[PDF] Explaining the Onset of Cohabitation under Semi

View - Download Explaining the Onset of Cohabitation under Semi

Previous PDF

Next PDF

ExplainingtheOnset ofCohabitation under

Semi-presidentialismpost_8701..20

RobertElgieandIain McMenamin

DublinCityUniversity

Semi-presidentialism-where theconstitutionpr ovidesforbothadir ectlyelectedfix ed-term presidentandaprime

ministerandcabinet collectiv elyresponsib letothelegislature-isan increasinglycommonformofgov ernment.F or

manyobservers cohabitationistheAchillesheelofsemi-presidentialism.Thisarticleaims toidentify theconditions

thatare associatedwiththeonsetof cohabitation.We specifyan umberofh ypothesesthat predictthe conditionsunder

whichcohabitationshould occur. Wethen testourhypothesesonthebasisofanewdata setthat records ever ycase

ofcohabitationin allsemi-presidential electoraldemocraciesfr om1989to 2008inclusive. Weconfir mthatcohabi-

tationismor elikely tooccurincountrieswith apremier-presidentialform ofsemi-pr esidentialismandsho wthatit

ismore likelytofollowan electionthatoccursmidwa ythr oughaparliamentaryorpr esidentialterm, andthatwhen

cohabitationfollows apresidentialelection,it islikely todosoinacountr ywherethere isonlya veryweakpr esident.

Overall,wefindthattheconditionsunderwhichcohabitationis mostlikely tooccur arealso theoneswher eitis most

easilymanaged.Thus, ourfindingsimply thatcohabitationisnotlikely tobe asprob lematicasthe existingliterature

wouldsuggest. Keywords:semi-presidentialism;cohabitation;premier-presidentialism; president- parliamentarism;elections presidentandaprime ministerandcabinet collectively responsibletothe legislature -isan increasinglycommonform ofgov ernment.Particularlysince thefourth waveofdemocra- tisationin1990, manycountr ies,notablyinCentralandEastern Europe,theformerSoviet Unionandsub-Saharan Africa,ha veadoptedasemi-presidentialconstitution(Elgie,2007b). Oneof themostr ecognisable featuresof semi-presidentialismiscohabitation-wherethere isapr esidentfrom onepartyandapr imeministerfromanopposing party andwherethe president'spartyis notrepresentedinthe cabinet.Thister mwas firstusedtodescribethe powerwitharight-winggov ernmentled byPr imeMinisterJacquesChirac .Sincethistime, suchasMongolia, Polandand Portugal,cohabitationhasnowbecome acommonfeatur eof thepoliticalpr ocess. Formanyobserv erscohabitationistheAchilles heelofsemi-presidentialism.Whenthe presidentandtheprime minister,via theassembly ,canbothclaimto bethelegitimate sourceofpoliticalauthority ,thenin youngdemocracies thesceneisseteitherfordeadlock betweenthetwocamps, perhapsencouraging themilitarytointerv eneto breakthe doi:10.1111/j.1467-9248.2010.00870.x

