TheDemandforDairyProductsinTurkeyTheImpactofHouseholdCompositiononConsumptiondoc.docx

TheDemandforDairyProductsinTurkeyTheImpactofHouseholdCompositiononConsumptiondoc.docx

《TheDemandforDairyProductsinTurkeyTheImpactofHouseholdCompositiononConsumptiondoc.docx》由会员分享，可在线阅读，更多相关《TheDemandforDairyProductsinTurkeyTheImpactofHouseholdCompositiononConsumptiondoc.docx（19页珍藏版）》请在冰豆网上搜索。

TheDemandforDairyProductsinTurkeyTheImpactofHouseholdCompositiononConsumptiondoc

TheDemandofDairyProductsinTurkey:

TheImpactofHouseholdCompositiononConsumption

A.AliKoç

ResearchConsultant

AgriculturalEconomicsResearchInstituteAnkara,e-mail:

aakoc@

SibelTan

AgriculturalEconomist,

AgriculturalEconomicsResearchInstituteAnkara,e-mail:

sibeltan@aeri.org.tr

Türkiye’deSütÜrünleriTalebi:

HanehalkıNüfusYapısınınTüketimÜzerineEtkisi

Özet

Buçalışmadafarklısütürünleriiçinfarklımodeltanımlanarak（Working-Leser,AIDSveÇift-Logaritmikfonksiyon）Engelfonksiyonutahminedilmiştir.Çalışmadayeralansüt,yoğurt,peynirvetereyağiçinharcamaesnekliklerihespalanmıştır.Peynirvesütiçinfiyat-talepesneklikleridetahminedilmiştir.Çalışmailehanehalkınüfusyapısınınpeynirvesüttüketimüzerineetkisidebelirlenmiştir.Sonuçlarhaneyeyenibirbireydahilolduğundapeynirvesüttüketimharcamasınınazalacağınıveyanegatifetkileneceğinigöstermektedir.Bunegatifetkibireyinyaşıiledoğruorantılıolarakartmaktadır.

Abstract

ThisstudyestimatesdifferentEngelCurves（intheformsof“Working-Leser”,AIDSwithunitvalue,andDouble-Log）fordifferentdairyproductsinTurkey.Studyprovidesexpenditureelasticityforfourdairyproducts（milk,yogurt,cheeseandbutter）.Own-priceelasticityisalsocalculatedforcheeseandbutter.Householdcompositioneffectsoncheeseandmilkconsumptionaredeterminedbythestudy.Resultsindicatethatadditionofanextrapersontoahouseholdhasnegativeimpactonpercapitacheeseandmilkexpenditure.Thisnegativeimpactgetsbiggerwithage.

KeyWords:

RepublicofTurkey,DairyProductDemand,HouseholdCompositionEffects,Working-Leser,andAIDSwithUnitValue

1.Introduction

2.

CommonlyusedfooddemandprojectionmethodisthesyntheticDouble-logspecificationthatemploysincomeelasticityandpopulationgrowthforfooddemandprojection,particularlyindevelopingcountries.Thismethodmaygeneratebiasedfooddemandprojectionsandpolicyresults,ifpopulationcompositionandotherdemandshifters（relativecommodityprice,urbanization,education,etc.）arealsochangedwithrapidly.ThisisthecurrenttrendinTurkeyasinmanymiddleincomedevelopingcountries.Ontheotherhand,intelligentpolicydesignforindirecttaxationandsubsidiesrequiresknowledgeofpriceandincomeelasticitiesfortaxablecommodities（Deaton,1988）.Suchknowledgewouldnormallybeobtainedbytheanalysisoftime-seriesdataforaggregatedemand,prices,andincomes.

Unfortunately,inTurkey,aswellasinmanydevelopingcountries,time-seriesfooddisappearancedataisnotreadilyavailabletoeconomistsoranalysts;however,manydevelopingcountriesregularlycollecthigh-qualityhouseholdsurveydataonexpendituresandquantitiespurchasedforawiderangeofcommodities.Inprinciple,thesehouseholdsurveyscontaininformationaboutthespatialdistributionofpricessothat,ifthisinformationcouldberecoveredintheusableform,thereisgreatpotentialforestimatingthedemandresponsesrequiredforpolicymaking（Deaton,1988）.Ifunitvalues,obtainedbydividingexpenditurebyquantity,areadjustedwithqualitydifferences,thenthisdatapermitstheestimationoffooddemandatdisaggregatedlevels,whichisofinteresttopublicpolicymakers,agribusinessindustries,andproducerorganizations.

InTurkey,theStateInstituteofStatistics（SIS）conductedlarge-scalehouseholdconsumptionexpendituresurveysin1979,1987and1994.Unfortunately,SISdidnotmakethesurveyresultsavailabletousersatindividualhouseholdlevel.Thepublishedformoftheconsumptionexpendituresurveydataisaggregatedatincomepercentiles.

Thisstudyestimatesahouseholddairyproductdemandatadisaggregatedlevelwithhouseholdcompositionvariablesandqualityadjustedunitvaluesusingdatafromthe“1994HouseholdConsumptionExpenditureSurveyResultsforSelectedProvinceCenters”（SIS,1997）.

2.Model

AformoftheEngelcurve,whichhasperformedwellintheempiricalanalysisofcross-sectiondata,expressesbudgetshareasafunctionofthelogarithmofincome（YoungandHamdok,1994）:

（1）

whereWiisthebudgetshareoftheithgoodinfullincome,andyisthehousehold’sfullincome;iandiareparameterstobeestimated.Thisform,oftenknownasthe“Working-Leser”curve,isconsistentwiththeAlmostIdealDemandSystemwhenpricesareconstant（ChesherandRees,1987）.

