ContentslistsavailableatScienceDirect
NJAS
-
Wageningen
Journal
of
Life
Sciences
j ou rn a l h o m e p a g e :w w w . e l s e v i e r . c om / l o c a t e / n j a s
Efficiency
of
European
Dairy
Processing
Firms
Rafat
A.M.E.
Soboh
a,
Alfons
Oude
Lansink
b,∗,
Gert
Van
Dijk
c aEconomicDepartment,AlQudsUniversitybBusinessEconomicsGroup,WageningenUniversity
cChairCooperativeEntrepreneurshipandBusinessAdministration,NyenrodeBusinessUniversity
a
r
t
i
c
l
e
i
n
f
o
Articlehistory: Received1April2013
Receivedinrevisedform23April2014 Accepted2May2014
Availableonline17June2014 Keywords: Dairyprocessing Cooperatives Investorownedfirms Productionfrontier
a
b
s
t
r
a
c
t
Thispapercomparesthetechnicalefficiencyandproductionfrontierofdairyprocessingcooperatives andinvestorownedfirmsinsixmajordairyproducingEuropeancountries.Twoparametric produc-tionfrontiersareestimated,i.e.forcooperativesandinvestorownedfirmsseparately,whichareused toevaluatetheintra-andinter-typetechnicalefficiencyofeachfirm.Themodelsareestimatedusing accountancydatafromtheAMADEUSdatabaseovertheyears1995-2005.Resultsshowthatdairy coop-erativeshaveamoreproductivetechnology,butareslightlylessefficientthaninvestorownedfirms duringthisperiod.Differencesinproductiontechnologyandtechnicalefficiencyofthecooperatives acrosscountriesreflectdifferencesinlocalmarketconditionsandcharacteristicsofthecompanies.Both cooperativesandinvestorownedfirmsarecharacterizedbydecreasingreturnstoscale.However,returns toscaleareonaveragemuchsmallerforcooperatives,implyingthatthescaleofoperationofcooperatives istoolarge.
©2014RoyalNetherlandsSocietyforAgriculturalSciences.PublishedbyElsevierB.V.Allrights reserved.
1. Introduction
Evaluatingtheproductionefficiencyandperformanceof mar-ketingcooperatives(Coops)andinvestorownedfirms(IOFs)1isof
interesttopolicymakersandfirmmanagers,whoareseekingto enhancecompetitivepoweroffirms.Thosewhoareinterestedin thistopicrecognizethatacomparisonbetweenIOFsandCoopshas toaccountforfundamentaldifferencesinobjectivesand organiza-tionalstructures[1].Underlyingthedifferences,betweenCoops andIOFs,is themore complexrelationshipbetweenCoopsand theirmemberscompared totherelationship betweenIOFsand theirshareholders[2].Membersofthecoopsareoftentheowners, controllers2,and themain(ifnottheonly)usersoftheservices
∗Correspondingauthor.
E-mailaddress:alfons.oudelansink@wur.nl(A.O.Lansink).
1Intheeconomicliterature,differentdefinitionsofcooperativesexistamong whichisthedefinitionthatcooperativesareanextensionofthefarm-business (Sobohetal.,2009).Also,cooperativesaredefinedasuser-controlled,user-owned anduser-benefitorientedfirms(Sobohetal.,2009).ChaddadandCook(2004) pro-videdanownershiptypologyofcooperativeswhicharediverseintermsofthelevel ofcontrol,ownershipandbenefitsofthemembersinthecooperativefirms.IOFsare oftendefinedasentitiesthataimtomaximiseshareholdervalue.
2Members,asowners(residualclaimants),areentitledtothenetincome gener-atedbythefirmandalsotheresidualriskbearersofthefirm’snetcashflows.While members,ascontrollers(havetheresidualrighttocontrol),havetherightstoany
of cooperatives[2].Literature hassuggesteda varietyof objec-tivesofCoopssuchasserviceatcost,maximizingtheaggregate members’profitand maximizingtheaggregatecooperativeand membersprofit[1].Hence,objectivesofmembersandCoopsare morealigned3 thantheobjectivesofIOFsandshareholders,who
primarilyfocusonmaximizingreturntoequity.Thealignmentof objectiveswithinCoopsshapestheorganizationalstructureofthe cooperativeinthelong-run[3].Thedifferencesinobjectivesand organizationalstructurebetweenIOFsandCoopsimplythatIOFs and Coopsdiffer inthe waytheytrade offrevenues, costsand investments[4]).Moreover,Coopsaremainlyfinanced bytheir members,whoaremorereluctanttoengageinriskynewventures. Asaresult,Coopsfocuslessonthemoreriskyvalue-added seg-mentsofthemarketsthanIOFs[26].Anexceptionisinvestmentin theirabilitytoprocessthepotentialcapacitygrowthofits mem-bers.Thismeansthat Coops’successfactorsarerelatedtotheir competenciestosustain-ablyfulfillingthemembers’objectivesat thesametimebeingabletofunctioninthecompetitivemarket[26]. Thesesuccessfactorsarecomparabletothesuccessfactorofthe
assetsthatarenotassignedtootherpartiesnorattenuatedbylaw(Chaddadand Cook,2004).
3However,thereareexamplesintheliteraturewhichshowsthatthisalignmentis notperfect.Forinstance,withaheterogeneousmembership,non-membersowners, distributionalissues(see[27]and[28]).
http://dx.doi.org/10.1016/j.njas.2014.05.003
IOFsinmeetingtheirshareholdersobjectives,yetitshouldbekept inmindthattheIOFsgoalsaremorecloselylinkedto profitabil-ity(Gentzoglanis,1997).TheCoopsprimaryconcernistoprocess allofitsmembers’productsandvalorisethisoptimally.Thislimits thecooperativeswiththeirchoicesofmarketsegments,whereas theIOFsdeterminetheirmarket(segments)basedon maximis-ingprofit(KlepandvanDijk,2005). Thesizeoftheiroperation dependsontheirmarketstrategyandtheyarenotobligedto pro-cessanypredeterminedquantity,contrarytotheexpectationfrom theCoops.
