DEMOCRATIZING-PARTY-LEADERSHIP-SELECTION-IN-SPAIN-AND-PORTUGAL- - -- Oscar-Barberà- UniversitatdeValència o.barbera@uv.es Marco-Lisi- UniversidadeNovade Lisboa mlisi@fcsh.unl.pt Juan-RodríguezETeruel- UniversitatdeValència Jrteruel@uv.es Papertobepresentedtothepanel DemocratisingPartyLeaderSelectionPrimaries:ChallengesandOpportunities BeyondIntraGPartyDemocracy ECPRGeneralConference Bourdeaux,5G7september2013 Draft-Version- - Please-do-not-quote-without-the-autors -permission-
1. Introduction Despite their neighbourhood, Spanish and Portuguese political parties have been following two somewhat different paths in terms of the democratization of their selectionprocedures.inspain,partyprimarieswereintroducedinthelate1990s and its limited use has been mainly restricted to candidate selection. Party leadershipselectionhasremainedapartycongressaffairforthemainstategwide Spanish political parties (Barberà et. alt., forthcoming). However, since the mid 2000ssomeregionalandnewpartieshaveintroduceddirectvotestoselecttheir party leaders (Barberà and Rodríguez, 2012). And, since the early 2010s, StateG wide parties such as the PP or the PSOE have already started to test the use of partyprimariesforleadershipselectionpurposesattheregionallevel.incontrast, candidate selection in Portugal has remained firmly anchored to the decision of partyleadershipandnationalpartybodies,whereastheselectionofpartyleaders has experienced a process of democratization with the introduction of closed primaries. This trend has concerned the main governing parties, while more extreme and antigsystem parties have maintained the traditional mechanisms of political recruitment based on the power of the party congress and national organs. The broad political context of both countries may account for part of these differentpaths,especiallywhenconsideringsomefeaturesofthepoliticalsystem and their party systems. Portugal and Spain share an imperial past originated in thexvthcenturyandalongdeclineduringthexixandxxthcenturies.inthexxth century liberal democratic regimes were introduced in both countries, but were unabletoinstitutionaliseandwerefollowedbydecadesofdictatorship.inthemid 1970s, both Portugal and Spain successfully conducted transitions to democracy. Theyalsohaveincommonseveralfeaturesoftheirpoliticalculture,mainlyrelated their religious denomination and the influence of the Catholic Church in the politicalsystem. Despite this common historical legacy, there are remarkable differences with regardtothepatternofdemocratization.thetwoiberiancountriesareoftentaken as an example of two completely different transition models, namely the consensual type of the Spanish case, on the one hand, and the military coup implemented by the Portuguese army, on the other. The different trajectories of statebuildinganddemocratizationprocessweretoestablishonecrucialdifference between the two countries. While the centregperiphery cleavage plays a fundamentalroleinthespanishpartysystem,thisdimensionismuchlessrelevant in Portugal. To address these territorial demands for selfggovernment the new democratic Spain implemented an ambitious devolution process that led to the presentmultiglayeredstate.inportugal,boththeinstitutionalcharacteristicsand the dimensions of competition contributed to reinforce the national character of party organizations and the centralization of powers in party leaders at the nationallevel. 2
ThemainfeaturesoftheSpanishelectoralsystemweredesignedduringtheearly stagesofthetransitiontowardsdemocracy.theyweremeanttoaddressthreekey issues: to secure stable governments; to guarantee the representation of the nationalminoritiesfromtheperipheryandtostrengthenthenewandfragilemain partyorganizations.asetofinstitutionalarrangementssuchasarestrictedvoteof nongconfidencefortheprimeministerorastategfundingregimeofpoliticalparties was also introduced to protect and enhance both governments and parties. In Portugalthenewelectoralsystemwasdesignedtoguaranteeafairrepresentation of the different social groups and political forces through the adoption of a PR system based on the Hondt method, with relatively low district magnitude (average 10.5) 1. As in Spain, one of the main objectives was to strengthen the consolidation of party organizations. To achieve this goal, the new constitution established that only political parties had the possibility to select candidate for nationalseats,whileindependentscanrunforofficeatthelocallevel. Spain s national party system has evolved through three stages. The first period (1977G1982) was characterized by two big statewide parties, the UCD (Unión% de% Centro% Democrático) and the PSOE (Partido% Socialista% Obrero% Español); two medium statewide parties, the PCE (Partido% Comunista% de% España) and the AP (Alianza%Popular); and several non statewide parties(notably, CiU and the PNV). Thesesixpartiesreachedawideconsensusoninstitutional,economicandwelfare policies.however,thispartysystemcametoanendatthe1982generalelections wherethepsoeemergedasadominantparty.thecurrentpartysystemstartedby the 1993 general elections and comprises two main statewide parties: the PSOE and the PP (Partido% Popular); a medium statewide alliance, IU (Izquierda% Unida) whichincludedthepceandotherleftistpartiesandsocialmovements;andseveral non statewide parties(ciu,convergencia%i%unió; PNB, Partido%Nacionalista%Vasco; ERC,Esquerra%Republicana%de%Catalunya,andothers).- - Although the Spanish party system remains very much the same as in the early 1990s(Oñate,2008),somesignspointoutthatitmaybeevolvingthroughafourth period.thechangesinthepublicmood,speciallysincethebeginningofthedeep economic crisis; the emergence of new statewide parties such as UPyD (Unión% Progreso% y% Democracia); or the steps taken by some of the political parties improving their accountability and responsiveness (i.e. implementing party primaries)mightbeinterpretedasfeaturesofthisnewstageofspain spolitics. InthecaseofPortugal,afterthe CarnationRevolution of25april1974,astable partysystemquicklyemerged,andby1976fourpartiesrepresentedalmost90per centoftheelectorate.apartfromabriefperiodduringthemidg1980swhenthe centregleft Party of Democratic Renewal (PRD) emerged and disappeared, the party system has remained relatively stable. The general tendency (1987G2005) wasforthevotetoconcentrateonthetwocentrist catchgall parties:thecentreg leftsocialistparty(ps, Partido%Socialista), and the centregright Social Democratic Party (PSD, Partido% Social% Democrata). These parties have always controlled the government alone or in coalition with the CDSGPP (Centro% DemocráticoESocialE 1Gallagherindexofdesproportionalityis4.65fortheperiod1976G2007. 3
