin1980

    to15percentin2006(DehmerandPardey2010). Notably,China,IndiaandBrazilaccountforagrowing

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    28/55

    andnowdominantshareofthisdevelopingcountrytotal61percentofthedevelopingworldstotal

    R&Dspendingin1980,increasingto83percentin2006.

    ThedynamicsbetweenfoodandagriculturalR&Dandsciencespendinggenerallyarelikelyto

    continuechanginginfutureyears,mostnotablyforthoselow andmiddleincomecountrieswith

    growingsciencesectors. Figure10showsthatforthepastseveraldecadesatleast,spendingonfood

    andagriculturalR&Dinhighincomecountrieshasbeenlessthan5percentoftotalsciencespending.

    Onaverage,researchdirectedtowardfoodandagriculturalR&Dinthelow andmiddleincome

    countrieswasaround20percentofthetotal(publicandprivate)researchconductedinthatpartofthe

    worldduringthe1980s,butbythemid1990sthatsharestartedtodeclineandnowaveragesnearer10

    percent.

    Therecontinuestobeahugegapbetweenrichandpoorcountriesintermsoftheintensitywith

    whichtheyinvestinfoodandagriculturalR&D. Figure11,Panela,showsthatthepublicagricultural

    researchintensity(ARI)forlow andmiddleincomecountriesbarelybudgedduringthe1980sand

    1990sandwaslessthanhalfthecorrespondingrichcountryfigureduringthisperiod. Moreover,the

    intensitywithwhichhighincomecountriesinvestinfoodandagriculturalR&Dhastrendedupwards

    sincethe1970s;andaveraged$2.95ofR&Dspendingforevery$100ofagriculturalGDPduringthe

    period20002007. Theintensitygapbetweenricherandpoorercountriesisevenmorepronouncedin

    terms

    of

    public

    plus

    private

    spending

    (Figure

    11,

    Panel

    b).

    Onaverage,theprivateshareoftotalfoodandagriculturalR&Dinrichcountieshastrended

    upwardsfromaround36percentintheearly1970sto50percentin2007(Figure12,Panela). About60

    percentofthisresearchrelatestofoodprocessingandbeverageproducts,ratherthanchemical,

    biologicalandmachineryrelatedR&Dthathelpsspurfarmproductivity. Infact,researchintendedto

    maintainorenhancefarmproductivityhasbeenagenerallydecliningshareofpubliclyperformedR&D

    intheUnitedStates(wheredatawereavailabletoassessthistrend)(Figure11,Panelb). By2006,less

    than57

    percent

    of

    all

    R&D

    conducted

    by

    the

    state

    agricultural

    experiment

    stations

    had

    afarm

    productivityorientation. IndicationsarethatthisU.S.trendmirrorsdevelopmentsinotherhighincome

    countries.

    Notonlyhasrichcountryresearchshiftedawayfromproductivityorientedendeavors,the

    overallrateofgrowthofreal(i.e.,inflationadjusted)spendinghassloweddramatically;fromaround3

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    29/55

    percentperyearduringthe1970stobarely1peryearforthepastseveraldecades(Figure13). While

    therateofgrowthofspendinginlow andmiddleincomecountriesishigher,ittoohassuccessively

    slowed,atleastuntiltheendofthe1990s. Ifthesespendingtrendspersist,itraisesrealquestionsasto

    whetherthegrowthinagriculturalproductivityrequiredtosustainablymeetbasicfoodrequirementsin

    thedecadesaheadwillberealized.

    4. TheWayForwardLinkingGlobalR&DtoNationalNeeds19Thedemandfor(public)internationalagricultureresearch(IAR)continuestobestrong. Moreover,as

    thecostsofinternationalcollaborationdecline(astravelandcommunicationscostsfall)andscaleand

    scopeeconomiesbecomemoreprominent,supplysidedevelopmentswillcontinuetopushformore

    notlessIAR. TheroleandcontributionsofIARtodevelopingcountryagriculturewillvarysignificantly

    amongcountriesaccordingtotheirrespectivestageofdevelopmentandthesize,structureand

    sophisticationoftheirnationalsciencecapacities. Forcountriesatthelowendofthestructural

    transformationprocess,mostlycountriesinsubSaharanAfrica,thetraditionalfocusonfoodstaples

    willcontinuetobeespeciallyimportant. Broadbasedproductivitygainsinstaplecropscanhavefar

    reachingimpactsontheruralpoor(BinswangerandMcCalla2010). Thetaskcontinuestobedaunting

    giventheheterogeneityofcropsandproductionenvironments,substantialexposuretoclimaterisks

    (whichmaygetworse),historicallyandcontinuinglylowlevelsofinvestmentininfrastructureand

    agricultureresearchcapacity,andapoorenablingenvironmentforenhancingproductivitygrowth.