Differencesinobjectivesandorganizationalstructuresbetween IOFsand Coopsare assumedtoaffecttheirproduction technol-ogyandtechnicalefficiency.BontemsandFulton[3]arguedthat thestrongalignmentofCoopsandtheirmembers’objectivesputs CoopsataproductionefficiencyadvantagecomparedtoIOFs. Con-trarytotheargumentofBontemsandFulton[3]; Oustapassidis
etal.[5]arguedthatcooperativesendorseover-supplyofmembers’ inputsandthereforeoperateonaless efficientscale.Moreover, PorterandScully[6],FerrierandPorter[7],andGentzoglanis[8]
arguedthatcooperativesarelessscale-efficientduetohighercosts ofcontroltotacklethepresenceoffree-riders.
Thepresentstudyemploysa(parametric)stochasticproduction frontiertomeasure differences in technical efficiencyand pro-ductiontechnologybetweendairyprocessingCoopsandIOFs.The parametricapproachtoefficiencymeasurementhasbeen imple-mentedwidelyinseveralindustries(seee.g.Schmidt[9],Lovell andSchmidt(1988),andBauer[10]eachforanoverview,and Bat-tese[11]forareviewoftheliteratureofstochasticfrontieranalysis intheagriculturalsector).Applicationoftheparametricfrontier approachtocomparingefficiencyonCoopsandIOFshasbeen pre-viouslyundertakenbySextonetal.[12],DoucouliagosandHone
[13],Singhetal.[14],Mosheim[15]4.Existingstudiesonmeasuring
andcomparingtechnicalefficiencyofCoopsandIOFsare,however, restrictedbytheassumptionthattheperformancesofIOF’sand cooperativeshavethesameproductionfrontier,i.e.thetechnology rulingthetransformationofinputsintooutputsisthesameforIOFs andCoops.
Thispaperestimatesproductionfrontiersfordairyprocessing CoopsandIOFsseparatelyandtestswhetherdifferencesintheir productionfrontiersarestatisticallysignificant.Firmtype-specific productionfrontiersandfirm-specificefficiencyparametersallow obtainingintra-andinter-typetechnicalefficiencymeasurements. Intra-typetechnicalefficiencyinvolvescomputingtheefficiencyof aparticularfirm’s(ofaparticulartype,eitherCooporIOF)relative toeithertheCoopsortheIOFsfrontier.Inter-firmtechnical effi-ciencyinvolvescomputingtheefficiencyofthatsamefirmrelative tothefrontierofthefirmtypethatrepresentsthebestpractice (i.e.thismayormaynotbethefrontierofthealternativetype). Inter-type(IE)efficiencyisdecomposedintointra-typeefficiency (IA) and inter-typecatch-up(CU) components.The CU compo-nentmeasuresdifferencesinproductiontechnologyacrossthetwo typesoffirms(CoopandIOF),whichmayarisefromdifferencesin investmentsinnewtechnologiesanddifferencesinthequalityand characteristicsoftheinputsused[16].Theempiricalapplication focusesondatafromCoopsandIOFsinthedairyprocessing indus-tryinBelgium,theNetherlands,Ireland,Denmark,Germanyand France.
2. EconometricsMethodology
Thestochasticfrontierproductionmodelwaspresented inde-pendentlybyAigneretal.[17]andMeeusenandvandenBroeck
4 SeeSobohetal.,(2009)foranextendedlistofliteratureusedsimilartechniques whilecomparingtheperformanceofcooperativestoIOFs.
* IOF IOF Coop YI Y’ I Y” I Y X XI
Fig.1. AnIllustrationofanIOFProduces(YI)BelowBothFrontiers.
(1977)[25].Thestochasticfrontieranalysis(SFA)hasbeen theoret-icallyandempiricallyimplementedonfirmsofdifferentindustries andsectors(see[18]).Measuringproductionefficiencyallowsone totestcompetinghypotheses regardingsources ofefficiencyor differencesinproductivity[19,20].Moreover,suchmeasurement allowsforquantifyingthepotentialincreasesinoutputthatmight beassociatedwithanincreaseinefficiency[19].
Inthispaper,twoTranslogproductionfrontiers(i.e.forCoops and IOFs separately) are estimated using data onthree inputs and one output of dairy processing firms in the six European countriesmentionedabove.Thestochasticfrontiermodel, assum-inganexponentialdistributionofthenon-negativerandomterm
uiisexpressedas:
Yi=f(xi,ˇ)e(vi−ui),i=1,...,N (1)
whereYiistheoutputofthei-thdairyfirm;xiisak×1vectorof
theinputquantitiesofthei-thtypeoffirm;isaparameter vec-tor.Moreover,
v
irepresentstherandomdisturbancetermthatisassumedtobei.i.d.N(0,2
v).Thisrandomterm(
v
i)isincorporatedintothemodelandisindependentofui.Thenon-negative
ran-domterm(ui)representstechnicalinefficiencyinproductionand
isassumedtobei.i.d.andhasanexponentialdistribution.The rela-tionshipbetweenuiandtheoutputorientedtechnicalefficiency
(TE)is:TEi=exp(−ui).
Estimatingthetwotype–specificproductionfrontiers(Coops andIOFs)allowsformeasuringtechnicalefficiencyofthefirm rel-ativetothefrontierfor:(1)itsowntypeand(2)thebestpractice frontier.Usingthisapproachwecanmeasuretheinter-typecatch upcomponent,whichindicatestheratioofthefirm’sefficiency fromitsowntype’sfrontiertothefrontierrepresentingthebest practice.Measuringthetechnicalefficiency,inthisapproach, indi-catesthatthebestpracticefrontiercaneitherbethesametype ofthefirmortheothertype.Moreover,measuringthetechnical efficiencyofthefirmsrelativetotheothertype’sfrontier(ifitwas thebestpracticefrontier)indicatesapotentialimprovementinthe technicalefficiencyofanyfirmofaparticulartypewhenadapting theothertype’stechnology. Inthispaper,weuseterminologies usedbyOudeLansinketal.OudeLansinketal.(2000,2001),the firstmeasurementiscalledintra-typeefficiency(i.e.efficiencyof thefirmrelativetoitsowntypefrontier),whilethesecondiscalled inter-typeefficiency(i.e.efficiencyofthefirmrelativetothebest practicefrontier).Toillustratetheseconcepts,takethecaseofan IOF(Fig.1):
TheactualoutputproducedbytheIOFis YI,Y’Ireferstothe
maximumobtainableoutputfromthefrontieroftheIOF’stype:
Y
---: Coop; ______ : IOF * IOF1 * IOF2 * Coop1 * Coop2 X Y
Fig.2.CoopsandIOFsRepresentedbyTwoFrontiers.