Partido%Popular),withtheexceptionofashortperiod(1976G1978)ofpresidential governments(freire2005;jalali2007). Sincethe1987 realignmentelections,therehasbeenanincreasingconcentration of the vote in two major parties. The change from consensual to majoritarian democracy has several features: movement from a fragmented to a kind of bipartisanpartysystem;fromcoalition(orminority)andunstablegovernmentsto singlepartymajorities(mostofthetime)andratherstablegovernments;andfrom astrongparliament(andpresident)toastronggovernment(andprimeminister) (Freire2005;Bruneauetal.2001). The economic and financial crisis that has hit Portugal since 2008 put an end to governmentstabilityandsinglegpartymajoritygovernments.itisstilltooearlyto claim that a new phase begun with the 2009 elections and the formation of a socialist minority government, but since then the pattern has been characterized byadecreaseintheconcentrationofvotesinthetwomoderatepartiesandmore unstable governments. It should also be noted that the two radical left parties PCP (Partido% Comunista% Português) and BE (Bloco% de% Esquerda) have always remainedexcludedfromthegovernment,withoutanycoalitionpotentialvisgàgvis theotherparties. 2. Contexts and rationales for adopting primary elections The previous section has highlighted some of the main characteristics and transformations of Spain and Portugal institutional setting. After a synthetic description of the main organizational party features, this section will consider boththestrategicrationalesandthepoliticalcontextinvolvedintheintroduction ofpartyleaderprimaries. Aspreviousstudieshaveshown,thepartyincentralofficeisthemostimportant faceofspanishandportuguesepartyorganizations,controllingtheparliamentary party (Gangas, 1995; Verge, 2007; Oñate, 2008; Lobo 2003; van Biezen 2003; Teixeira2009).Thisismostlytheresultofthehighcentralizationinvestedtothe partyleader(presidentorsecretaryggeneraloftheparty).infact,throughthemain partybodies,thepartyleaderusuallyhasthefinalsayoncandidateselectionand is also directly responsible for choosing the parliamentary leader, who mainly playsthefunction of intermediation between the parliamentary group and the main party bodies. Party discipline is usually very high, strengthening leaders controlovertheirparliamentarygroup.ingeneral,partyleadershaveusuallybeen abletocontrolboththepartyincentralofficeandtheinstitutionalcomponent.the power of party leaders is clear in terms of strategic decisions related, for example, to governmental alliances or ideological changes G or the nomination of ministersandotherpublicofficials.thisisbecausewhenthepartyisinpowerthe party leader usually is the Prime Minister or the most influential partyfigurein coalitiongovernments. Despite these common characteristics, there are differences between Portuguese andspanishpartyorganizationsworthyofexamination.themainorganizational 4
variances between Spanish parties are the divides between both statewide and nongstatewide parties and new and established ones. While the former mainly referstosizeandstrategy,thelaterhassofarmoreorganizationalconsequences. In organizational terms, despite slightly different original models both PP and PSOE have converged towards highly centralized organizations (Méndez, 1996; Garcia Guereta, 2001). In both statewide parties the regional party leaders have growninternalinfluencesincethe1990s,speciallywhentheyareinofficeandthe nationalpartyisintheopposition. IU still remains less centralized due to its originalmodelasanelectoralallianceandthemorelimitedpowerofitsregional leaders (Ramiro, 2002). The main differences for the Portugueses parties are mainly related to their original model. Both the PS and CDSGPP followed a topg downprocessofpenetration,leadingtoaverycentralizedstructure.ontheother hand,localpartystructureswereextremelyimportantintheformationofthepsd, withnotablesalsoplayingasignificantroleatthenationallevel(jalali2007:146g 147). This foundational phenomenon has important implications for the party leaderpositioninternally.whiletheleadingroleofthepartychairwithintheps and CDSGPP is evident, within the PSD party leaders have essentially been more managers rather than uncontested decisiongmakers. This means that when in opposition, the party assumes a more stratarchical structure and party leaders have more difficulty in asserting their power over local bosses(jalali 2006: 368G 370). Consequently, the effective role of the party leader can vary according to institutionalpositionandelectoralperformance. Asithasbeenstatedintheintroduction,sofarthemainSpanishstatewideparties have been very reluctant to open their leadership selection methods. Up to now, only some neworregionalpartiesand party branches have implemented party primaries. The Catalan ICV was the first Spanish party to ever introduce party primaries to select its party leader. They were adopted in hard times for ICV becauseofitsbreakingoffwithiu,itssistergpartyinspain.thatledtoaninternal split and the formation of a new party that competed against ICV in the 1999 (regional) and 2000 elections. Although ICV managed to maintain its representation at both levels the cost was a considerable electoral setback.that fostered criticism to ICV s party leader and the emergence of a challenger. The challenger asked for primary elections to select the 2000 general elections top candidateandshortlyafterwardsfortheselectionofthepartyleader.thatledto the withdrawal of the incumbent party leader. The main rationale for the implementation of ICV s primary elections was the move to persuade the incumbentpartyleadertoretire.butalsotoinvestthechallengerwithanewand muchneededlegitimacyandauthorityagainstthepotentialrebellionofrestofthe partyboard. Inthe2003regionalelections,ERCdoubleditsshareofvotes(from8%to16%) andjoinedthenewleftcoalitiongovernmentincatalonia.however,shortlyaftera scandal led to the resignation of its party leader from government. In a sound tactic,erc sleaderdecidedtorunforthe2004generalelectionsandachievedthe bestresultsforthepartysincethe1930 s.thepartyleaderthenproposedsome organizational reforms to restrict the way the party Congresses were held. The introduction of party primaries was meant to be a sort of exchange with dissatisfiedpartymembers.buttheywerealsodesignedtoreinforcethepowerof 5
