We explain such differences by focusing on political parties57557 access to resources supply side and voters57557 dependence on 64257scal largesse demand side We show how these differences affect the patron57557s choice of public sector wages and em ID: 75243
Download Pdf The PPT/PDF document "Who Delivers Partisan Clients in the Arg..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
744ERNESTOCALVOANDMARIAVICTORIAMURILLO1PublicandPrivateSectorWagesandYearsofEducation(Argentina1997) Note:Fittedcubicregressionlinesofwagebyeducation,forbothpublicandprivateworkers.EstimatedfromSiempro(1997)NationalHouseholdSurvey,18.643validobservations.partisanpatrons,andtheyshouldhavedifferentstrategiesofpatronagetomaximizetheirelectoralreturns.Thesupply-sideadvantagestopatronagearethere-sultofthepartiesdifferentaccesstopublicfunds.Becauseanyclientelisticpartypreferslargerbudgets,supply-sidedifferencesemergefrombiasesintheinstitutionalmecha-nismsthatregulatetheincumbentsaccesstopublicfunds.Weshowhere,forthecaseofArgentina,thatdifferentelectoralandfiscalrules(MonroeandRose2002)gen-eratepartisanbiasesinthesubnationaldistributionofpublicfunds.Onthedemandsideofpatronage,theabsenceofcivilservicerulesallowsthediscretionaryuseofpublicem-ploymentasadistributivemechanismthatrewardspublicemployeeswithawagepremiumabovetheirlikelyprivatesectorwages.Becausethispremiumcanbetargetedtogroupsofvotersendowedwithdifferentskills,factorsthataffecttheexpectedincomeormarketalternativesofdifferentgroupsofvoterswillalsoshapetheeffectivenessofpatronage-drivenelectoralstrategies.Forinstance,so-cioeconomicdevelopmentisusuallyassociatedwithin-creasesintheskilllevelsofthelaborforce,therebylower-Wedefinethewagepremiumasthedifferencebetweenthepublicsectorwageandtheprivatesectorwageforgroupsofvoterswithequivalentskills.ingthereturnsthatpatronscanderivefromredistributionthroughthediscretionaryallocationofpublicjobs.Figure1providesanexampleofthisargumentbyplottingindividual-levelsurveydataonwages(privateandpublic)andyearsofeducation(asaproxyforskilllevels)inArgentina.Asshownbythesolidline,theme-dianpublicwageissignificantlyabovethemedianprivatewageforthelow-skilledemployeesandslightlybelowtheprivatesectorwageforhigh-skilledemployees.Asaresult,althoughthemedianprivatewageisaround370pesosandthemedianpublicwageisaround500pesos,this33%wagepremiumisunequallydistributedamongworkersTherefore,wemayexpectpatronagetoincreaseinhighlyunequalmiddle-incomedemocracies,incontrastwithbothlow-incomecountries,inwhichthereislittleredistribution,andconsolidateddemocraciesindevelopedcountries,wherethereislimitedparti-sancontrolovertheallocationofrelativelycostlypublicjobstocomparativelyskilledworkers.Thereisevidence,however,thatre-distributionthroughpublicemploymentisstillquitesignificantinsomedevelopedcountries,suchasSweden(IversonandWren1998;IversonandSoskice2002)andItaly(Alesina,Danninger,andRostagno2001),andthatpublicemployeesprovideelectoralre-turnstothepartieswithredistributivepolicypreferences.How-ever,eveninthosecasesthepatronsdiscretionhasgenerallybeenreplacedbyuniversalbudgetarymechanismswhileconstituenciesdependencehasbeenalleviatedbywelfareentitlements. 