WhileY
I referstothemaximumobtainableoutputfromthe
bestpracticefrontier:
Y
I =maxI,C [fI(XI),fc(XI)] (3)
Theintra-typeefficiency(IA)isdefinedas: IA= YI
Y I
;0<IA≤1 (4)
Theinter-typeefficiency(IE)isdefinedas: IE=YYI
i
;0<IE≤1 (5)
Thecatch-up(CU)componentistheratioofoutputobtainedat theowntypefrontierandtheoutputobtainedatthebestpractice frontier:
CU= YI
Y I
;0<CU≤1 (6)
TherelationbetweenthesemeasuresisIE=CU×IA,hence
IE=YYI I × YI Y I = YI Y (7)
ThecriticalpointiswhenCUofanytype(i.e.CooporIOF)is equalto1.Firms,ofcertaintype,withCU=1performbetterwith theirowntypetechnologythanwiththeothertype’s.Whilefirms, ofcertaintype,withCU<1performworsewiththeirowntype’s technologyratherthanwiththeotherfirmtype’stechnologyand thereforehaveapotentialfortechnologicalimprovement.
Fig.2illustratesthefourpossiblesituations:aCoopabovethe IOF’sfrontier(Coop1),anIOFabovetheCoop’sfrontier(IOF2),a Coopbelowbothfrontiers(Coop2),andanIOFbelowboth fron-tiers(IOF1).InFig.2,theobservationCoop1islocatedinthepart ofthefrontierwheretheCoopsfrontierislocatedabovetheIOFs frontier.Hence,theCUofthecooperativewillbeequalto1.IOF1 isalsolocatedinthepartwheretheCoopsfrontierdominatesthe IOFfrontier.Hence,CUissmallerthan1forIOF1.
3. DataandEmpiricalModel
3.1. Data
Paneldataondairy processingfirms(ofbothtypes)fromsix Europeancountries(Belgium,Denmark,France,Germany,Ireland andtheNetherlands)wereusedinthisstudy.Thesecountriesare consideredtobetheEuropeanstatesthatproducethebulkofthe EUcow’smilk[22].Thepaneldatacoverstheperiod1995-2005 andisobtainedfromAMADEUS.AMADEUSisaEuropeanfinancial databasepreparedbyBureauvanDijkandcontainsmorethan5
millionprivate,cooperativeandpubliccompanies.Thedata-base iscollectedfromreportsproducedbythechamberofcommerce ofthedifferentEuropeancountries.AMADEUSstandardized the figuresofthefinancialstatementsofthedifferentcountries.Our sampleconsistsof1221observationsamongwhich861areIOFs and360arecooperatives.
The model distinguishes one output (total turnover), three inputs(materialcost,employmentcost,andfixedassets)andatime trendvariable(T).TheoutputsandinputsareexpressedinEurosof 1996(baseyear)bydeflatingthemonetaryvalueswiththeirprice indexes(providedbyEurostat).
Thedairyprocessingfirmsinthesecountriesaretypically pro-ducingmultipledairyoutputs,suchas:cheese,yoghurts,butter, freshmilk.However,theavailabledatareportsonlythetotal rev-enuesanddonotdistinguishbetweenrevenuesfromthedifferent sub-categoriesofthetotaloutput.Outputismeasuredastotal oper-atingrevenuefromsellingallproductsproducedbytheprocessing company,deflatedbythecountry-specificpriceindexofconsumer prices (a producer price off-factorywas not available) of milk, cheeseandeggs.
Fixedassetismeasuredasthevalueofphysicalland,buildings, machinery,andthenon-physicalfixedassets(suchasthe good-will,patents,brands,andmarketshares).Thevaluewasdeflated usingtheaveragevalueofthecountry-specificpricesindexofthe agriculturalgrossfixedcapitalformationandthepriceindexofthe agriculturalmachineryandequipmentpercountry.Thedatabase providesuswithaninputtitledasmaterialcost.Thisvariablerefers tothecostofpurchasingtheinputmaterialsbeforethestarting oftheprocessingoperation.Thisinputconsistsmainlyofthecost ofpurchasingrawmilkbythedairyprocessingfirm.Weusedthe deflatedEC-indexofproducerpricesofthecows’milkper coun-tryasthedeflatorforthematerialcost.Laborcostisdeflatedusing thecountry-specificnominalvalueofthelaborcostindexintotal industries(excludingpublicadministration).
Table1providesthemean,minimumandmaximumvaluesof turnover,fixedassets,rawmaterial,andlabor.Itshowsthat coop-erativeshave,onaverage,ahighervalueoftheoutputandthree inputsthanIOFs.
3.2. EmpiricalModel
Thetype-specificfrontierforeachtypeoffirmsinthesampleis assumedtofollowaTranslogspecification:
ln(Yit)=ˇ0+ 3
i=1 ˇiln(xit)+0.5⎛
⎝
3 i=1 3 j=1 ˇijln(xit)ln(xjt)⎞
⎠
+1(T)+0.5(2T2)+0.5 3 i=1 iTln(xit) +(it−uit)Wherexitareinputquantitiesattimet,withi=1(fixedassets),
2(rawmaterials),and3(Labour).Atimetrend(T)isincludedinthe empiricalmodeltoaccountforexogenoustechnologicalchangein theestimation.Thecoefficientsofthe3inputs(ˇi,i=1,2,3),the
quadraticpartofthemodel(ˇij,iandj=1,2,3),thetimetrend(1),
itsquadraticterm(2),andthecrosstermofthetimetrendwith
theinputs(i,i=1,2,3)aretobeestimated.