the party leader and his deputy against the party elites. The first primaries were held at the 2008 congress in a context of bad electoral results and deep factionalism: On one side there was a fight for power between the discredited partyleaderanditsdeputy.ontheotheronetherewereminorityfactionswanting to withdraw the regional coalition. Four teams of candidates run for the post of party leader and its deputy. The deputy leader won the challenge, ERC stayed in government and the minority factions decided to split shortly afterwards (Argelaguet,2009). UPyD was created shortly before the 2008 general elections. BornasanantiG stablishment party, UPyD relied heavily on its party leader(rosa Díez) that was theonlymembertoachievedrepresentationinthe2008elections.shepromoted primaryelectionstoalllevelsasawaytodifferentiateupydfromthe established parties and, also, to assert its influence over the rest of the party board and the partyregionalelites.however,itsearlysuccessdidn tstopinternaldissent.atthe 2009 congress the party leader was elected with closed primaries. Although anothercandidatechallengedher,shewonbyalandslidevictory.asstatedbythe partystatutes,upydhasimplementedpartyprimarieseversince. ThePPintheBalearicIslandslostthegovernmentinthe2007regionalelections. The party leader was then replaced in the 2008 regional congress. However, the new party leader resigned just one year after being elected. That led to the appointmentofathirdpartyleader(josér.bauzá)whowasquicklychallengedby other party elites. To overcome its democratic deficit the appointed party leader calledforaclosedprimaryelectionin2010,thefirstoneeverheldbytheppatthe regional level. Implementing party primaries the leader rightly expected, as well, toovercomethepowerofthedividedlocalpartyelites.thepartyleaderwonthe primary,successfullyregunitedthepartyandshortlyafterwonthe2011elections andregainedtheregionalgovernment. Afterlosinggovernmentinthe2009regionalelections,thePSOEinGaliciaelected anewpartyleader.althoughthepsoesufferedanationwideelectoralsetbackat the 2011 local elections, that specially undermined the authority of the Galician party leader. Few months later, at the 2011 general elections serious internal conflictsemergedthatworsenedbythe2012regionalcongresswheretheparty leader won by a very narrow victory against a former national minister. As expected, the PSOE suffered a serious electoral setback in the early called 2012 Galician elections. That impelled the emergence of new challengers to the party leader. He then called for a party congress with closed primaries as the party leaderselectionmethodandrevealedhisintentiontoleavethepartyleadershipby 2013 2.Interestingly,callingforprimarieswasnotonlymeanttobeawaytosolve aregionalproblembuttoincreasethepressureforreformontheselectionofthe nationalpartyleader. InthePortuguesecase,therecentchangesinthemethod of leadership selection represent a remarkable phenomenon, especially if we consider that party 2ThePSOEstatutesdonotallowprimariestoselecttheregionalpartyleaders.Alegalformula(a multigstagemethod)wasfoundtocircumventtheproblem.thecallforprimarieswasmadeinthe summerof2013. 6
organizational changes are very unusual and that party elites have been mainly concerned with the institutional arena, rather than with the improvement of members participation and inclusion in decisiongmaking processes (Jalali 2007; Lisi2011). Theargumentinfavoroftheadoptionofclosedleadershipprimariesemergedfirst inthepsin1992inthecontextofintragpartyconflictsbetweenwouldgbeleaders (Lisi2011:219G221).Inthiscase,democratizationreformsweretoalargeextenta Trojan horse in the attempt to reinforce internal support and to gain party leadership (see also Panebianco 1988). Yet this attempt failed and this change remainedonpaperuntilthemajororganizationalreformthattookplacein1998 (vanbiezen2003;lisi2009).thistimethesocialistleader,antónioguterres,was forcedtoreachacompromisewiththeoppositionontherulesfortheelectionof the main party bodies and their powers. The final arrangement established the directelectionofthepartyleaderbymembers,whilereturningtothetraditional hierarchicalstructurebasedontherepresentativeprinciple. Ontheotherhand,bothrightGwingpartiesintroduceddirectleadershipelections following the 2005 electoral defeat and their return to opposition. In the case of the PSD, the new leader (Marques Mendes) decided to force the introduction of closedprimaries despitestronginternaldivergences inordertoenhanceparty legitimacy, the accountability of party leaders and to boost participation and mobilization (Lisi 2010: 137G138), As for the CDSGPP, the adoption of OMOV for leadership selection was due mainly to the weakness of the incumbent leader (Ribeiro e Castro), who was elected for a very tight margin during the previous partycongress.bothpartiesdecidedtoadoptapluralitysystem,butin2012the PSDdecidedtochangetoamajorityrunoffsystem. With the partial exception of the PS, which adopted OMOV following a major organizational restructuring, the move towards more democratic methods of leadershipselectionwasmainlybasedonstrategicconsiderations.therefore,the raison%d être for the introduction of direct leadership selection hasbeentwofold. On the one hand, this change followed electoral defeats, with the aim of demonstrating party renewal in order to appeal to voters and recruit new party members. On the other hand, democratizing reforms have been used to contest internal power.oppositionleadershaveoftendefendedthedeepeningofintrag party democracy to challenge the incumbents, whereas weak leaders (such as Marques Mendes or Ribeiro e Castro) have tried to strengthen their position throughtheuseofomov.thefactthatthesameleadershaveoften(andsuddenly) changed their opinion on leadership selection procedures depending on their circumstancesconfirmsthestrategicuseofdemocratizationreforms. As mentioned above, in Portugal all the main governing parties have recently adopted closed primary for the selection of leaders. This means that only party membersareentitledtoparticipateforthechoiceofthepartyleader.twofactors are responsible for this decision. On the one hand, there are institutional constraints,giventhatthenewpartylaw(2003)establishesthattheselectorateof officialpartybodiesisformallylimitedtopartymembers.ontheother,itshould benotedthatopenprimariesdonotfitthepoliticalcultureofportugueseparties. 7