POLITICALSTUDIES: 2011

©2011The Authors.PoliticalS tudies©2011PoliticalStudiesAssociation stalemate,orforeitherthepresidentor theprime ministerto seizepow erunilaterally.Linda Kirschke(2007)hasprovidede videncetosuggest thatcohabitationhas indeedbeen damagingtodemocracyin sub-SaharanAfrican countries.Indeed, inNigerwhat mightbe calleda'textbook' caseofcohabitation coincidedwiththe collapseofdemocracy in1995 (VillalónandIdrissa,2005). Inconsolidateddemocracies, cohabitationdoesnotposea threattothedemocraticsystem, but itcancause tensionswithinthe executiv e.InFrance, theconstitutionw asamended in2000todecreasethe likelihoodof cohabitationinthe future.Inshort,whetherornot cohabitationisultimately destructiveofdemocracy ,it certainlyposesproblems ofcoordination withinanyexecutive.Ther efore ,understanding theconditionsunder whichcohabitationis mostlikely tooccuris animpor tantpolitical issue. Todate,there havebeenscarcely anysystematiccross-nationalstudies ofcohabitation. Certainly,inindividualcasesthere hasbeenplenty ofanalysis,perhapsmostnotab lyin Francewhere thetopichasbeenstudiedin considerabledetail, thoughpr edominantlyby constitutionallawy ers(e.g.Cohendet,1993).Therearealsoplenty ofcasestudiesthatrefer totheconsequences ofcohabitationas partof amore generalnar rative aboutthedevel- opmentofthe politicalsystem. 1 canbefound inSamuels andShugart (2010).Aspar tofabroader projecton presidents, primeministersandpar ties,theyprovide descripti vestatisticsshowingthat cohabitationis muchmorepre valentincountrieswith thepremier-presidentialsubtypeofsemi- presidentialismthaninthosewith thepresident-parliamentar ysubtype(Sam uelsand Shugart,2010,ch.2).Thatsaid,whileDavidSamuelsandMatthe wShugart clearlyidentify thegeneralinstitutional framework inwhichcohabitation ismostlikelytooccur,the ydo notexplore themorespecificconditionsunder whichitis found.Asaresult, therear e plentyofquestions thatremain unanswer ed.Within agiven regimetype,iscohabitation merelythestochasticr esultofexo genouspoliticalf actorssuchasthepar tysystem, orisit endogenoustoparticularinstitutional arrangements,such astheelectoraltimetable? What isther elationshipbetw eenpresidentialpower andthefrequencyofcohabitation? Inthis article,wearenotconcernedwiththedurationof cohabitation.Instead,weaimto identify theconditionsthat aremost likelyto beassociatedwiththeonset ofcohabitation.Fir st,we identifynecessary ,orvirtuallynecessary,conditionsfor cohabitation.Then, wepropose someprobabilistic hypothesesabouttheconditionsunder whichcohabitationismore likely.Wetestourhypothesesonthebasisofane wdata setthatr ecordsevery caseof cohabitationinall semi-presidentialelectoral democraciesfrom 1989to2008inclusiv e.Our analysisisa combinationofdescr iptive statisticsandlo gisticr egression.Wefindthat cohabitationisstr onglyassociatedwith particulartypesofelectoralsituation. We alsofind thattheonset ofcohabitationis oftenassociatedwith theinteractionof certaintypes of electoralsituation andthedeg reeof presidentialpo werinacountry. Ourfindingsha vepolicy implications.Weshowthatcohabitationcangenerally bea voided cohabitationismost likelyto occurare alsotheoneswhereit ismost easilymanaged.Thus, eventhoughwedonotstudy theeffectsof cohabitationdirectly,our findingsstrongly imply thatcohabitationis notlikely tobeas problematic astheexistingliterature wouldsuggest.

2ROBERTELGIEAND IAINMcMENAMIN

©2011 TheAuthors.P oliticalStudies©2011PoliticalStudiesAssociation

POLITICALSTUDIES:2011

Semi-presidentialismandCohabitation

Thedefinitionof semi-presidentialismhas longbeena sourceofdebate. Inthisar ticle,w e usethedefinition thatisno wstandard (Elgie, 1999;2007a;Sam uelsandShugart,2010; Shugart,2005;Skach,2005).A countryhas asemi-presidential constitutionwhen thereis bothadir ectlyelectedfix ed-termpresidentanda primeminister andcabinetcollectively responsibletothelegislature. Theke ypointto noteaboutthisdefinitionisthatthereisno requirementforthepresidentto have 'quiteconsiderable powers'(Duverger ,1980, p.166). Inotherw ords,ther eisnobehaviouralelementtoit.The advantageofsucha definition isthat,when decidingwhichcountr iesare semi-presidential,w edo nothavetomak ea subjectivejudgementcallastowhatconstitutes 'quiteconsiderable pow ers'.W esimplyneed toread theconstitution. Whateverdefinitionweuse, semi-presidentialismis inextricablylinkedwith thenotionof cohabitation.Cohabitation ismore thanthesituation whererepresentativ esfrom different partiesholdthetw omainpositions withintheex ecutive.For example,cohabitation isnot whereonecoalitionparty holdsthepr esidencyandanother coalitionpartyholdsthe premiership.Instead,cohabitationisthesituationwherea presidentfr omone partyholds poweratthesametimeasa primeminister froman opposingpar tyandwher ethe president'spartyisnot representedinthecabinet. Giv enthatther earetwosourcesof legitimacyundersemi-presidentialism- thepresidential electionandtheparliamentary election-ther eisal waysthepotentialthat onepoliticalfor cemaywinoneelectionand thatanopposing forcema ywinthe other.Thus,thepotential forcohabitationisalwa ys presentundersemi-presidentialism.F orthisr eason,debatesaboutthepr osandconsof semi-presidentialismalways involveatleastsomediscussionoftheeffectsof cohabitation. Themostcommon argumentinf avour ofsemi-presidentialism isthat itcanallowfora degreeofpowersharing betweenopposing forces(Moestrup,2007).Onepar tycanoccupy thepresidency ,anothercanoccupythepremier shipand,ther eby, bothcanhavea stake in theinstitutionalsystem. Intheirw ork,SamuelsandShugar t(2010)pr ovideavariantofthis argument.They state(p.337):'the mostimportant potentialadvantage[ofthepr emier- presidentialsubtypeofsemi-pr esidentialism]isthe prospectof cohabitation,whichdoesnot necessarilyofferpartiesthe "bestofboth worlds"butat leastoffersthepossibilityto oscillate betweenthepresidentialand parliamentaryw orlds'.ForSam uelsandShugar tthereforethe institutionalflexibilityof semi-presidentialismis anadvantage .Thisflexibilityexpresses itselfthrough thepotentialforashiftfr omaunified executi veto cohabitation.To usetheir terms(SamuelsandShugar t,2010,p.338),the 'parliamentarisation'of thesystemunder cohabitationcanact asacounterw eighttothe presidency. Thestr engthofSamuelsandShugar t'sargumentlies inthefactthatit results froma rigorousempiricalstudy.How ever,they arealmostalone voiceintheirsupportof cohabitation.Thestandar dwisdomis thatcohabitationisproblematic .F orexample, Alfred StepanandEzra Suleiman(1995,p .399)see 'thepossibilityof constitutionalconflict betweentwoelectorallyleg itimatedexecutives[as] thecentral problem[ofsemi- presidentialism]'.SergioFabbrini(1995,p.133)statesthat '[h]ereinliesthemain weakness ofsemipresidentialism: thepossibilityofarift betweenthe president withhispopular