4. Results
Thelog-likelihood-testwhichteststhenullhypothesisofno dif-ferenceinproductiontechnologybetweenIOFsandcooperatives
Table1
DescriptiveStatistics(MillionEuros).
Cooperatives(n=360) Investor-ownedfirms(IOFs)(n=861)
Mean Minimum Maximum Mean Minimum Maximum
Output 429.26 2.91 5513.82 96.45 2.58 1605.03
Fixedassets 108.30 0.03 2243.13 19.59 0.01 406.47
RawMaterial 202.57 0.04 4604.08 7.25 0.03 222.31
Labor 24.85 0.01 742.61 1.12 0.01 24.09
Table2
ResultsofEstimationoftheTypeSpecificFrontiers(p-valuesinbrackets). Variable Cooperatives Investor-OwnedFirms(IOFs) Constant 12.138(0.000) 4.979(0.000) Ln(x1) −0.560(0.000) −0.105(0.403) Ln(x2) −0.536(0.018) −0.344(0.049) Ln(x3) 0.573(0.004) 0.810(0.000) T −0.007(0.959) 0.382(0.000) Ln(x1)Ln(x2) 0.011(0.550) 0.020(0.164) Ln(x1)Ln(x3) −0.020(0.162) −0.036(0.030) Ln(x1)T 0.004(0.621) −0.024(0.000) Ln(x2)Ln(x3) −0.052(0.000) 0.004(0.747) Ln(x2)T −0.001(0.915) 0.015(0.054) Ln(x3)T −0.009(0.435) −0.024(0.001) ½[Ln(x1)Ln(x1)] 0.124(0.000) 0.124(0.000) ½[Ln(x2)Ln(x2)] 0.086(0.000) −0.002(0.924) ½[Ln(x3)Ln(x3)] 0.032(0.248) −0.028(0.186) ½T2 −0.001(0.970) −0.010(0.052) x1=fixedassets;x2=rawmaterial;x3=labourandT=timetrend.
yieldsavalueof685.Thisimpliesthatthenullhypothesisisrejected
at5%.Inwhatfollows,wediscusstheparameterestimatesandthe estimationoftechnicalefficiencybasedontwoseparateproduction frontiersforIOFsandcooperatives.
Table2providestheestimationresultsofthetwoproduction frontiers.Thep-valuesindicatethatthesixcoefficientsofCoops’ productionfrontieraresignificantatthe5%criticalvalue.Allsingle termsaresignificantintheCoops’productionfrontier,exceptfor timetrend,whereasmostinteractionandquadratictermsarenot significantfortheCoopsfrontier.Ninecoefficientsweresignificant atcritical5%levelfortheIOFs’productionfrontier,suggestingthat theIOFsaremorehomogeneousthantheCoops.Onlythree param-eters(singletermsofmaterialcost,labor,andthequadratictermof fixedassets)aresignificantforbothfrontiers.Theparameter esti-matesofsixcoefficientshaveoppositesignsforthetwofrontiers (suchas:thetimetrend).
Table 3 provides the annual mean, maximum and mini-mumvalueofthethreedifferentefficiencymeasurementsacross countriesandoftheaverageoftheentiresample,(intra-firm(IA), Catch-Up(CU)andInter-firmefficiency(IE)).Additionally,these measurementsarepresentedbycountrytoinvestigateany differ-encesbetweencountries.
TheresultsinTable3 showthattheaverageintra-firm effi-ciencyscoreoftheCoops(IAC)islowerthanintra-firmefficiency
scoreoftheIOFs(IAI).ThisindicatesthattheCoopsare,onaverage,
lessefficientrelativetotheirowntechnologythanIOFstotheirs. TheresultsacrosscountriesshowthatCoopsinBelgium,Germany, IrelandandtheNetherlandshaveabetterintra-firmefficiencythan theIOFsinthesamecountry.Therefore,Coopsinthesecountries bettersucceedinexploitingtheirtechnologytoitsfullpotential thantheirIOFscounterparts.Theefficiencymeasurementsreflect,
5 Theresultofthelog-Likelihoodratiotestyieldsavalueof(68)intestingthe hypothesisofdifferenceintheproductiontechnologyoftheIOFswiththe coopera-tiveswithcomparableturnoversize(i.e.excludingthecooperativesthatarelarger thantheIOFs).Thisimpliesthatthesizeoftheturnoverdoesnotalterthedifference betweentheproductiontechnologyofIOFsandcooperatives.
notonlytheefficiencyofthephysicaltransformationofinputsinto outputsbutalsoimplicitlyreflectthesuccessofotherintangible dif-ferencessuchasmarketingandmanagementstrategiesaimingto generatemorevalueaddedandoutputqualitydifferencesbetween companiesofthesametype(eitherCooporIOF).Itisimportant tonotethatitisimpossibletotestwhetherthedifferencesinIA (norCUandIE)betweencountriesandbetweenIOFsandCoops aresignificant.ThisisbecausethedifferencesinIA(CUand IE) havebeendeterminedusingtheresultsoftheestimationoftwo separateStochasticProductionfrontiers.
Theimplicationof theresultsonintra-firmefficiencyisthat Coopscan,onaverageincreasetheiroutputby24%whileusing thesamebundleofinputs,whereasIOFscanincreasetheir out-putby21%,usingthesamebundleofinputs.Theimprovement oftechnicalefficiencycanbeachievedbetterinternal organisa-tion,suchthatmaterials,andfixedassetsareusedmoreefficiently. Also,technicalefficiencycanbeimprovedbyimprovingthe qual-ityand/orpriceoftheiroutputs,e.g.byincreasingthemarketing effortsorbystrengtheningthemarketpositionvis-a-visretailers andwholesalers.