Thefactthatlocalnotableshavecontrolledtheaccessofpartymembersasaway toinfluencetheelectionofpartycadresandcandidatestolocalandnationaloffices suggests that there are strong antibodies against the openness of party organizations to party sympathizers or independents. Moreover, open primaries are considered to be detrimental to the image of parties in public opinion and unnecessarygiventhelackofdeepculturalandsocialcleavages. 3. The primary election process Thissectionwillfocusonthemainfeaturesofthepartyleaderprimaries.Itwill analyse both the formal procedures and their real implementation over time in bothcountries. It is worth noticing that most Spanish parties have hold primaries without introducingtheseproceduresintheirpartystatutes.thisisespeciallythecaseof theppandpsoe sregionalbranches,butalsoappliesforthefirstprimariesheld byicv andupyd. On the other hand,the regulation of party primaries is quite similar for the five Spanish parties selected in this chapter(table 1). All of them link the selection of the party leader to the celebration of a party congress. ICV, ERC,PSOE(Galicia)holdtheirprimariessomedaysbeforewhileUPyDandthePP dothematthesametimeofthecongress.inportugaltheelectionofpartyleaders takes place some time (approximately one month) before the congress. Some partieslinktheelectionofthepartyleaderwithpartoftheboard:thepartyleader andtheboardareelectedonthesameballotinupyd,whilethepartyleaderand his/herdeputyareelectedatthesametime(butwithdifferentballots)inerc.in all the other cases, the newly elected party leader submits the list of the party boardtotheapprovalofthepartycongress. Almost all Spanish parties are using their members as selectorates (closed primaries). ICV is the only party using a combination of both members and registeredsympathisers 3.Thissingleselectorateranksquitehigh(18outof24)in HazanandRahat(2010:48)inclusivitycontinuum.ItisworthnoticingthatinICV and PSOE (Galicia) first primary elections a multigstaged method was used: the party leader was elected first by all members and then ratified by the party congress.portuguesepartieshavealsoadoptedclosedprimaries.however,during the campaign to elect the new socialist leader after Sócrates resignation(2011), oneofthetwocandidates(franciscoassis)defendedtheneedtointroduceopen primaries for selecting the primegminister and mayoral candidates. This was mainly a strategic move aiming to increase supporters by deepening party democratization 4. Yet this issue was a divisive one as emerged during the last congress (April 2013). In the attempt to neutralize internal opposition and to enhancepartydemocratization,thenewsocialistleader(antóniojoséseguro)was forcedtoaccepttheadoptionofopenprimariesforlocaloffices.moreover,thenew statuteapprovedinmarch2012establishedthepossibilitytoadoptprimariesalso 3Thesympathisers (nonpayingfeemembers)votesarecastedinadifferentboxbut,apparently, allofthemarecountedaltogether. 4The same happened when Ribeiro e Castro, former CDSGPP leader, proposed in 2007 to adopt partyprimariestoselectmps,aswellaswhenthepsddecidedtoadoptclosedprimariesin2005. 8
fortheselectionofmps.yetthedirectelectiononlycanbeusedonlywhenthereis more than one list of MPs at the district level and both the federation (district level)andthenationalpoliticalcommissionhaveavetopower. Keydifferences can be found in Spanish parties on the candidacy requirements. ThescalerangesformUPyDwherevirtuallyallpartymemberscancontesttothe PSOEorERCwhereasubstantialamountofsupportisneed.Thisfeatureisvery importantbecausemostparties(excepticvandupyd 5 )statethatifthereisjust onecontestantnoelectionisneeded.thatis,incidentally,oneofthereasonswhy in Spain many (candidate and leader) primaries are called but not held. Sofar therehasnotbeenanyscandalrelatedtomassenrolmentduringthecampaigns.in Portugal,partystatutesestablishthatacandidacymustbesupportedbyacertain numberofpartymembers.thisrequisitehasvariedacrosspartiesandovertime. ThemostimportantchangetookplaceinthePSD,whichrequiresthesubmission of1500signaturesofpartymembers,whilstthisbarrierismuchlowerfortheps (100signatures) 6.Thishasbeenaproblematicissueinthetwomainparties(PS andpsd)duetothemanipulationofenrolmentfilesbylocalnotables.therehave beenseveralcasesofmassenrolmentintheperiodthatimmediatelyprecededthe electionofpartyleaders.usuallyincumbentshaveusedtherulestomakeitmore difficult for nongmembers to join and participate in the leadership contest, while challengershavefavouredthepossibilityofregisteringnewmembersuntilafew days before the election. Another requirement for the participation in leadership primaries is a membership period of several months (between one year and18 months) 7. AllSpanishpartieshaveoptedforapluralityformula.Themaindifferenceinterms of the electoral system can be found in how many ballot boxes are available on ElectionDay.Mostpartiesplaceoneballotboxperlocalbranch.However,others such as ICV and UPyD have occasionally allowed the use of electronic voting. PluralityformulaisalsothemostcommonmethodadoptedbyPortugueseparties. However, there are strong internal divergences regarding the most appropriate method to select party leadership. In particular, strong criticism has emerged withinthecdsgpp,leadingpartydelegatestoreintroducein2011theelectionof the party leader by the congress through plurality vote. The OMOV method for leadershipselectionismoreconsolidatedwithintheps,whileinthepsdthereare stillstrongdivergencesofopinionandin2012partyleadershipdecidedtochange fromapluralityformulatoarungoffmajoritysystem(table2) 8. 5UPyDstatesthatwhenonlyonecandidateconteststhevotescastedforhis/herhavetobehigher thantheblankvotes,otherwisethecandidateisrejected.sofarthishasneverhappened. 6Unlike what happens in parties insome other countries, in none of the Portuguese parties is a votingfeerequiredinordertoparticipate. 7Itisworthnotingthat,ingeneral,therehasbeenanincreaseinthemembershipperiodrequired torunforpartypositionsatthenationallevel. 8According to a recent survey of socialist party delegates (2011), 68 per cent of PS activists support the selection of the leader by party members, while 20 per cent would prefer an open ballotforsympathizersandvoters(lessthan10percentsupportthepreviousprocedurebasedon partydelegates).therearenosimilardataforthepsd,butinboththe2010and2012congresses an intense debate took place about the opportunity to return to the election of leaders by party delegates. 9