746ERNESTOCALVOANDMARIAVICTORIAMURILLOargumentonpartisanpreferencesfordifferentdistribu-tivepathspublicsectorsizeandwagederivedfromthebudgetconstraintsfacedbypoliticiansoveralimitedrangeoffeasiblevoterstowhichtheyhavepriorities.SubnationalPoliticsandPatronageinArgentinaArgentinas1853constitutionestablishedafederalandpresidentialrepublicwithabicameralCongress.Thiscon-stitutionaldesignwasrelativelystableuntilthefirstmili-tarycoupin1930.Afifty-yearperiodofpoliticalinstabil-ityensued,whicherodedthestatescapacitytogeneratestablesocialexpectations.Intheabsenceofacivilservice,thisinstabilityledtohighlevelsofrotationintheupperbureaucraticandjudicialcadres(Iaryczower,Spiller,andTommasi2002).Duringthesecondhalfofthetwentiethcentury,Argentinaspartysystemwasbasedontwopoliticalcoali-tionsledbytheUnionCvicaRadical(UCR)andthePero-nistParty(PJ).TheUCRwasacentristpartythatemergedfromtheurbanmiddleclassesinthe1890s,whilethePJemergedinthe1940sfromacoalitionofurbanworkersinthemostdevelopedareasofthecountryandlocalbossesinthemostruralprovinces.TheUCRwonallpresiden-tialelectionsbetweentheelectoralreformof1912,whichmadeuniversalmalesuffrageeffective,andthe1930coup.ThePeronistswonallsubsequentelectionsinwhichtheywereallowedtorununtil1983.Thecombinationofalackofeffectivecompetitionandpoliticalinstabilitypreventedeitherofthetwopartiesfromestablishingdistributivestrategiesinacompetitivecontextuntil1983.Thereturntodemocracyin1983markedthebe-ginningofaperiodofdemocraticstabilitywithcom-petitiveelectionsandeffectivepoweralternationsinthenationalexecutive.Intheperiodbetween1983and2001,boththePJandtheUCR(withallies)wontwopresiden-tialelectionseach.Regardlessofwhowonthepresidency,PeronistscontrolledtheSenate,amajorityofgovernor-ships,andusuallyamajorityofmunicipalgovernmentsandapluralityofseatsintheLowerHouse.Bycontrast,non-Peronistpresidentswhohadgatherednationalma-joritiesnevercontrolledtheSenateandonlybrieflythepluralityofseatsintheLowerHouse(Table1).Duringtheirfirstthreeadministrations(191630),theRadicalsdidnotestablishanylargeprogramofsocialpolicyredistributionbutenlargedthepublicsectorandopenedittothemiddleclasses.Incontrast,duringtheirfirsttwoadministrations(194655)thePeronistsnotonlyexpandedthestatebutalsocommittedtoadra-maticredistributionthroughsocialpolicyandlabormarketregu-lations(McGuire1997).1PartyofthePresident,HouseandSenateSeats,andGovernorships 1983Ð19851985Ð19871987Ð19891989Ð19911991Ð19931993Ð1995 PartyofthePresidentUCRUCRUCRPJPJPJPJPJUCR-FrepasoPeronistDiputados(%)43.738.438.447.148.649.851.446.638.82AlianzaDiputados(%)50.650.644.735.332.932.935.341.647.0PJSenators(%)45.645.645.661.261.261.255.755.755.7UCR-AlianzaSenators(%)39.143.543.526.121.721.728.628.628.6PeronistGovernorships52.252.273.973.958.358.358.358.362.5AlianzaGovernorships30.4330.438.78.716.716.720.820.829.16PeronistMayors(4000inhabitants)43.343.355.955.955.755.758.958.951.5UCRMayors(4000inhabitants)45.845.832.832.830.430.428.828.836.2 Sources:DireccionNacionalElectoral,MinisteriodelInterior;JonesandHwang(2003).Thenumberofsenatorsincreasedfrom48to72aftertheconstitutionalreformof1994.Alianzadiputadosafter1997includeseatsfromtheUCRandtheFrepaso. 