Additionally,theresultsinTable3showthat,onaverage,the catch-upcomponentoftheCoops(CUC)equalsonewhereasthe
catch-upcomponentofIOFs(CUI)isonaverageequalto0.8.The
differencebetweenthetwocatch-upcomponents(CUCandCUI)is
onaverageequalto0.1858.Thisresultindicatesthesuperiorityof theCoops’technologyovertheIOFs’technology,whichalsoresults onaverageinahigheroverallInter-firmefficiencyforCoops.These differencesinfavoroftheCoopsvis-a-vistheIOFsareattributedto afewaspects;differencesinthetechnologygoverningthephysical transformationofinputsintooutputs,qualitydifferencesand dif-ferencesinthesuccessofmarketingstrategies(betweenIOFsand Coops).However,thesizeofthedifferencebetweentheCoops’-and IOFs’-catch-upcomponents(CUCandCUI)differsalsoacrossthe
differentcountries.Forexample,Denmark,hasthelowestvalueof thedifferenceinthecatch-upcomponentsbetweenIOFsandCoops (0.1407).ThissuggeststhatthesuperiorityoftheCoops’ technol-ogyoverthatoftheIOFsislessapparentinDenmark.Irelandhas thehighestvalueofthedifferenceincatch-upcomponent(0.3128), suggestingthatthecooperatives’technologyisthemostsuperior totheIOFs’technologyinIreland.
Table3
EfficiencyScoresandCatch-UpComponents*
Cooperatives Investor-owned firms(IOFs) Difference IAC IEC CUC IAI IEI CUI CUC-CUI Belgium 0.81 0.81 1 0.78 0.65 0.8166 0.1834 Denmark 0.62 0.62 1 0.75 0.65 0.8593 0.1407 France 0.74 0.74 1 0.80 0.70 0.8580 0.1420 Germany 0.82 0.82 1 0.78 0.67 0.8547 0.1453 Ireland 0.83 0.83 1 0.81 0.58 0.6872 0.3128 Netherlands 0.78 0.78 1 0.72 0.62 0.8094 0.1906 Total(Mean) 0.76 0.76 1 0.79 0.68 0.8142 0.1858 *IA,IEandCUindicateintra-firm,Catch-UpandInter-firmefficiency, respec-tively.ThesubscriptsCandIrefertoCooperativesandIOF,respectively.
Thedifferenceinthecatch-upcomponent betweenIOFsand Coopsacrosscountriesisexplainedbycross-countrydifferences inthe(i)characteristicsofcooperatives/IOFsandthe(ii)market conditionsforcooperatives/IOFs.Anexceptionaretheresultsof Denmark,whichrequiresanalyzingthecharacteristicsofthe Dan-ishfirms,dairymarketandthedegreeofproductionspecialtyof theDanishIOFs.TheCoopsdifferacrosscountriesintermsoftheir financialstructure(partiallytradedinthepublicmarketornot), marketorientationandmembership(targetinginternational mar-ketsornot and havingmembersfrom differentcountries).The differencesinthefinancialstructureofCoops,refer tothelevel of controland ownership of membershave of thecooperative. Forinstance,cooperativesinIrelandcantradeabouthalfoftheir equityinthestockexchange,whichimpliesmorediversityinthe sourcesofcapital[23].This,relativediversifiedsourceofcapital fortheIrishcooperatives6,inadditiontothecharacteristicsofthe
Irishdairysector,mayexplain thelargedifferencein thecatch upcomponentsofIrishCoopsandIOFs.Thedifferencesin mar-ketorientationofcooperativesacrosscountriesrefertowhether thecooperativeistargetinginternationalmarketsornot. Coop-erativesinDenmark(whichexportsaround2/3ofthetotalmilk production),theNetherlands,andBelgiumaretotargeting inter-nationalmarketsmorethanFrenchandGermancooperatives[24]. TheIrishcooperativesaremainlyexportingtheirproductionto theUK,whichisencouragedbythegeographicallocationandthe strongUKpoundrelativetotheEuro.
ThelocalmarketconditionsforCoopsrefermainlytotheCoops’ marketsharewhichreflectsthedegreeofcompetitivenessinthe localmarketandtheopportunitytoexercisemarketpower.Asan example,onelargecooperativeinDenmarkproducesmorethan 90%ofDanishmilkproduction,whereastherearemorethan10 othersmallcooperativeswhichhandlearound6%ofDanishmilk production.ThisleavestheDanishIOFstoprocessless than3% of thetotal Danishmilk production (Danish dairy board7).The
largeproportion of the processed milk bycooperative (around 97%-thelargestinEurope)reflectsthatthemajorityofthe Dan-ishfarmersarecooperative-orientedand thatDanishIOFsmust behighlyspecializedwithloyalcustomers.Thecharacteristicsof theDanishmarketsuggestthatthecountervailingpowerof coop-erativesiswellfunctioningandthatthedairymarketismainly exploitedbyCoops,whichleavesno(orlittle)marginforany fur-thercooperativeexploitation.Thesituationofthedairyprocessing industryinIreland,however,islargelyfragmentedanddominated by three relatively not so large Coops8. The Irish cooperatives
haveweakincentivestoconsolidateandtheyachieveefficiencyby co-processingarrangementandmilk-sharingarrangementsrather thanbyexpandingrevenueareas.TheIrishdairysectorisnotfully exploitedandhaspotentialtoimprovecooperatives’shareofthe dairy processing sector [23]. In theNetherlands, there are two largecooperatives(presentlymergedintoasingleone),processing about85%ofthetotalmilkproductionintheNetherlands,whereas threesmallcooperativesaccountforanother6%oftotalDutchmilk production(ProductschapZuivel).TheGermanandFrench cooper-ativesaccountforlessthan60%and40%ofthelocaldairymarket, respectively(ProductschapZuivel)9.
Tofurtheranalyzetheresults,wecomparethenumberandthe characteristicsofthefirmswithcatchupcomponentequaltoone tothosewithcatchupcomponentlowerthanoneinTable4.Only
6Termedaspubliclimitedcompanies(PLC).ThePLCisnotobligedtotakeinmore milkofmembersornewmembership(Harte,1997).