- Intermsofdeselectionprocedures,theSpanishandPortuguesecasesaresimilar in that party statutes are completely omitted. As a consequence, the only way to punish incumbent leaders is to start a new electoral process by the call of a new congressbypartydelegatesormembers.fundingtheelectionofpartyleadersis alsoomittedfromspanishandportuguesepartyregulation.theonlymentionsto thistopicestablishthatallcandidatesmusthaveattheirdisposalpartystructures on equality bases, as well as equal access to membership files. In Portuguese partiesthisfacthasledtointragpartyconflictsduringthecampaign. [TABLE-1-AROUND-HERE]- [TABLE-2-AROUND-HERE]- - So far, the main unitsofanalysis have been the political parties. The rest of this chapter will focus on processes, not on parties. This will give a more detailed accountofwhathashappenedineachpartybeyonditisstatedintheirregulation. However,itmustbenotedthatinSpainnotallpartieshaveheldthesamenumber ofprocesses.infact,thereisadisproportionbetweenthenumberofpartyleader primariesheldbyupyd(75%oftheprocesses)andthatoftherestoftheparties (ICV,PP,ERC) 9.Thereisalsoadisparityregardingthelevelatwhichtheprimaries areheld.almostallofthemareorganizedbyregionalpartiesorbytheregional branches of statewide parties. UPyD is the only party that has organized leader primariesatthenationalleveluntilnow.inportugal,closedprimarieshavetaken place exclusively to select party leaders at the national level. Besides this, since 2003 theps has adopted closed primaries to select party leadersatthedistrict level 10. However, these elections have only an internal character without any visibilityoutsidetheparty.moreover,these middleglevel leadershaveamarginal roleinthepartyfunctioningandtheirinfluenceonthedecisiongmakingprocessis secondary.asmentionedearlier,thepsisexperiencingforthefirsttimeaprocess of- primariesfortheselectionofcandidatesatthelocallevel.thisexperiencehas shown two main features. On the one hand, only a very small proportion of contestshavetakenplace(tenoutof308municipalities).thismeansthatinmost cases there is no real competition, as local leaders are selected by middleglevel elites without any challengers. On the other, these contests adopted OMOV especially when there were deep intragparty divisions.from this viewpoint, the adoptionofprimarieshasnotonlyallowedtoincreasingthedegreeoflegitimacy, aswellastosolvinginternaldivergencesandtokeeppartycohesion. Democratization of the party leadership selection has indeed increased the numberofpartymembersinvolvedinthesedecisionsinbothcountries.andyet, dataonparticipationandmobilizationmustbeinterpretedwithcautionduetothe factthatfiguresonpartymembershiparehighlyunreliableinbothcountries.that 9ThePSOE(Galicia)willhelditsfirstpartyleaderprimaryinSeptember2013. 10The requirements to present a candidacy are 2% of signatures of the party members at the districtlevelor80members. 10
said, the participation rates have fluctuated substantially between parties and electionsbothinspainandportugal(tables3and4).inspain,upydandicvhave the lowest turnout rates (29%), while ERC and UPyD have the highest ones (slightly above 70%). The mean score for the overall processes is close to 50% (48,6%) of the members involved. In Portugal, the level of mobilisation in closed leadershipprimariesvariesgreatly,rangingfrom14to66%.onaverage,thepsd displays the highest rate of participation(52.6%), while members of the CDSGPP arelesslikelytoparticipate (23.8%).The PS falls between these two (40.4%). These figures are not significantly different from data on internal participation before the reform of leadership selection methods. 11 Overall, the Portuguese experience shows that only a small portion of the parties supporters actually participatesinthesecontests. On the other hand, party primaries have not always meant competition between twoormorecontestants.inspain,onlyinlessthanhalf(47%)oftheprimarieshas been more than one candidate,theresthave been coronations. ICVistheonly partythathasneverhadacontestedpartyleaderprimary.nevertheless,contested primarieshavehigher(18%)competitionratesthanpartycongresses(barberàet. alt, forthcoming). In this respect, party primaries do have represented a change. The overall picture of Portuguese parties is the low number of candidates and a relatively high proportion of coronation. The evidence suggests that in the PS therehasbeenadecreaseinthenumberofcandidatesaftertheadoptionofomov procedures,whileinthepsdthischangehasbeenassociatedtomorecentrifugal dynamicsandlowerlevelsofpartycohesion.theshortexperienceofthecdsgpp displays a mixed trend, with the alternation of both highly competitive elections and party leaders coronations. If the institutional position is a key factor for explaining this different pattern, the degree of competitiveness is also associated to leaders popularity or electoral challenges (Lisi and Freire 2013). This is true especiallyforthepsd,whilecdsgpppartyleadersseemlessvulnerable,showing greatercapacitytoinsulatethemselvesfromelectoralpressures. Regardless of the party, when competition has taken place the number of competitorshasremainedlow.inspainonly3outof33processeshavehadmore than3contestants,whileinportugalthisproportionisslightlyhigher(4outof19 contests).despitethenumberofnongcontestedprimariesinthespanishcase,the meanshareofvotesofthewinnerishigherinportugalthaninspain(86.5%and 83.7%, respectively). This phenomenon can be explained by the plebiscitary nature of leadership contests when the party is in office. This means that incumbencyisalwaysassociatedwithverylowlevelsofcompetitiveness,whilethe shifttooppositionfostersmorecompetitivecontests.nevertheless,ifweonlytake intoaccounttheshareofvotesofthecontestedelectionsthemeandropsto70%in Portugal and 60% in Spain. The mean results of the runner up are 33.6% for Portuguesepartiesand27.6%forSpanishones.Thisindicatesthatevenwhenthe competitiondoestakeplacetherunneruphardlyrepresentsaseriousthreatfor thewinners. 11Data from internal referenda show the stability of turnout rates. The CDSGPP called for a referendumin1992,inwhich26percentofpartymembersparticipated.turnoutinthesocialist referendumin1982was30percent. 11