748ERNESTOCALVOANDMARIAVICTORIAMURILLOnumberofpartiesreceivehalfoftheLowerHouseseatsinoverrepresentedmajoritariandistricts,whilealargenumberofpartiesreceivehalfoftheLowerHouseseatsinunderrepresentedproportionaldistricts.Indeed,over-representationisevenworseinaSenatewithequalrep-resentationperprovinceregardlessofpopulation(twoSenatorsuntil1994andthreeafterwards).ThegeographicconcentrationofPJvotersinless-populated,andmoreoverrepresented,provincesoverlapswiththemajoritarianbiasproducedbytheelectoralsys-tem,therebyfavoringthispartyinthedistributionofseatsintheLowerHouse,theSenate,andprovincialgovernor-ships.Therefore,evenwhenthePJdidnotgatherna-tionalmajoritiestowinthepresidency,itobtainedmoregovernorshipsandthusaccesstofiscalresourcestheRadicals.Inadditiontotheeffectproducedbytheelectoralbiasonsubnationalincumbency,fiscalinstitu-tionsfurtherreinforcedpartisandifferencesinaccesstoresourcesasdiscussedbelow.FiscalFederalismandPartisanBiasinAccesstoResourcesThegeographicdistributionofthePeronistvoteincombi-nationwiththedistributiveeffectsofArgentinefiscalfed-eralismprovidesanadvantageforPeronistpoliticiansinobtainingfiscalresourcestoinvestinpatronageThispar-tisanbiasproducedbythecombinationofinstitutionaleffectsandthePJconcentrationofthevotecannotbeac-countedforbyotherfactorsaffectingfiscalredistribution,suchaspopulation,provincialincomeredistribution,orevenelectoraloverrepresentation.Since1934,theArgentineprovincesdelegatedfis-calauthoritytothefederalgovernmentforlevyingandcollectingtaxes.Fiscalrevenuesweredividedbetweenthefederalandprovincialgovernmentsusingarevenue-sharingformulasetbyCongress,whichovertimein-creasedtheprovincialshareofresourcesattheexpenseofthefederalsharetothepointthatbothshareswerealmostequalby2001.ThisdistributionfavoredtheSince1991,Argentinahas24provinces,whichserveaselectoraldistrictsforthe257membersoftheLowerChamberandreceiveanumberofdeputiesinproportiontopopulation,providedthatnodistrictreceivesfewerthanfivedeputiesorfewerlegislatorsthanitsshareduringthe197376democraticperiod.Asaresult,sixprovincesintheleastpopulousquartilehave3.9%ofthepopula-tion,11.7%oftheseatsintheLowerChamber,and25%ofseatsintheSenate.Incontrast,theprovinceofBuenosAires,whichhas39%ofthepopulation,onlyreceives27%ofdeputiesand4%ofSenators.Priorto1991,TierradelFuegowasnotaprovinceandtheCityofBuenosAireswasafederaldistrict.Between1980and2001,theprovincialexpendituresdiscountingpublicdebtgrewfrom8%to13%oftheGDPwhilefederalexpendituresdecreasedfrom17%to13%oftheGDP(Min-istryoftheEconomy2002).Peronistpoliticianswhocontrolledmoreprovincialgov-ernorshipsthantheirRadicalscounterparts.ThePeronistsadvantageinobtainingaccesstofiscalresourcesisnotonlybasedontheirlargerelectoralsuccessattheprovinciallevelbutalsointhefactthatthedistribu-tionoffiscalrevenueisbiasedtowardPeronist-controlledprovinces,eventakingintoaccountotherfactorsinflu-encingrevenue-sharing.Usingapooledcross-sectionaldatasetofeconomicandpoliticalprovincialindicatorsweshowthatPeronist-dominatedprovincesareabletoextractlargerfederalresourcesthantheirUCRcounter-parts,evenwhencontrollingfordifferencesinincomeandpopulationintheredistributivecomponentoftherevenue-sharingformula,aswellasfortheoverrepresen-tationgeneratedbytheelectoralsystem.Ourdatasetincludescross-sectionsofthe24provincesandtheCityofBuenosAiresfortheyearsof1987,1990,1995,and2000.TwoyearswereunderUCRpresidencies(1987and2000)andtwoyearsunderPeronistpresidencies(1990and1995).Inordertotestfortheexistenceofapartisanbiasintheallocationoffiscalresources,weanalyzetheef-fectofthePeronistandUCRvoteshareonourdependentvariables:Shareofexpendituresnancedbythefederalgovern-ment:Describesthepercentofprovinceistotalex-pendituresfinancedbybothrevenuesharingandotherspecialtransfersfromthefederalgovernment.Therelativeshareofscalresourcesrevenuesharingratio)receivedbyeachprovince:Measuresprovincesshareofthetotalfederalresourcesovertheirpop-ulationshare.OurindependentvariablesaretheUCR/AlianzaandPeronistforprovince,andweusetheusualcon-trolsforexplainingrevenuesharingandfederalfinancing:Anincome-adjustedrevenue-sharingformulacanbeanindica-tionofPeronistlegislativestrength.Althoughcontrollingforsuchendogeneityshouldprovideevenstrongerresultsthanthosepre-sentedinthissection,theconservativeestimatesdisplayedhereshowapartisaneffectevenwhencontrollingforthemaincompo-nentsoftherevenuesharingformula.Between1997and2001,theUCRmadeanelectoralalliancewiththecenter-leftFREPASO,forthatreasonwehavepooledtheirvotestogetherforthe1995electiontheyearaftertheestablishmentofFREPASO.Themeasurementsoffederalgovernmentfinancingandrelativerevenue-sharingratioaresimilartothoseusedbyGibsonandCalvo(2000)andRemmerandWibbels(2000).TheeffectivenumberofcompetitivepartiesinArgentinarangesfrom2.7to3.3,withrelativelyhighlevelsofcompetitionin1991,1995,andafterthepresidentialcrisisof2001.Wecontrolledfortheresidualcategoryotherpartiesandfoundnosignificantdiffer-ences.Furtherstatisticalresultsareavailablefromtheauthorsuponrequest. 750ERNESTOCALVOANDMARIAVICTORIAMURILLOandtheterritorialdistributionofvotersprovidePeronistswithapartisanadvantageinaccesstofiscalresourcesthatcanbedeployedforpatronage.Thissupplyeffect,though,isindependentfromthedemandforredistributionwhichexplainsthehighervote-per-buckthePJreceivesfrompa-tronagespendingduetoitsvotersdependenceonpubliclargesseasdiscussedinthenextsection.PartisanAdvantagesintheEfÞciencyofPatronageExpendituresWehavealreadyshowntheexistenceofpartisandiffer-encesinaccesstofiscalresourcesinArgentina,andwenowturntoourmainargumentaboutthepartisanvariationintheelectoralreturnsreapedfrominvestingpublicfundsonpublicemployment(demandside).Weshowherethatduetothehigherdependenceoftheirconstituenciesonpubliclargesse,publicemploymentprovidesbetterelectoralreturnsforthePJthantheUCR-AlianzaintheArgentineelectoralmarket.PartisanReturnstoPatronageandConstituenciesDependencyonPatronageOnceagain,thissectionusesthecross-sectionalprovin-cialdatasetofTable2toexplaintheeffectofpatronageonthePJandUCRvote.Inthisanalysis,ourdependentvariablesarethepercentofcongressionalvotesbyPeronism(Model1)andtheUCR-Alianza(Model2)ineveryprovinceandyear.Theexplanatoryvariablesare:(1)IncumbentGovernor,adummyvariableindicat-ingwhetherthegovernorisPeronistornot(Model1)orUCR-Alianza(Model2);(2)IncumbentPresidentdummyvariableindicatingwhetherthepresidentisPer-onist(Model1)orUCR-Alianza(Model2);(3)MedianVoterIncome,describingtheprovincialmedianworkerincome,asreportedbytheArgentineCensusBureau(Indec)inprovinceandyear;(4)PublicEmploymentdescribingthenumberofprovincialpublicemployeesper1000inhabitantsofprovinceinyearPublicExpen-diturespercapita,describingthegrossprovincialexpendi-turepercapitainArgentinepesosforprovinceandyear;and(6)effectivenumberofcompetingparties,whichcon-trolsforthedifferentexpectedvotesharesunderdifferentprovincialcompetitionlevels.WemeasureitusingLaaksoandTaagepera(1979)index: ,whereistheshareofvotesforeveryparty.PublicEmploymentdatafromProvInfo(http://www.mininterior.gov.ar).AgainwepoolthevotesfortheUCRandFrepasoforthe1995electionandincludedummiestoestimatetheeffectofincumbencyforbothincumbentgovernorsandthepresident.AsintheestimatedmodelsofTable2,theresultsobtainedbyOLSweresimilartotherandomeffectmodels,sowedisplaythesimplermodels.Table3showsthatchangesinpublicemploymenthaveasignificantandpositiveeffectonthePJvotebutnotontheUCR-Alianzavote.Theeffectofpublicemploy-mentper1,000inhabitantshasapositiveandsignificanteffectforthePeronists.A1%increaseinprovincialpublicemploymentleadstoa.066%increaseinthePeronistvote.Therefore,weshouldexpectthatdoublingthenumberofprovincialpublicemployees,from5%oftheeconomi-callyactivepopulationto11%oftheeconomicallyactive3PublicEmploymentandElectoralReturns PJUCR-AlianzaVote(model1)Vote(model2) IncumbentGovernor775)(2IncumbentPresident725)(4MedianVoterIncome2)(3PublicEmploymentper1,000(LN)89)(2EffectiveNumberofCompetingParties45)(21995569503)(42000508)(2Constant7899)(24.72.57N8383 Note:IncumbentgovernorisadummyvariablethatindicatesaPeronistgovernorinthePJequationsandanUCR-AlianzagovernorintheAlianzaequation.IncumbentpresidentisadummyvariablethatindicatesaPeronistpresidentinthePJequationandaUCR-AlianzapresidentintheAlianzaequation.Publicemploymentisavariabledescribingthenaturallogofthetotalnumberofprovincialpublicemployeesper1,000citizensfor1987,1990,1995,and2000.Publicexpendituresdescribesthenaturallogofthetotalexpendituresinpesosforeveryprovinceineveryyearofthesample.EffectivenumberofcompetingpartiesisareducedversionoftheLaksoandTaageperaformulaincludingonlymajorparties. 752ERNESTOCALVOANDMARIAVICTORIAMURILLO3PublicandPrivateSectorWagesandYearsofEducationinPeronistandNon-PeronistProvinces(Argentina1997) Mean Public Wage Mean Private Wage Mean Public Wage Mean Private Wage Note:FittedcubicregressionlinesofwagebyeducationinPeronistandnon-Peronistprovinces,publicandprivateworkers.EstimatedfromSiempro(1997)NationalHouseholdSurvey,18,643validobservations.maximizebothredistributionandthetotalnumberofpublicemployeesbenefitingfromit.Yet,becausehigherrelativepublicsectorsalariesresultinfewerpublicsectoremployees,thereisabenefitincultivatinglow-skilledvot-ersandinmoderatingpublicwagesinfavoroflargerpub-licsectors.Thetrade-offbetweenwagesandemploymentshouldthereforegeneratepressureforwagemoderationinfavoroflargeremploymentinPeronistprovinces.Wetestthispartisaneffectonthesizeofthepublicsectorusingaggregate-leveldatatotesttherelationshipbetweenpublicsectorwages,publicsectorsize,andparti-sanship.Becausebudgetconstraintslimitthetotalbillforpublicsectorwages,controloverthenumberofemploy-eesisgivenbyregulatingpublicsectorsalaries.Hence,higherpublicsectorwagesshouldleadtoreductionsinthenumberofpublicemployeesandlowerpublicsec-torwagestoincreasesinpublicsectoremployment.Thestatisticalanalyses,therefore,focusesonpartisanchoiceoverpublicsectorwages,whichinturnleadtolargerorsmallerpublicsectorsize.Weuseagainthecross-sectionalpooleddatasetoftheprevioussection,andweincorporateinformationonprovincialpublicsectorwagesandprivateincomeaswellasthetotalnumberofprovincialpublicemployees.dependentvariablesarethetotalnumberofpublicem-ployeesper1000citizensandtherelativeprovincialpublicIncomedatafromthePermanentHouseholdSurveyofINDECandpublicwagesfromProvInfo(http://www.mininterior.gov.ar).sectorwage.Operationalizingourargumentisrelativelystraightforward,asweneedtoshowthatpublicemploy-mentdecreasesasafunctionofrelativelyhigherprovincialpublicsectorwages.Becauserelativepublicsectorwagesareafunctionoftheprovincesprivatesectorwageandthenationalpublicsectorwage,weoperationalizerela-tivepublicwagesas