7Thesefiguresareestimatedbasedondataavailableon(www.mejeri.dk). 8WhencomparedtothelargecooperativesinDenmarkandtheNetherlands. 9Thecooperatives’marketsharepercountryisorderedfromlargest:DK,NL,Ire, B,D,Fr.(basedonestimationbasedondatafrom(www.prodzuivel.nl)
Table4
TheCharacteristicsoftheFirmswithDifferentValuesoftheCatch-Up. Item(average)MillionEuros Cooperatives InvestorsOwned
Firms CUC=1 CUC<1 CUI=1 CUI<1 Numberoffirms 359 1 7 854 Turnover 430.38 450.73 155.57 95.96 Fixedassets 108.43 100.09 9.75 19.67 RawMaterial 203.13 187.79 29.97 7.07 Labor 24.92 30.36 0.93 1.12 Intra-FirmEfficiency 0.76 0.86 0.87 0.79
onecooperative(ofthe360cooperatives)hasacatchupvalueless thanone,whereas,onlysevenIOFs(ofthe861IOFs)havea catch-upvalueequaltoone.ThesesevenIOFshave,onaverage,ahigher turnover,fixedassetsandrawmaterialsbutloweruseoflaborthan theotherIOFs.Also,thesesevenIOFshave,onaverageahigher effi-ciencyscore(0.87)thantheotherIOFs(0.79).Whereas,thesingle cooperative(withCUC<1)doesnotappeartohavedifferent char-acteristics(intermsofvalueofoutputandinputs)thantheaverage characteristicsoftheothercooperatives(withCUC=1).The intra-firmefficiencyscoreofCoops(IAC)oftheonlycooperativewith CUC<1isequalto0.86andishigherthantheaveragevalueofthe intra-cooperativesefficiencyscore(0.76),oftheothercooperatives (withCUC=1).
Almost100percentofthecooperativeshaveacatch-upvalueof one,whereasaround99percentoftheIOFshaveacatch-upvalue lessthanone.Theseresultsshowthatthetechnologyofrelatively smallIOFs(intermsofturnover)isdominatedbythetechnology ofthelargerCoops.TheIOFswithacatch-uptermsmallerthan onehavearelativelyhighvalueofbothfixedassetsandlabor com-paredtothesevenIOFswithacatch-uptermequaltoone.This relativelyhighvalueoffixedassetsandlaborofthesmallIOFmay beduetotheirhighcostsinproducingalimitedquantityof spe-cializedoutputfromarelativelysmallquantityofrawmaterials. ThebusinessimplicationsoftheseresultsarethatsmallIOFneed toimprovetheirtechnicalefficiencyinordertopreventlossesdue toinadequateorganization.Theymayadaptstrategieswhicheither reducetheuseofbothfixedassetsandlabor,orincreasethelevel ofoutput.
ThedifferencesinIntra-firmEfficiency(IA)ofCoopsandIOFs suggest that Coops have different internal incentives and are focusedonothermarketsegmentsthanIOFs.TheCoopsfirst inten-tionistocollectandprocessestheirmembers’entireproduction whileIOFsdrawtheirincentivesfromcollectingandprocessing thatmilkquantitythatmaximizestheshareholders’value.Asa result,IOFs’areinclinedtotargetdifferentmarketsegmentsthan thoseofthecooperatives.Thehighvalue-addedsegmentsofthe market requirehighquality production,packagingand market-ing, whilequantitieswill bemuch lower thanin Coops, which explainstherelativelyhighvalueoffixedassets tototal output ofthesmallIOFs.Second,targetingthevalueaddedsegmentsof themarketrequirehighlyskilledandspecializedlaborand man-agement,whichexplainstherelativelyhighvalueoflabortototal outputofthesmallandthelowerInter-firm(overall)efficiencyof IOFs.Thecooperativeisrequiredtomeetmembers’interestsby processingtheirentireproductionofrawmilkwhilepayingthem (forprovidingtherawmaterial)acompetitiveprice.
Table5providestheaveragevalueoftheoutputelasticitiesof cooperativesandIOFsandthep-value,testingwhetherthe elas-ticityis(statistically)significantlydifferentfromzero.Theoutput elasticity,ofthefirms(ofbothtypes)withcatch-upequaltoone, differsfromtheelasticityoffirmswithcatch-uplessthanunity.
Theresults,presentedinTable5,showthattheoutputelasticity offixedassets(whichincludesbrandvalue)hasthehighestvalue
Table5
OutputandScaleElasticities(p-valuesinbrackets).
Outputelasticityof(p-value) Cooperatives Investor-ownedfirms(IOFs)
CUC=1 CUC<1 CUI=1 CUI<1 N=359 n=1 n=7 n=854 Fixedassets 0.569(0.000) 0.135(0.000) 0.629(0.000) 0.593(0.000) RawMaterial −0.102(0.001) −0.114(0.001) 0.056(0.159) 0.034(0.133) Labor 0.071(0.013) 0.100(0.010) −0.017(0.301) 0.022(0.359) Time −0.037(0.005) −0.045(0.006) −0.021(0.234) −0.012(0.194) Elasticityofscale 0.623(0.020) 0.121(0.017) 0.668(0.034) 0.646(0.028)
amongtheinputs.Therefore,thisinput(fixedassets)isimportant forbothIOFsandCoops.However,fixedassetsismoreimportantto IOFs(withCUI=1andCUI<1)thancooperatives(withCUC=1and CUC<1).Theoutputelasticityoffixedassetsissignificantlydifferent fromzeroforCoopswithCUC=1andtheIOFswithCUI<1.
Theoutput elasticityof raw materials is negativefor Coops and is significantly different from zero at the critical 5% level for those Coops withcatch-up component equal to one, while it is insignificant and positive for IOFs. The negative value of theelasticityofrawmaterialssuggeststhatCoopsoveruseraw materials (raw milk), to such an extent that it decreases the producedoutputoftheCoops.Thisisexplainedbythe coopera-tive’sobligationtoprocesstheentireproductionofitsmembers. In addition,the commonagriculturalpolicy (CAP) allowsdairy processingcompaniestoprocesstheirsurplusmilkintolowvalue added commodities that can bestored (e.g. butter, milk pow-der).However,inordertoexplainthenegativeoutputelasticity ofmaterialsfor Coops,more information onquality differences of materialsand more details onthe composition of output is needed.