Deaton（1997,p.231）;YoungandHomdok（1994）introducedhouseholdsizeandhouseholdcompositionbyre-defininghouseholdincomeinpercapitatermsandbyre-specifyingtheintercepttermtoallowfortheinfluenceofhouseholdcompositionasfollow:

（2）

wherendenoteshouseholdsize,andthreehouseholdmembertypearedistinguished（k=1,2,and3）,n1thenumberofchildrenlessthan12yearsold,n2thenumberofteenageragedbetween12and17yearsandn3thenumberofadultaged18andover;ikiandidenoteparameterstobeestimated.Inequation

（2）,thehouseholdcompositionvariablesactasexplicitdemandshifters.Householdsize（n）entersasaseparateexplanatoryvariable（inlogform）,aswellasinpercapitaincometerm.Thisistoensurethatthewayinwhichincomeaffectsbehaviourisunrestricted（YoungandHomdok,1994）.Equation

（2）mayfurtherbeimprovedbymeansofintroducingpriceterm（hereisunitvaluethatisobtainedbydividingexpenditurebyquantitypurchasedofthegoods）:

（3）

whereP*istheunitvalueoftheithgoodthatneedstobeadjustedifaggregatequantityofthegoodsiscomposedfromheterogeneousproduct.

Oneproblemassociatedwiththemodeldescribedbyequation（3）isthattheunitvalue,obtainedbydividingexpenditurebythequantitypurchased,isnotadirectsubstitutefortheactualmarketprice.Unitvaluesnotonlyreflectspatialvariationinpricesduetothetransportcostdifferentials,theyalsoreflectconsumerqualitychoicesintheirpurchasesanderrorsinmeasuringexpendituresandquantities（Deaton,1988）.Ifunitvaluesareuseddirectlyindemandestimation,thepriceelasticitiesarenotstandardelasticitiesofdemand.Theyalsoreflectqualityelasticitiesofdemand（Theil,1952;Houthakker,1952;Cramer,1973;CoxandWohlgenant,1986;Deaton,1987,1988and1990;Nelson,1991;Parketal.,1996;andDongetal.,1998）.

Thecausesofcross-sectionalpricevariationmustbeidentifiedinordertointerpretcorrectlytheeffectsofpricesintheanalysisofhouseholdbudgetdata（PraisandHouthakker,1955,p.110）.Polinsky（1977）pointedoutthatfailuretospecifycross-sectionalpriceeffectsadequatelycouldresultinbiasedandmisleadingdemandelasticities.Thus,traditionalEngelanalysismaybeinappropriateifthepricesfacedbyindividualconsumersarenotconstant.AccordingtoPraisandHouthakker（1955）,pricevariationsacrossregionsmaybeduetopricediscrimination,servicesbundledwiththecommodity,seasonaleffects,andqualitydifferencescausedbytheheterogeneouscommodityaggregate.

Theopportunitycostsofconsumers'time,themarginalcost/benefitofinformationsearch,retailingstrategies,andbrandloyaltymayalsocausecross-sectionalpricedifferences.Amongtheabovefactors,qualitydifferencescausedbyheterogeneouscommodityaggregatesmaybemoreproblematicalintheestimationofdemandfunctions（CoxandWohlgenant,1986）.Qualityeffectsincross-sectionalpricevariationresultmainlyfromcommodityaggregation（Houthakker,1952）.Thepotentialdistortionfromnotadjustingcross-sectionalpricesforqualityeffectswillincreasewithheterogeneityofthecommodityaggregate（Cramer,1973）.

Thesimplesumofphysicalquantitiesusedasthedemandinthequalityliteratureisatheoreticallyarbitrarymethodofaggregationandispotentiallyamisleadingmeasureofdemandwhengoodsareheterogeneous（Nelson,1991）.AccordingtoNelson,theimportanceofproperlyadjustingforqualityvariationdependsontheimportanceofqualityeffectsinthedataunderexamination.

Thisstudyadjuststheunitvaluesintermsofincomeandhouseholdsize.Theunitvaluesoftheaggregatedcommodityareestimatedusingthefollowingequation（CoxandWohlgenant,1986;Parketal.,1996）.

（4）

where

isunitvalueoftheithaggregatedcommodity,yishouseholdfullincome,andnisthehouseholdsize.Itiscommonlyassumedthattheintercepttermofthehedonicpricefunctionreflectsthequantityprice.Ifweassumetheaveragesamplepriceistheintercepttermofthehedonicpricefunction,thentheadjustedunitvaluecanbeobtainedfromequation（4）.

（5）

Inequation（5）

istheaveragesampleunitvalueandisresidualfromequation（4）.

IncomeandMarshallianpriceelasticitiesfromtheestimatesofequation（3）arecomputedusingfollowingformulas（GreenandAlston,1991）.

Expenditure

（6）

Price

（7）

Asthewayinwhichchangesinfamilycompositionaffectdemandisquitecomplex（theadditionofafamilymemberoftypejincreasesnaswellasnj）,theparametersinequation

（2）aredifficulttointerpretdirectly.Following（Chesher,1991）foreachcommoditygroup,theimpactonhouseholdexpenditureoftheadditionofahouseholdmemberoftypertothehousehold,“ceterisparibus”,maybecalculatedasfollows（YoungandHamdok,1994）.

（8）

where

denotesthechangeinthebudgetshareofgoodi（orequivalently,thechangeinexpenditureiasaproportionofhouseholdincome）.Itmeasuresthe“totaleffect”ofachangeofhouseholdcomposition,i.e.thecombinedimpactofthe“specificeffects”and“incomeeffect”ref