    Foremergingeconomies,ontheotherhand,IARcouldcapitalizeonthegrowingstrengthofnational

    publicinstitutionsandprivatefirmsthatinvestintechnologygenerationanddeliveryandfocusits

    effortsinareaswhereitcanprovideuniqueinternationalpublicgoods. Inthecaseoffavorable

    productionenvironments,prebreedingmaterialsforshiftingyieldfrontiersforthemajorstaples,

    managingtransboundarypests,andsustainingintensiveproductionsystems,aresomeoftheareas

    whereinternationalagricultureR&Dcouldcontinuetobeanimportantandcosteffectiveoption.

    Focusedresearchonstressproneenvironments(forexample,droughtandhightemperature)may

    alsohaveimportantinternationalresearchcomponents..

    ThesupplyofpubliclyprovidedIARtodevelopingcountryresearchprogramsishowever,

    becomingincreasinglyconstrainedbyvariabledonorsupport,agrowingdisconnectwithprivate

    19ThissectiondrawsonPingali(2009).

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    30/55

    sectorpriorities,apushtowardsdownstreamproductadaptationanddisseminationactivitiesrelative

    toinnovationandproductdevelopment,andalackofclearlinksbetweeninternationalpublicgood

    researchandnationalagriculturedevelopmentpriorities. Countryleveldonorcoordinationand

    alignmentmechanisms,asspecifiedintheParisDeclarationonAidEffectiveness,donotexplicitly

    accountfortheroleofinternationalagricultureresearchinthedevelopmentprocess. Thissection

    presentssomeoptionsforrebuildingsynergiesbetweeninternationalpublicgoodresearchand

    nationalagriculturedevelopmentpriorities.

    FindingSynergiesbetweenPublicandPrivateR&D

    Asdocumentedabove,privatesectorinvestmentinagricultureR&Dhasincreasedinrichcountries

    andfortheincreasinglymarketorientedpartsoftheproductionsystemsinemergingeconomies.

    Largemultinationalcorporationspartneringwithnationalagribusinessfirmsarebecomingaviable

    alternativetopublicsectortechnologydelivery,mostnotablyinthecaseofhighvalueagriculture

    (cotton,vegetables,andlivestock),hybridsofstaplecropssuchasmaize,andpestanddisease

    managementandmachinerytechnologies. Postfarmprocessingtechnologiesarealsolargelyinthe

    privaterealmandmakinginroadsinselectedemergingmarkets. Theabilitytocapturetherentsfrom

    agricultureR&Dinvestments,throughtheuseofintellectualpropertyrightsandothermeanshas

    increasedtheprivatepresenceininnovationintensivemarketsrelatedtofoodandagriculture,but

    only

    in

    certain

    segments

    of

    those

    markets

    and

    with

    an

    emphasis

    on

    certain

    countries

    (Pingali

    and

    Traxler2002;PardeyandAlston2010). Thisexpandedprivatepresencehasoccurredatthesame

    timeasgrowthinpublicR&Dspendinghasstalledorstumbled,shiftingtheoveralltrajectoryof

    innovationinfoodandagriculturefurtherinthedirectionofcommercialfarmerswithsignificant

    productivitygrowthpotentialandincreasinglyintegratingproductionagriculturewiththerapidshifts

    inpostfarmfoodprocessingandmarketingoperations.20 Thispresentsadilemmaforthenational

    andinternationalnonprofitsectorshouldtheyusetheprivatesectortoleveragetheirown

    investmentsinbreadbasket

    areas

    or

    should

    they

    redeploy

    resources

    to

    protect

    poor

    farmers

    growingorphancropsinmarginalareasfromdeterioratingtermsoftrade? Whileoneperspectiveis

    thatpublicresearchshouldemphasizetheinterestsofmarginalfamers,therecentcrisishasbrought

    20 Notablehereistherapidriseofsupermarketsinmanydevelopingcounties,aswellastheincreaseddemandforfood

    consumedawayfromthehomeandconveniencefoodsconsumedinthehomeaspercapitaincomesriseforcertain

    segmentsofcertainmarkets,particularlyinAsiaandLatinAmerica.