Asaconclusionshouldbepointedoutthattheintroductionofclosedprimariesto selectpartyleadershascertainlystrengthenedthepersonalizationofpoliticsand enhanced candidategcentred campaigns. Although party leaders have always playedacentralroleintheelectoralarenaandhavebeenanessentiallinkbetween party images and voters (Oñate, 2008, Lobo 2006), the democratization of leadership selection has led the media to increasingly focus on party candidates and to relegate to a marginal role intragparty debates on policies and strategic orientations that usually take place during the party congress. This is especially trueinportugal where all mayor parties have adopted closed primaries as their leadershipselectionmethods,butcanalsobeappliedforthespanishparties.one important indicator of this increasing personalization is the higher level of professionalization of campaigns for party leadership. Since the 2004 election of the PS leader the first competitive contest G, internal party campaigns have attractedwidespreadattentionfromthemassmediaandhavebeenorganizedas electoral campaigns with a campaign staff and external consultants, raising the costsnormallyassociatedwithinternalelections 12. And yet, Portuguese leadership contests have not displayed high levels of negativity. Generally speaking, candidates prefer to highlight the divergences between the different programmatic and strategic options rather than stressing the negative aspects of ones opponent. Two main reasons can explain this phenomenon. One is related to the political culture, which has always rejected personal attacks as an instrument to increase leaders popularity. It is worth notingthatalsomassmediadonotusuallyrelyonthistypeofcommunicationin thebeliefthatitfostersdemobilizationandmistrusttowardspolitics.ontheother hand,negativecampaignsareconsidereddetrimentaltopartyimageandleaders tend to downplay intragparty conflicts. There is very little evidence on the negativityofthecampaignsinvolvedinpartyleaderprimariesinspain.mostofthe media coverage has been until now focused on the results, not on the internal campaigns.thisisduetothefactthatthishasmainlyconcernedsmallstatewide andregionalpartiesandsomeregionalbranchesoftheppandpsoe. [TABLE-3-AROUND-HERE]- [TABLE-4-AROUND-HERE]- 4. Political consequences on parties Theliteraturehasidentifiedseveraldimensionsthatcanbeinfluencedbytheuse ofpartyleaderprimaries:thecontinuityoftheincumbentpartyleader;theprofiles 12AllmainPortuguesenewspapershavestartedtoconductopinionpollssincethefirstcompetitive contest in 2004. In general, the higher the competitiveness, the higher is the visibility and the coverageofleadershipelectionsinthemedia.thecandidatesofthetwomainpartiesusuallyform ad%hoc teamsfortheorganisationofinternalcampaigns.unfortunately,nodataorestimatesare availableforthefundingofleadershipcontestsandinternalcampaigns. 12
of the party leaders(women, age, seniority); the number of party members; and, eventually, the electoral results of the party. This section discusses the main consequences and outcomes of this kind of primaries on the Portuguese and Spanishparties. Sofar,partyleaderprimariesseemtohaveledtoalowrateofincumbentstability in Spain (Table 2). Less than 40% of the incumbents do not contest on the next primary 13.However,ithastobetakenintoaccountthatagreatdealofprocesses came from UPyD, a statewide party formed by 2007. This means that instability hasprobablymoretodowiththeformationofanewpartythanwiththeprimaries themselves.portuguesepartiesdisplayahigherlevelofstabilitywithanaverageof 62.5% of contests where incumbents compete for regelection (52.6% if we consider also the first primaries). Usually incumbents do not contest party leadership when they suffer electoral defeats or are not able to increase party popularitywhentheyareinopposition.regardlessoftheparty,thechancesofan incumbentwinningthecompetitionareveryhigh.uptonow,onlyoneincumbent outof8(12.1%)haslostaprimaryinspain,whileinportugalthepercentageis evenlower(8.3%). InthePortugueseandSpanishcasepartyleadershiphasbeendominated ifnot monopolized by men. In Spain, only seven women (20.6%) have been encouragedtocontestthepartyleaderprimaries,allofthemfromupyd(table3). ThisprobablyhastodowiththefactthatUPyDnationalpartyleaderisawoman (RosaDíez).ButalsoseemstopointoutanotherdividebetweenUPyDandtherest ofthepartiesaswomenrepresenttherealmostonethird(28%)ofthecontestants. Interestinglyenough,in UPyD only two women (25%) have lost their races, a record almost as low as that of the incumbents. In Portugal, the only woman electedaspartyleader(psd)wasselectedthroughclosedprimaries,butthiscase can be considered as an exception and the choice obtained a wide support also frompartycadres. Party leaders in Spain are generally younger than their European counterparts. The average age of the winners is about 48 years old in Spain.There are no big differencesbetweenpartiesinthisregard,notevenwiththeageofpartyleaders elected through party congresses (Barberà et al., forthcoming). In Portugal the average age of party leaders elected under closed primaries is 55.9 years old, which is slightly above the average age of party leaders under the democratic regime(lisiandfreire2013).despitethisapparentcontinuity,itisworthnoting that the introduction of direct elections shifted the emphasis from intragparty conflicts to less ideological candidates with more attention to external dynamics. After the first and second generations of leaders more directly associated to the institutionalizationofthedemocraticregime,anewyounggeneration especially withinthepsandpsd withnoconnectionstothetransitionperiodhasassumed power.therefore,wecanarguethatthischangefacilitatedtheadoptionofomov andvicegversa becauseitbrokethetraditionalmechanismsbasedontheelection ofpartydelegatestothecongress. 13Firsttimeprimariesareexcluded,otherwisetheratewillbeevenlower(30%). 13