EXPLAININGTHEONSET OFCOHABITA TION3

POLITICALSTUDIES:2011

majorityandthepremier withhisleg islative majority.Sucha splitcouldhamper orev en paralyzethe executiv e'.JuanLinzandAlfredStepan(1996,p.286) aremoreexplicit still: Whensupporter sofoneortheothercomponentofsemi-pr esidentialismfeelthat thecountr y wouldbebetteroffif onebranchof thedemocraticallyleg itimatedstructur eof rulew ould disappearorbe closed,thedemocratic systemisendanger edandsuffer san overall lossof legitimacy,sincethosequestioningoneortheotherwilltend toconsider thepoliticalsystem undesirableaslongasthe sidethey fav ordoes notpre vail... [I]nasemipresidentialsystem, policyconflictsoften expressthemselv esasa conflictbetweentwobranches ofdemocracy. Forthesewriter scohabitationcan beharmfulfortheverysur viv alofdemocracy .How ever, inconsolidateddemocracies toocohabitationis oftenasour ceofdissatisf action.This isseen mostclearlyin France. Thefirst experienceofcohabitationfrom 1986to1988 'didnot produceafundamentalpolitical orconstitutionalcr isis',but itdidpr oduce'sev ere strains betweenthepresident andthepr imeminister'(Pierce,1991, p.287). PresidentMitter rand refusedtosigncertain gover nmentdecrees; hereturnedan umberofgovernmentbillsto parliamentforfur therconsideration;and theprimeministerencroached uponthe presi- dent's'reser veddomain'offoreignanddefencepolicy.Thethir dperiod ofcohabitation from1997to2002,which beganjusttw oyear sintoPresidentChirac's seven-y earterm, was evenmore'devastating'(Quer monne,2007,p.6) inthatit'progressi velysloweddownthe government'sworkandplacedFrance'sexternalr elationsinadifficultposition, aswit- nessedinEur opeanaffairs bytheunfortunateT reatyofNice'. 2

Thepoliticalr esponsewas

theintroduction ofthefive-year presidentialter min2000 andthedefactosynchronisation ofpresidential andlegislativeelections. Forthose infavourofthisconstitutionalreform, the desiretoav oidcohabitationw asthe'mainbenefit'(Bigaut,2000,p. 11)thatw ouldaccrue fromit. Tosummarise,cohabitation hasbeenthesourceofconsiderableacademic andpolitical debate.Todate ,thisworkhaso verwhelminglyfocusedontheeffectsofcohabitation.It has shownthatcohabitationcanha veimpor tantpositiv eor ,moreusually,negativeconse- quences.Inshor t,ithas beenshownthatcohabitationmatter s.In thisarticle ,we havea differentaim.We wishtoidentifytheconditionsunderwhichcohabitationis mostlikely tooccur. Specifically,withwhattypesof conditioniscohabitationmostlik elytobe associated?If wecan identifyregularitiesinthe onsetofcohabitation, thenwe canplacethe debateaboutthe effectsofcohabitation -goodor bad-in itsappr opriateinstitutional context. CohabitationinSemi-presidentialElectoral Democracies Wedefinecohabitationasthe situationwhere apresident from oneparty holdspow erat thesametime asapr imeministerfr omanopposing partyand wherethepresident's party isnotr epresentedin thecabinet.Whateverthepartycomposition ofthe cabinet,ifeither thepresident ortheprimeministeris non-partisanor independent,then wedonotclassthis asacase ofcohabitation.T odetermine theparty identificationof thepresidentandprime minister,werelyontheaffiliationsrecorded inwww.worldstatesmen.org.This isar eliable sourceanditstandardises suchaffiliationsas faras possible.To determinethe party