TheoutputelasticityoflaborispositiveandishigherforCoops thanforIOFs.IthasanegativevalueforIOFswithcatch-up com-ponentequaltoone.Moreover,theoutputelasticityoflaboris significantlydifferentfromzeroonlyforCoopswithcatch-up com-ponentequal toone. Thehighvalueof theoutput elasticityof CoopsindicatesthatlaborismoreimportanttoCoopsthantoIOFs inproducingoutput.TheimportanceoflabortoCoopsvis-a-vis IOFsmaysuggestthatCoopshavefewbuthighlyqualifiedlabor. Additionally,thevalueoftheoutputelasticityintermsoflaboris negativeforthesevenIOFswithcatch-upcomponentequaltoone, thisimpliesthataslaborincreases,theoutputdecreasesforthese IOFs.
Also,technicalchangeofCoopsandIOFsisnegative,which indi-catesthat, thetechnologyof dairyprocessing firmsonaverage getsworseover thetimeperiod underconsideration.However, theCoops’technologydecreasesmorequicklyovertimethanthe IOFstechnology.Also,IOFswithacatch-upcomponentlessthan one(CUI<1)havethelowesttechnicalregress,implyingthatthese IOFs(althoughnothighlyefficientandusinginferiortechnology) arerelativelybetteroff.Technologicalregressisunlikelytooccur duetoamoreinferiortechnologyovertime.Amorelikelyreasonfor technologicalregressisthetighteningofgovernmentregulations regardinge.g.foodsafetyandpackagingrequiringdairyprocessing firmstotakecostlyadditionalmeasures.
Finally,resultsinTable5showthatfirmsofbothtypeshave pro-ductiontechnologiesthatarecharacterizedbydecreasingreturns toscale.TheCoops’technologyisfacingamorepronounceddegree ofdecreasingreturnstoscalethanIOFs.Furthermore,resultsshow thatIOFswithcatch-upcomponentequaltounity(CUI=1)have
largerreturnstoscalethanfirmswithcatch-upcomponentsmaller thanunity.These findingsimply thatanincrease in thesize of operationhasamorepositive effectonIOFsratherthanCoops, and may be a direct consequence of the larger size of Coops
and theirobligation toprocessallmembers’ productionof raw milk.
5. Conclusions
Thisstudycontributestotheliteratureinusingtwodifferent frontierstocompareandmeasurethetechnicalefficiencyand pro-ductiontechnologyofdairyprocessingcooperativesandIOFsinsix Europeancountries.Themethodologicalapproachusesadifferent frontierforeach typeoffirms,allowingformeasuringtechnical efficiencyofthefirmwithrespecttoboth:itsowntypefrontier (Intra-typeefficiency)andthefrontieroftheothertype(Inter-type
efficiency),andthedifferencebetweenthetwofrontiers(Catch-up
component).Thecooperatives,onaveragehavealowerIntra-type
efficiencythanIOFs.However,thecooperatives,onaveragehave a higherCatch-up componentand a higher Inter-typeefficiency thanIOFs.ThesuperiorityofthecooperativesintheirInter-type
efficiencymeasurementreflects,inadditiontothephysical pro-ductivity,themarketingefficiencyasaresultofnormalizingthe outputsandinputs,ofcooperativesandIOFs,withthesameprice indices.
TheCatch-upterm differsacrosscountries. Thisis explained bydifferencesincooperativescharacteristicsandmarket condi-tionwhichimplydifferencesintheinternalincentives.InIreland thesuperiorityofthecooperatives’technologycomparedtoIOFsis thehighestwhichisexplainedbythreedifferentaspects;thefact thatIrishcooperativeshavealegalformtermedaspubliclimited company,thefragmentednatureoftheIrishprocessingsectorand theproximitytotheUKmarket.SevenIOFsareexceptionalwith
Catch-upcomponentequaltoone.Theyhavealargerturnoverand quantityofrawmaterials,whiletheyhavesmallerfixedassetsand laborthantheotherIOFs.Moreover,thesesevenIOFsarenotonly themostefficientwhencomparedtotheirowntechnology,they alsohavethehighestoutputelasticityintermsoffixedassetsand rawmaterial,andtheleastdecreasingreturningtoscaleproduction technology.OtherIOFswouldbenefitfromreviewingthe charac-teristicsof thesesevenIOFsandimprovingtheirowntechnical efficiencywithrespecttoboth;thetechnologyoftheIOFsandthe oneofthecooperatives.Thecooperatives’superiorityhasastrong impactonthefuturestructureofthedairymarketinthesesix Euro-peancountries,i.e.cooperativesmayincreasetheirmarketshare inthenearfutureattheexpenseoftheIOFs.InDenmark, cooper-ativesalreadyaccountfor97percentofthetotalmilkproduction, intheNetherlandsmorethan85%.
Futureresearchwouldbenefitfromdetaileddataonthe qual-ityandsizeofthecompositionofinputsandoutputsinthedairy processingfirmsanalyzedinthisstudy.Havingthisdetailed infor-mationwouldallowforabetterrepresentationoftheheterogeneity attributabletoinputandoutputcomposition.Atpresent,the cur-rentlyavailabledatadoesnotallowforfurtherinvestigatingthe potential impact of output heterogeneity or input qualities on performance.Also,futureresearchshouldfocusonanalyzingthe socio-economicandenvironmentalfactorsthatexplaindifferences
ininefficiencybetweendairyprocessingcompaniesacrossthe dif-ferentcountries.Finally,anextensionofthisanalysiswouldbeto includecountrieswithdifferentcooperatives’structureand busi-nessmodelsacrosscentralandeasternEurope.
AppendixA. Supplementarydata
Supplementarymaterialrelatedtothisarticlecanbefound,in theonlineversion,athttp://dx.doi.org/10.1016/j.njas.2014.05.003.