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    31/55

    renewedappreciationoftheeffectivenessofR&Dasaninstrumenttomoderateupwardpressureon

    staplefoodcommodityprices. Thismayencouragegovernmentstoreorientresourcestoward

    productive,commercialfarmerswiththegreatestpotentialimpactonmoderatingfoodprices

    (althoughtheevidencefrompastspikesinglobalfoodpricesisthatthesetypesofresponseshave

    beenshortlived).

    Anenablingpolicyenvironmentthatincludesappropriateintellectualpropertyprotection,

    reducedtradebarriers,andatransparentbiosafetyprocedurewillleadtofurtherprivateresearch

    investmentsforcommercialproductionsystemsintheemergingeconomies. However,many

    developingcountries,especiallyinsubSaharanAfrica,remainoutsidetheorbitofprivatesector

    interests. Theprivatesectorisalsounlikelytoinvestmuchinresearchfortraditionalcropsgrowingin

    especiallydifficultenvironments,suchasdroughtprevalentorhightemperatureenvironments,even

    intransformingeconomies.21 Theprivatesectorsrecordindeliveringnaturalresourcemanagement

    (NRM)technologiesisalsolimited,eveninadvancedcountryagriculture. Publicresearchinvestments

    couldbejudiciouslyandcreativelydeployedtoleverageprivatetechnologydevelopmentanddelivery

    capacitiestohelpmeettheneedsofthepoor(FAO2004).

    ChangingAidArchitecture

    Thenatureofoverallaidsupplytodevelopingcountrieshasbeenchangingdramaticallyoverthepast

    decade

    in

    terms

    of

    the

    quantities

    provided,

    the

    plurality

    of

    funding

    sources,

    and

    donor

    coordination

    andalignmentmechanisms. ThesechangeshavesignificantimplicationsforthewayIARisconducted

    andtransferredtodevelopingcountries. ArecentOverseasDevelopmentInstitute(ODI)report

    indicatesthattotalaidvolumeshaverisenfromaround$60billionperyearinthe1990stoaround

    $100billionin2005andareanticipatedtoriseto$130billionby2010(BurallandMaxwell2006).

    AveragedacrossOECDcountries,overseasdevelopmentassistance(ODA)asapercentageofgross

    nationalincomehasrisenbackto0.33in2006afterhavingdroppedtoalowof0.22in1997

    (OECD/DAC2006).

    New

    donor

    countries,

    such

    as

    China,

    India,

    Korea,

    as

    well

    as

    private

    foundations

    (suchastheGatesFoundation)andmultilateralfunds(GEF),haveaddedtotheoverallaidtotals.

    21Thisiscertainlynottoarguethattheprivatesectorwillnecessarilyignoresuchresearch. Witness,forexample,the

    partnershipbetweenCIMMYT,MonsantoCorporationandmanyotherstodevelopwaterefficientmaizevarietiesfor

    Africansmallholderfarmers(seewww.monsanto.com/droughttolerantcorn/WEMA.asp).

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    32/55

    Growthinthevolumeofaid,thenumberofdonors,andmultiplicityofagendashasspurredcallsfor

    greatercoordinationandalignmentofdonorsupportatthecountrylevel.

    TheParisDeclarationonAidEffectiveness,sponsoredbytheDevelopmentAssistance

    Committee(DAC)oftheOECD,hasbeenasignificantstepinthedirectionofenhancingdonor

    coordination. Donorswhosignedontothedeclarationagreedtofollowgovernmentplansand

    priorities(alignment)andtoworktogetherinthatprocess(harmonization). TheParisDeclaration

    emphasizesbudgetsupporttopriorityprogramsatthecountrylevelratherthansupportfordiscrete

    projectsthatmayormaynotbepartofthegovernmentplansandpriorities.