This generational turnover has also relevant consequences in terms of political background. From this viewpoint, two considerations are crucial for understandingtheportuguesecase.thefirstisthatpartyleadershiphasfollowed a process of professionalization based on the occupation of important positions withintheparty.thismeansthatagoodpartisancurriculumisaprerequisitefor contesting the leadership at the national level. All the leaders elected through closedprimariesareexperiencedpoliticianswithastrongbackgroundinsidethe party for example, as previous leaders of the youth organization or at the institutional level (as MPs). In other words, there is little room for outsiders or leaders with a more technocratic profile. As shown in table 4, all leaders elected throughclosedprimaries withnoexceptions haveheldanationalposition.in contrast with Portugal figures, the rates of senior party leaders elected via primariesareverylow(15.2%)inspain.thishastodowiththefactthatmostof them are leaders of the regional branches of state widegparties and, more specifically, the newly created UPyD. Nevertheless, the same happens with the party leaders of the PP and PSOE regional branches. Although it may seem paradoxical,theregionalicvisthepartywithmoreseniorleaderselectedthrough primaries.thisismainlybecauseofthecareerpathsofthelasttwopartyleaders (servingfirstasmpandthenaspartyleaders). As closed primaries formally empower party members they are usually expected to boost party membership. However, the literature has warned of some pathologiessuchas instantmembership (membersenlistedjustfortherace)and theexistenceof votecontractors (Kenig,2009;CrossandRahat,2012). Inthe caseofspainandportugalitisalsoworthnotingthatpartiesdonotkeepaccurate enrolment files, and membership figures are likely to be inflated. Therefore, it is extremely hard to assess the real impact of direct leadership selection. The still very limited evidence from Spanish parties shows that party membership is usuallybiggerinelectiondaythanitwasoneyearbefore(table3).thenumberof missing cases does not allow commenting much further. This seems consistent, nonetheless,withtwootherevidences:thegrowingnumberofmembersrecorded in comparative studies of party membership (Van BiezenandMair,2001;Van Biezen et. alt. 2012); and the steady growth of UPyD since its origins. One year after the primaries the (limited) data shows mixed results. In some cases the membership figures grow far beyond the 10%, and yet in others there is a small decline. Whether this has to do with the existence of instant membership pathologies seems still unclear. In Portugal there is almost no relationship betweenpartymembershipfiguresandleadershipcontests:datagoupanddown regardless of the election of new leaders. Usually, party membership increases whenthepartyisinpower,whileinoppositionthesedataareupdatedleadingto exclusionofasignificantproportionofmembers(especiallythosewhodonotpay regularly party fees). Rather than the effect of different methods of leadership selection, it seems that the evolution of party membership is strictly linked to electoralandinstitutionaldevelopments. Regardlessoftheparty,therelationbetweenmembersandvotersisveryweakin bothportugaland Spain (Tables 5 and 6). Taking the parties into consideration, partymembersrepresentaround2%g3%oficv s,upydanderc svotersand9%g 10%ofthePP(BalearicIslands)voters.UPyDratiosrangefrom0.5%atCastillala 14
Manchato5.6%attheCanaryIslands.Duetothelimitedtimeframethereisstill little evidence of changes over time linked to the use of party primaries. In Portugal the evolution of party membership suggests that members are likely to increasewhentheyareinpower.overall,inthecaseofthepsdandcdsgppwecan see a decline of M/V ratio due to the growth of their electoral support, but in severalcaseswecannotachieverelevantconclusionsdue Thissectionconcludeswithananalysisofchangesintheelectoralperformanceof partiesandapreliminaryassessmentofwhethertheycanbeconnectedtotheuse ofprimaries.atfirstsightitisworthnoticingthatthespanishpartiesdoimprove their electoral performance in the next national election after holding a primary (Table 5). ERC is the only exception to that pattern.however, there are other factorsthathavetobetakenintoconsideration:themostimportantpointisthe long time (almost two years on average) amid the primary election and the subsequentnationalelection.thattimespanmakesitdifficulttoascertainadirect relationbetweenthetwofigures.theotheroneisthatmostoftheselectedparty leaders in this chapter didn't actually take part in the national elections. This is becausetheyaremainlyregionalpartyleaders(theexceptionbeingupyd srosa Díez)andtheycompeteastopcandidatesintheregionalelections 14.Thelastone hastodowiththeverylowmediaprofileofthepartyleaderprimariesheldbythe selected parties. That said there might be indirect effects derived from the legitimacy of the process and, of course, from the performance of those party leaders. Moving to the Portuguese case, we can see that there is almost no relationship between the electoral performance and leadership contests. As noted above, strategicconsiderationswerecrucialfortheintroductionofomovproceduresfor right parties. This followed the successful example of the 2004 election of the socialist leader, which contributed to increase its electoral performance in the 2005elections.However,closedprimariesdonotdisplayanysystematiceffectin termsofelectoralresults.althoughtheremightbeshortgtermbenefitswithregard to party popularity and leader visibility, the performance of the main governing partieshasoscillatedupanddownregardlessofthetimingofleadershipcontests. Due to the relatively limited experience of closed primaries in Portugal, it is difficulttoanalyzetheeffectontherelationshipbetweenleadersandmembers.in thecaseoftheps,theevidencesuggeststhatthisprocedurehasstrengthenedthe degree of centralization, with a marginalizationofbothlocalpartybranchesand the congress (Lisi 2006). It has also reduced the programmatic divergences, leadingtomorepragmaticandflexibleintragpartyalignments.ontheotherhand, thepoweroflocalnotablesisgreaterinthecaseofthepsd,whichhasdisplayeda significantterritorialcleavageintermsofpartyorganizations.thismeansthatthe conflict between centre and periphery may achieve higher levels of salience, especially if closed primaries tend to enhance the centralization and the concentrationofpowerinthenationalpartyelite.eveniftherearesignsthatthis is what is happening, it is still too early to make a definitive assessment on the impact of OMOV procedures on intragparty dynamics. As in Portugal, party 14FewUPyD sregionalpartyleaderscontestedastopcandidatesinboththe2011regional(may) andnational(november)elections. 15