References
[1]R.A.M.E.Soboh,A.OudeLansink,G.Giesen,G.VanDijk,Performance Measure-mentoftheAgriculturalMarketingCooperative:TheGapBetweenTheoryand Practice,ReviewofAgriculturalEconomics31(4)(2009)446–469.
[2]R.J.Sexton,J.Iskow,FactorsCriticaltoSuccessorFailureofEmerging Agri-culturalCooperatives,DepartmentofAgriculturalandResourceEconomics, UniversityofCalifornia,Davis,1988(No.88-3).
[3]P.Bontems,M.Fulton,Organizationalstructureandtheendogeneityofcost: cooperatives,for-profitfirmsandthecostofprocurement,InstitutNationalDe LaRechercheAgronomique-Unitéd’EconomieetSociologieRurales,2005. [4]R.G.Hamermesh,P.W.Marshall,T.Pirmohamed,NoteonBusinessModel
AnalysisfortheEntrepreneur,HarvardBusinessSchoolPublishing,2002 (9-802-048).
[5]K.Oustapassidis,A.Vlachvei,K.Karantininis,GrowthofInvestorOwnedand CooperativesFirmsinGreekDairyIndustry,AnnalsofPublicandCooperative Economics69(3)(1998)399–417.
[6]P.K.Porter,G.W.Scully,EconomicEfficiencyinCooperatives,JournalofLaw andEconomics(30)(1987)489–512.
[7]G.D.Ferrier,P.K.Porter,TheProductiveEfficiencyofU.S.MilkProcessing Coop-eratives,JournalofAgriculturalEconomics42(2)(1991)161–173.
[8]A.Gentzoglanis,EconomicandFinancialPerformanceofCooperativesand Investor-OwnedFirms:AnEmpiricalStudy,in:J.Nilsson,G.vanDijk(Eds.), Strategiesandstructuresintheagro-foodindustries,VanGorcum,Assen,1997, pp.171–183.
[9]P.Schmidt,Frontierproductionfunctions,EconometricReviews4(2)(1986) 289–328.
[10]P.Bauer,Recentdevelopmentsintheeconometricestimationoffrontiers, Jour-nalofEconometrics46(1–2)(1990)39–56.
[11]G.Battese,FrontierProductionFunctionsandTechnicalEfficiency:ASurvey ofEmpiricalApplicationsinAgriculturalEconomics,AgriculturalEconomics7 (1992)185–208.
[12]R.J.Sexton,B.M.Wilson,J.J.Wann,SomeTestsoftheEconomicTheoryof Coop-eratives:MethodologyandApplicationtoCottonGinning,WesternJournalof AgriculturalEconomics(14)(1989)55–66.
[13]H.Doucouliagos,P.Hone,TheEfficiencyoftheAustralianDairyProcessing Industry,TheAustralianJournalofAgriculturalandResourceEconomics44 (3)(2000)423–438.
[14]S.Singh,T.Coelli,E.Fleming,PerformanceofDairyPlantsintheCooperative andPrivateSectorsinIndia,AnnalsofPublicandCooperativeEconomics72 (2001)453–479.
[15]R.Mosheim,OrganizationalTypeandEfficiencyintheCostaRicanCoffee ProcessingSector,JournalofComparativeEconomics30(2)(2002)296–316. [16]A.OudeLansink,E.Silva,S.Stefanou,Inter-FirmandIntra-FirmEfficiency
Meas-ures,JournalofProductivityAnalysis,200115(3)(2001)185–199. [17]D.Aigner,C.A.K.Lovell,P.Schmidt,FormulationandEstimationofStochastic
FrontierProductionFunctionModels,JournalofEconometrics6(1977)21–37. [18]T.Coelli,EstimatorsandHypothesisTests foraStochastic Frontier Func-tion: AMonte Carlo Analysis, Journal of Productivity Analysis 6 (1995) 247–268.
[19]M.J.Farrell,TheMeasurementofProductionEfficiency,JournalofRoyal Statis-ticalSocietyseriesA(120)(1957)253–290.
[20]C.A.K.Lovell,P.Schmidt,Acomparisonofalternativeapproachestothe mea-surementofproductiveefficiency,in:A.Dogramaci,R.Fare(Eds.),Applications ofModernProduction:TheoryEfficiencyandProductivity,KluwerAcademic Publishers,Norwell,MA,1988.
[21]A.OudeLansink,E.Silva,S.Stefanou,DecomposingProductivityGrowth allow-ingefficiencygainsandpriceinducedtechnicalprogress,EuropeanReviewof AgriculturalEconomics.27(4)(2000)497–518.
[22]Eurostat(2014).epp.eurostat.ec.europa.eu.accessedon10March2014. [23]J.V. Edward, W.D. Dobson, G.C. Frank, N.F. Olson, The Dairy Sector of
Ireland:ACountryStudy,BabcockInstituteDiscussionPaper,2007,Foundat: http://purl.umn.edu/37334(2007-2).
[24]O.Ebneth,L.Theuvsen,InInternationalizationandFinancialPerformanceof Cooperatives-EmpiricalEvidencefromEuropeanDairySector,in:Paper pre-sentedatthe15thAnnualWorldFoodandAgribusinessSymposiumandForum, Chicago25th-28thJune,2005.
[25]W.Meeusen,J.VandenBroeck,EfficiencyestimationfromCobb-Douglas pro-ductionfunctionswithcomposederror,InternationalEconomicReview18(2) (1977)435–444.
[26]R.A.M.E.Soboh,A.OudeLansink,G.vanDijk,EfficiencyofCooperativesand InvestorOwnedFirmsrevisited,JournalofAgriculturalEconomics63(1)(2012) 142–157.
[27]R.Carson,ATheoryofCooperatives,CanadianJournalofAgriculturalEconomics 10(4)(1979)565–589.
[28]R.J.Sexton,TheFormationofCooperatives:AGame-TheoreticApproachwith ImplicationsforCooperativeFinance,DecisionMaking,andStability.American JournalOfAgriculturalEconomics68(2)(1986)214–226.