    Dotheaboveeffortscontributetothepromotionoftransnationalpublicgoodresearchand

    strengthentheR&Dpipelineforfarmlevelimpact? Thereareseveralreasonstobeconcerned. First,

    therearenoobviousmechanismsfornationalplansandprioritiestoberesponsivetoemergingglobal

    agricultureR&Dopportunities. Second,nationalprioritiestendtofocusondownstream,highly

    adaptiveactivities,ratherthaninternationalpublicgoodresearch. Third,scaleeconomiesin

    technologygenerationmaybelostifcountriesembarkonunnecessarilyduplicativeeffortsaround

    similarproblems. Fourth,theCGIARitselfhasmovedmoredownstream(playingadevelopmentrole)

    inseveralcountriesinresponsetodonorsupportforcountryspecificactivities,weakeningits

    traditionalroleasasourceofinternationalR&Dspillovers. Finally,currentparalleleffortstowards

    increased

    harmonization

    of

    IAR

    (including

    the

    CGIAR

    reform)

    do

    not

    take

    into

    account

    donor

    efforts

    to

    alignwithandsupportnationalplansandpriorities. So,whilethemovementtowardsnational

    ownershipofdevelopmentagendasanddonoralignmentaroundthemisunquestionablygood,an

    unintendedconsequencecouldbeadisruptionintheR&Dpipelinethatsuppliespublicgoodresearch

    andtechnologiesforenhancingproductivitygrowthindevelopingcountryagriculture. Thelonglags

    inherentinmovingfromR&Dinputstothetechnologiestakenupbyfarmersmakestheebbandflow

    (andfaddishness)ofdonorfundingespeciallyproblematic.

    Technology

    Demand

    AssessmentBeyond

    Farmers

    Voices

    Muchofthediscussiononassessingtechnologydemandandpreferenceshasfocusedatthe

    communitylevelusingavarietyofparticipatorymethods(seePingali,RozelleandGerpacio2001;

    McIntyreetal.2009). Farmerassociationshavealsobeeninvolvedinmakingdecisionsonthe

    allocationofresearchfunds,asintheYaquiValleyofMexico. Elicitingfarmervoiceinprioritysetting

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    33/55

    isimportantatthemoreappliedandadaptiveendoftheresearchpipelineandforelicitinglocal

    preferencesontechnologydesign. However,aggregatingacrossfarmerpreferencesandeliciting

    informationonstrategic,longertermresearchprioritiesischallengingifthereisanexclusivereliance

    onparticipatorymethods.

    Thereareotheremergingtoolsandapproachesthatcanhelpassesstechnologydemandat

    thenationalandregionallevelandbeaffectivelyusedbytheinternationalagricultureresearch

    communityforsettingprioritiesandtargetingitsglobalpublicgoodresearchefforts. TheWorld

    BanksLivingStandardsMeasurement(LSMS)groupisembarkingonamassivehouseholdpanel

    surveyacrosssubSaharanAfrica,withafocusonruralhouseholds. Thisnationallyrepresentative

    householdsurveywillprovideawealthofinformationonthestateofAfricanfarmingsystems,

    technologyuse,andconstraintstoenhancingproductivitygrowth. TheLSMSdatacanbeinvaluable

    ingeneratinganalysisanddiscussionsonnationallevelresearchprioritiesandtechnologydemands.

    SincetheLSMSsurveysarestandardizedacrosscountries,aggregationatregionallevelsisalso

    possible,hencetheabilitytoderivetransnationalresearchdemands. TheHarvestChoicedata

    platformbeingjointlydevelopedbyIFPRIandtheUniversityofMinnesotaandawholeraftof

    collaboratorsprovidesspatiallydisaggregateddataonavarietyofvariablesthatareimportantfor

    assessingtechnologydemand.22 Agroclimatic,biophysicalandsocioeconomicdatacanbeoverlaidto

    identify

    priority

    constraints

    at

    the

    sub

    national,

    national

    and

    regional

    levels,

    and

    to

    target

    technology

    diffusionappropriately. TheHarvestChoiceplatformallowsforanexanteassessmentofpotential

    technologyinterventionsatthenationalandsubnationallevels. InformationfromtheHarvestChoice

    analysiscanbeusedforanexanteassessmentofpotentialtechnologiesatthegloballevelandover

    timebyusingIFPRIsIMPACTmodelwhichisalsobeingrevampedtobetterservethisrole. The

    challengeliesinincorporatingtheseimprovedanalyticaltoolsintotheshiftingpoliticaleconomies

    thatshape(strategic)prioritiesforinternationalagriculturalresearch.

    Improving

    the

    Links

    between

    International

    R&D

    and

    National

    Strategies

    Thechallengeforthenewaidarchitectureistocreatemechanismsthatimprovethelinksbetween

    internationalR&Dandnationalagriculturedevelopmentstrategies. Evenwithinacountry,the

    processforidentifyingtechnologyneedsandprioritizingthemforbudgetsupportisdifficultand

    22Formoredetailsseewww.HarvestChoice.org.