WhereRPW,relativepublicsectorwage,isthera-tioofthepublicsectorwage()inprovincetothenationalpublicsectorwage(),andtheprivatesectorwage()inprovincetothenationalprivatesectorwage).Consistentwithourargument,andhintedbyFig-ure3,higherlevelsofpublicsectorwagesleadtosmallerpublicsectors.Bycontrast,relativelyhighprovincialpri-vatesectorwagesleadtolargerpublicsectorsalariesinordertobecompetitiveandstillsatisfyredistribution.Becauseprovincialpoliticianscanonlysettheprovincialpublicwages,itisusefultopresenttheresultswithdistinct )andprivate )terms.Theexplanatoryvariablesforthenumberofpublicemployeeshavebeendescribedpreviouslyinexplainingtheprovincialbudgetequation:(1)percentofexpendituresThenaturallogoftheratioisusedtonormalizethedistributionoftheterms.ThestatisticalresultswiththefulltermofEquation1weresimilartothosepresentedwithseparateterms. 754ERNESTOCALVOANDMARIAVICTORIAMURILLO4PartisanChoicesonPublicSectorWagesandSize Model1,Model1,Model2,Model2,Equation1Equation2Equation1Equation2NumberofPublicRelativeNumberofPublicRelativeEmployees(per000)PublicWageEmployees(per000)PublicWage RelativePublicWageLn(uw/uw21)(5RelativePrivateWageLn(pw/pw)20054)(42RelativeCoparticipatedRatio(LN)2603)(6ProvincialPublicDeficit03)(3FederallyfinancedSpending(LN)69)(7Population(LN)47)(2PeronistSeatShare31)(PeronistGovernor14)(InteractionPJSeatsGov35)(UCRSeatShare19)(UCRGovernor36)(InteractionUCRSeatsGov74)(MedianPrivateWage07)(Constant19666)(46)(2734)(.84.53.84.53N83838383 Note:Model1estimatedby2SLS,withindependentlycorrelatederrors(RecursiveModel).Model2estimatedbySeeminglyUnrelatedRegression(SURS).constituencieswithhigherwagesthantheirprivatesec-torvalueand(b)trytomaximizeemploymentforlow-paidjobs.Politiciansthusactstrategicallyaccordingtotheirperceptionofpartisanreturnsfrompatronageandindoingsofurtherthedistributiveexpectationsoftheirtraditionalconstituencies.ComparativeImplicationsThisarticlebringstogetherpreviousworkondistribu-tivetaxation,partisanship,andclientelismtoexplainwhysomepartiesreceivelargerbenefitsfrompatronagespend-ing.Wedemonstratethatpatronagecanserveasadis-tributivemechanismtobenefitpartisanconstituencieswithhigherelectoralreturnsforpoliticalpartieswhoseconstituenciesaremoredependentonpubliclargesse(demandside).Wealsoshowhowinstitutionallydefinedterritorialbiasescanaffectaccesstopatronageifover-lappingwithgeographicconcentrationofelectoralsup-portforpoliticalparties(supplyside).Botheffectsareindependentbecausetheyimpacteitherthereturnsfrompatronageoraccesstopublicresourcestopayforit.WeillustrateourargumentinArgentinawherebotheffects 756ERNESTOCALVOANDMARIAVICTORIAMURILLOA1GroupedLogisticResultsofMajoritarianBiasedintheArgentineProvinces:CongressionalElections:PooledData1983 MetroPeripheralMetroPeripheralProvincesProvincesProvincesProvinces(large)(small)(large)(small) (.06)(.11)(.05)(.08)ENCP.16(.05)(.06).16.39.47(.20)(.22)(.17)(.16).13.44.52(.21)(.24)(.18)(.18)1991.03.17.54.45(.22)(.23)(.16)(.18)1993.001.15.48.46(.22)(.23)(.17)(.18).09.42.58(.21)(.24)(.17)(.18).19.32.42(.20)(.23)(.17)(.18)1999.01.20.46.42(.21)(.23)(.17)(.18)N7,2804,2797,2804,279 