primaries do seem to have enhanced party centralization. But again the Spanish experience is not really conclusive on the impact of primaries on the leadersg members relationships. To some extent they have ousted the power of party congress delegates (still relevant for programmatic matters). Yet, few media attentionmakeunclearifthathasdownplayedorenhancedtheroleoflocalparty elites.moreresearchisneededtoassesswhethertheyarestillneededtomobilise supportatthelocallevelortheyhavebeenbypassed. [TABLE-5-AROUND-HERE]- - [TABLE-6-AROUND-HERE]- - 5. CONCLUSIONS Thispaperhadtwomainobjectives.Ontheonehand,weexaminedtherulesand processes of leadership primaries in Spain and Portugal. On the other, we investigated the impact of leadership contests and compared the two countries, highlightingthemainsimilaritiesanddifferences.thefirstpointworthmentioning is that in both Southern European democracies the adoption of primaries has shownalowlevelofinstitutionalization.thismeansthatthismethodofleadership selectionisstillacontroversialissueinseveralparties,withpartyelitesdividedon the pros and cons of OMOV procedures. Moreover, party statutes lack clear and wellgdefinedrulesforprimarycontests.theabsenceofanaccurateregulationand thefactthatinspainfewcontestshavetakenplacearetworelevantindicatorsof thelowlevelofinstitutionalization. Overall, in both Portugal and Spain only members have been eligible to vote for leadershipselection.althoughtherehavebeensomeexceptions,partyelitesseem toconsideropenprimariespotentiallydangerousforthecohesionandthestability of party leadership. There has also been an overwhelmingly preference for the adoptionofapluralityformula,withnosignificantexperiencesofexperimentation. Yet primaries in Spain have been adopted for the selection of leaders at the regionallevel,whileinportugalthisprocesshasconcernedmainlynationalparty elites. What are the consequences of leadership primaries in Portugal and Spain? First, the evidence suggests that the introduction of party primaries did not have any impactontherepresentationandresponsivenessofpoliticalparties.ifwelookat gender,ageorpoliticalbackground,wefoundsmallornodifferencesbeforeand aftertheadoption of OMOV procedures. In Spain the only exception seems to be UDyP, a new party that has been able to attract disaffected voters from the mainstreamparties.second,therearemixedtrendswithregardtotheevolutionof party membership and electoral performance. It is hard to draw a definitive 16
conclusionduetothereliabilityofpartymembershipfigures,ontheonehand,and thetimespanfromleadershipcontestsandnationalelections,ontheother.finally, therearesomeexamplesofincreasingcentralization,thusstrengtheningtherole of party leaders visgàgvis members. However, these conclusions must be taken withanoteofcautionduetothelimitedexperienceofprimariesinbothcountries. 17
LIST OF ACRONYMS PP:People sparty(partidopopular),spain CDS-PP: Democratic-Social Centre People s Party (Centro Democrático-Social Partido Popular), Portugal ERC:CatalanRepublicanLeft(EsquerraRepublicanadeCatalunya),Spain ICV: Iniciative for Catalonia-Greens (Iniciativa per Catalunya-Verds), Spain PS: Socialist Party (Partido Socialista), Portugal PSD: Social-Democratic Party (Partido Social-Democrata), Spain PSOE:SocialistParty(PartidoSocialistaObreroEspañol),Spain UPyD:Union,ProgressandDemocracy(UniónProgresoyDemocracia),Spain REFERENCES Argelaguet, J.(2009) The power of grassgroots members in a political party: the caseofesquerrarepublicanadecatalunya Paperpreparedfordeliveryat the workshop on Promoting internal party democracy: a selling point, a seriousdanger,oraredundantexercise?,lisbon,2009. BarneaS.,andRahatG.(2007)ReformingCandidateSelectionMethods:AThreeG LevelApproach,in%«PartyPolitics»,XIII,pp.375G394. Barberà, O., RodríguezGTeruel, J., Barrio, A. and Baras, M (forthcoming) Spain, Cross, B. and Pilet, J.B. (eds) The% Selection% of% Political% Party% Leaders% in% Contemporary%Parliamentary%Democracies.%London:Routledge. Barberà,O.andRodríguezGTeruel,J.(2012) Theintroductionofpartyprimaries inspain,paperpreparedfordeliveryattheworkshopon PartyPrimaries in Europe: Consequences and Challenges, European Consortium for Political Research s Joint Sessions of Workshops, 11G15 April 2012, UniversityofAntwerp,Belgium Bruneau, T. C., et al (2001) Democracy, Southern European style, in P. N. DiamandourosandR.Gunther(eds.),Parties,%Politics,%and%Democracy%in%New% Southern%Europe,Baltimore:TheJohnHopkinUniversityPress. Cross, B. and Rahat, G. (2012) The pathologies of party primaries and their possiblesolutions,paperpreparedfordeliveryattheworkshopon Party PrimariesinEurope:ConsequencesandChallenges, EuropeanConsortium for Political Research s Joint Sessions of Workshops, 11G15 April 2012, UniversityofAntwerp,Belgium Freire,A.(2005) PartysystemchangeinPortugal,1974G2005:Theroleofsocial, politicalandideologicalfactors,portuguese%journal%of%social%science,4(2): 21G40. Gangas, Pilar (1995). El% desarrollo% organizativo% de% los% partidos% políticos% españoles% de% implantación% estatal.% Madrid: Instituto Juan March de Estudios e Investigaciones.Ph.Ddissertation. GarcíaGGuereta,ElenaM.(2001).Factores%externos%e%internos%en%la%transformación% de% los% partidos% políticos:% el% caso% de% APEPP. Madrid: Instituto Juan March de EstudioseInvestigaciones.Ph.Ddissertation. Jalali, C. (2006) The woes of being in opposition: The PSD since 1995, South% European%Society%&%Politics,11(3 4):359 79. 18