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    34/55

    uncertain,andR&Dcontinuestobeundervaluedinnationalstrategiesanddonorpriorities. More

    thantwodecadesago,VernonRuttanbroachedtheideaofformingNationalResearchSupport

    Groupstohelpassessandprioritizeresearchdemandsandchampiontheirsupplyatthenational

    level(Ruttan1987). Thesesupportgroupscouldalsobeaconduitforbetterlinkingnationaland

    internationalR&D. DataandanalysisgeneratedthroughtheLSMS,HarvestChoiceandother

    initiativesdiscussedabovecouldstrengthentheabilityofnationalresearchgroupstoidentifypriority

    problemsandtoidentifypotentialsolutionsontheglobalR&Dpipeline,andcoordinatetheir

    adaptationanddiffusionatthenationallevel. Finally,theresearchsupportgroupscouldachievea

    regionalandcontinentalvoicebyworkingcollectivelyinregionalgroupingssuchastheSouthern

    AfricanDevelopmentCommunity(SADC)orECOWAS,theEconomicCommunityforWestAfrican

    StatesandwithglobalalliancessuchasGFAR.

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    35/55

    References

    Alston,J.M.,B.BabcockandP.G.Pardey.Summary,Synthesis,andConclusion.Chapter15inJ.M.Alston,B.

    BabcockandP.G.Pardey,eds.TheShiftingPatternsofAgriculturalProductionandProductivityWorldwide,

    CARDMATRICElectronicBook,Ames:IowaStateUniversity,2010(inpress).

    Alston,J.M.,P.G.PardeyandJ.M.Beddow.GlobalPatternsofCropYieldsandOtherPartialProductivityMeasuresandPrices.Chapter3inJ.M.Alston,B.BabcockandP.G.Pardey,eds.TheShiftingPatternsof

    AgriculturalProductionandProductivityWorldwide,CARDMATRICElectronicBook,Ames:IowaState

    University,2010(inpress).

    Alston,J.M.,M.A.Andersen,J.S.James,andP.G.Pardey.PersistencePays:U.S.AgriculturalProductivityGrowth

    andtheBenefitsfromPublicR&DSpending.NewYork:Springer,2010.

    Alston,J.M.,J.M.Beddow,andP.G.Pardey.AgriculturalResearch,Productivity,andFoodPricesintheLong

    Run.Science325(4)(September2009):12091210.

    Alston,J.M.,J.M.BeddowandP.G.Pardey.MendelversusMalthus:ResearchProductivityandFoodPricesin

    theLongRun.DepartmentofAppliedEconomicsStaffPaperNo.P0901,StPaul,UniversityofMinnesota,

    January2009(revisedSeptember2009).

    Alston,J.M.,S.DehmerandP.G.Pardey.InternationalInitiativesinAgriculturalR&D:TheChangingFortunesof

    theCGIAR.Chapter12inP.G.Pardey,J.M.Alston,andR.R.Piggott,eds.AgriculturalR&DintheDeveloping

    World:TooLittle,TooLate?WashingtonD.C.:InternationalFoodPolicyResearchInstitute,2006.

    Alston,J.M.,M.C.Marra,P.G.Pardey,andT.J.Wyatt.AMetaAnalysisofRatesofReturntoAgriculturalR&D:Ex

    PedeHerculem?IFPRIResearchReportNo113.WashingtonD.C.:InternationalFoodPolicyResearchInstitute,

    2000.

    Beddow,J.M.AnEconomicPerspectiveontheSpatialDynamicsofU.S.CropProductionandYields.Department

    ofAppliedEconomics,UnpublishedPhDDissertation.St.Paul:UniversityofMinnesota,2010(inpreparation).

    Beintema,N.M.andGJStads.MeasuringAgriculturalResearchInvestments:ARevisedGlobalPicture.ASTI

    BackgroundNote.WashingtonD.C.:InternationalFoodPolicyResearchInstitute,October2008.

    Binswanger,H.P.,andA.McCalla.TheChangingContextandProspectsforAgricultureandRuralDevelopment

    inAfrica.Chapter70inP.L.PingaliandR.E.Evenson,eds.HandbookofAgriculturalEconomics,VolumeIV.

    Amsterdam:ElsevierPress,2010.

    Burall,S.andS.Maxwell.ReformingtheInternationalAidArchitecture:OptionsandtheWayForward.