TableA1displaystheresultofEquationA1forthetwogroupsofprovinces:largemetropolitanprovinces(BuenosAires,CapitalFederal,Cordoba,andSantaFe)andforthesmallProvinces(allotherprovinces).Thevaluesofmodels1and2wherethenreplacedinEquationA1inordertoobtaintheseat-votelinesdisplayedinFigure1.1describesaperfectlyproportionalelectoralsys-temwhilelargervaluesof1,describemajoritarianbiasesandvalueslowerthan1,1,describesubma-joritarianbiases.ThedatausedtoestimateEquation(A1)includedallprovincelevelcongressionalelectionresultsfrom1985to1999.Thefirstyearofdemocracy,1983,waseliminatedbecausetherenovationofallseats,equivalenttodoublingalldistrictmagnitudes,leadstosmallermajoritarianbi-ases.Eachobservationinthedatasetincludedeverypartyinprovinceinyearvoteandseatshares.Dummyvari-ablesforeveryelectionyearwereintroducedtocontrolfortimespecificeffects.ReferencesAlesina,Alberto,andDaniRodrik.1994.DistributivePoli-ticsandEconomicGrowth.QuarterlyJournalofEconomicsAlesina,Alberto,andHowardRosenthal.1995.PartisanPol-itics,DividedGovernment,andtheEconomy.Cambridge:CambridgeUniversityPress.Alesina,Alberto,StephanDanninger,andMassimoRostagno.RedistributionThroughpublicEmployment:TheCaseofItaly.IMFStaffPapers48(3).Boix,Carles.2001.Democracy,Development,andthePublicSector.AmericanJournalofPoliticalScienceBrusco,Valeria,MarceloNazareno,andSusanStokes.2002.SelectiveIncentivesandElectoralMobilization.PresentedattheconferenceonPoliticalPartiesandLegislativeOr-ganizationinParliamentaryandPresidentialRegimes,YaleUniversity.Calvo,Ernesto,andJuanPabloMicozzi.2004.TheGovernorBackyard:aBayesianModeltoEstimatePartisanBiasinMultiparty(Subnational)ElectoralReformsPaperedpre-paredtobepresentedatthe62ndAnnualConferenceoftheMidwestPoliticalScienceAssociation,Chicago,IL,April1518,2004.Canton,Dario,andRaulJorrat.2002.EconomicEvaluations,Partisanship,andSocialBasesofPresidentialVotinginArgentina,1995and1999.InternationalJournalofPublicOpinionResearch4(Winter):413Catterberg,Edgardo.1989.Losargentinosfrentealapoltica.Culturapolticayopinionpublicaenlatransicionargentinaalademocracia.EditorialPlaneta:BuenosAires.Cox,Gary,andMatthewMcCubbins.1986.ElectoralPoliticsasaRedistributiveGame.JournalofPoliticsDiaz-Cayeros,Alberto,FedericoEstevez,andBeatrizMagaloni.PrivateversusPublicGoodsasElectoralInvestments:APortfolioDiversificationModelofPolicyChoice.Pre-paredfortheworkshopCitizen-PoliticianLinkagesinDemo-craticPolitics,DukeUniversity.Garrett,Geoffrey.1998.PartisanPoliticsintheGlobalEconomyCambridge:CambridgeUniversityPress.Gervasoni,Carlos.1998.EstructurayevoluciondelasCoali-cionesElectoralesenlaArgentina:1989Unpublishedmanuscript.FacultaddeDerechoyCienciasSociales,Uni-versidadCatolicaArgentina.Gibson,Edward,andErnestoCalvo.2000.FederalismandLow-MaintenanceConstituencies:TerritorialDimensionsofEconomicReforminArgentina.StudiesinComparativeIn-ternationalDevelopmentGibson,EdwardL.1997.ThePopulistRoadtoMarketReform:PolicyandElectoralCoalitionsinMexicoandArgentina.WorldPolitics49(April):339Iversen,Torben,andAnneWren.1998.Equality,Employment,andBudgetaryRestraint:TheTrilemmaoftheServiceEcon-omy.WorldPolitics50(July):507Iversen,Torben,andDavidSoskice.2002.ElectoralSystemsandthePoliticsofCoalitions:WhySomeDemocraciesRe-distributeMoreThanOthers.Presentedatthe2002An-nualMeetingoftheAmericanPoliticalScienceAssociation,Boston.Iaryczower,Matias,PabloSpiller,andMarianoTommasi.2002.JudicialDecision-MakinginUnstableEnvironments.AmericanJournalofPoliticalSciencesJones,MarkP.1999.FederalismandtheNumberofPartiesinArgentineCongressionalElections.JournalofPolitics59(May):538