    WorkingPaper278.London:ODI,2006.

    Byerlee,D.,andG.Traxler.TheRoleofTechnologySpilloversandEconomiesofSizeintheEfficientDesignof

    AgriculturalResearchSystems.Chapter9inJ.M.Alston,P.G.Pardey,andM.J.Taylor,eds.Agricultural

    SciencePolicy:ChangingGlobalAgendas,Baltimore:JohnsHopkinsUniversityPress,2001.

    ChanKang,CandP.G.Pardey.ACenturyofWheatVarietalInnovationintheUnitedStates:AnEmpirical

    Assessment.DepartmentofAppliedEconomics.St.Paul:UniversityofMinnesota,2010(inpreparation).

  • 8/6/2019 3568 Pardey Pingali 2010 GCARD Text Figs Tabs 1

    36/55

    Dehmer,S.andP.G.Pardey.PrivateAgriculturalR&DintheUnitedStates,19502006.DepartmentofApplied

    EconomicsStaffPaper.StPaul:UniversityofMinnesota,2010(inpreparation).

    Economist.2009AgriculturalRobotsFieldsofAutomation,December10,2009.

    Evenson,R.E.andD.Gollin.AssessingtheImpactoftheGreenRevolution,19602000.Science300(May

    2003):758762.

    FAO(FoodandAgricultureOrganizationoftheUnitedNations).HungerontheRise:SoaringPricesAdd75

    MillionPeopletoGlobalHungerRolls.FAOBriefingPaper.Rome:FAO,September2008.Availableat

    http://www.fao.org/newsroom/common/ecg/1000923/en/hungerfigs.pdf.AccessedNovember2008c.

    ____.MorePeopleThanEverAreVictimsofHunger.FAOPressRelease.Rome:FAO,June2009.Availableat

    http://www.fao.org/fileadmin/user_upload/newsroom/docs/Press%20release%20juneen.pdf.AccessedJune

    2009.

    Frisvold,G.B.,T.M.HurleyandP.D.Mitchell.Overview:HerbicideResistantCropsDiffusion,Benefits,Pricing

    andResistanceManagement.AgBioForum12(2&3)(2009):244248.

    Gebbers,R.andV.I.Adamchuk.PrecisionAgricultureandFoodSecurity.Science327(February2010):828

    831.

    Griliches,Z.HybridCorn:AnExplorationintheEconomicsofTechnologicalChange.Econometrica25(1957):

    501522.

    Herdt,R.W.andC.Capule.Adoption,SpreadandProductionimpactofModernRiceVarietiesinAsia.LosBanos:

    InternationalRiceResearchInstitute,1983.

    IDA.AidArchitecture:AnOverviewoftheMainTrendsinOfficialDevelopmentAssistanceFlows.

    InternationalDevelopmentAssociation.ResourceMobilization(FRM),2007.

    James,J.S.,P.G.PardeyandS.Wood.TechnologicalDistance,InternationalSpillovers,andStocksof

    AgriculturalKnowledge.DepartmentofAppliedEconomicsStaffPaper.StPaul:UniversityofMinnesota,2010,

    (inpreparation).

    Just,R.E.,J.M.AlstonandD.Zilberman,eds.RegulatingAgriculturalBiotechnology:EconomicsandPolicy.New

    York:SpringerVerlag,2006..

    Jin,S.,S.Rozelle,J.M.Alston,andJ.Huang.EconomiesofScaleandScopeandtheEconomicEfficiencyof

    ChinasAgriculturalResearchSystem.InternationalEconomicReview46(3)(August2005):10331057.

    Liebenberg,F.,P.G.PardeyandM.Khan.SouthAfricanAgriculturalResearchandDevelopment:ACenturyof

    Change.StaffPaperP1001.St.Paul:UniversityofMinnesota,DepartmentofAppliedEconomics,January

    2010.

    McIntyre,B.D.,H.R.Herren,J.WakhunguandR.T.Watson.InternationalAssessmentofAgriculturalKnowledge,

    ScienceandTechnologyforDevelopment.WashingtonD.C:IslandPress,2009.

    Pardey,P.G.andC.ChanKang.PublicandPrivateFoodandAgriculturalResearchInvestmentTrends:

    UnpublishedDataFiles.St.Paul:DepartmentofAppliedEconomics,UniversityofMinnesota,2010(in

    preparation).

  • 8/6/2019 3568 Parde