4ROBERTELGIEAN DIAINMcMENAMIN

©2011The Authors.Political Studies©2011PoliticalStudiesAssociation

POLITICALSTUDIES: 2011

compositionofcabinets, weuse secondarysour ces,includingpublicationssuch asAfrica ResearchBulletinandtheEuropeanJournal ofPoliticalResearchPoliticalData Yearbook. Weidentifyever yoccur renceofcohabitationinallcountrieswithasemi-presidential constitutionthatar erecor dedaselectoraldemocraciesby FreedomHouseto2008inclu- sive.Theirtimeseriesforelectoraldemocraciesgoes backto1989 inclusive . 3

Whilewe

couldhav eincreasedourunitsofobservation byusingdatasets withlongertime series,for example,FreedomHouse'sFreedomintheWorlddemocracy scores orPolityIVscores, we relyontheelectoraldemocracy classificationpartly becausethetime series islongenough tocapture theincreaseinthen umberofsemi-pr esidentialcountriesafter1990 andalso becausewe donothaveto imposeacut-off pointasto whatscoreconstitutesthethreshold abovewhichacountrycanbe classedasa democracy.F orthepurposesof thisstudy, therefore,ifacountrywithasemi-pr esidentialconstitutionis classedasan electoral democracy,thenwe includeitin ourdataset.Ifsucha countryis firstclassified asan electoraldemocracysome timeafter1989, wer ecordit onlyfr omtheyearwhenitachieved thisstatus.If acountrylosesthe statusofan electoraldemocracy,thenwe ceasetoinclude itin ourdataset. Howe ver, ifthesame countryisreclassifiedasanelectoraldemocracyat alaterdate ,then werecordit inourdata setfromthatyear on. Inourdata set,thebasic unitofobser vationisanelectoral period.This comprisesthe time betweenapresidentialor parliamentaryelection. Eachunitofobserv ationisa potential periodofcohabitation.We arenot concernedwiththedurationofcohabitation,only with whetherornot cohabitationoccurs inthefir stplace. Therefore,w etreat eachelectoral period,whateveritslength,asoneunitofobservation.F orexample ,Lithuaniaisfirstclassed asanelectoral democracyin 1991.How ever ,itadopted asemi-presidentialconstitutionin September1992. Theobserv ationsbegin withthefirstdirectpresidentialelection in February1993.Therewasthena parliamentaryelection inOctober/November1996,a presidentialelectioninDecember 1997/Januar y1998,a parliamentaryelection inOctober

2000,apresidentialelectionin December2002/January2003, anotherpr esidentialelection

inJune 2004followingtheimpeachment ofPresident RolandasPaksas,aparliamentary electioninOctober 2004,anda furtherparliamentar yelectioninOctober2008. Thisgi ves useightunits ofobserv ationforLithuania (1993-1996,1996-1998,1998-2000, 2000-2003,

2003-2004,2004,2004-2008,and2008-).Therearetwo periodsof cohabitation.Thefirst

wasfrom1996-1998when PresidentAlgirdasBrazauskas fromthe DemocraticLabour PartyofLithuaniawasf acedwithPr imeMinisterGediminas Vagnoriusfromthe Homeland Union-ConservativesofLithuaniawhoheadedathree-partycoalitionin whichthe president'spartyw asnotrepresented.The secondwasfrom2003-2004 whenPresident RolandasPaksas fromtheLiberalDemocratic Partywasopposite PrimeMinister Algirdas Brazauskaswhothen repr esentedtheSocial DemocraticPartyofLithuaniaandwhoalso headedacoalition inwhichthe president'spar tywas notr epresented.In theotherseven unitsofobser vation forLithuaniaeitherthepresidentwasclassed asnon-partisan orthe president'spartyw asrepresentedinthe government.Inall, wehave218unitsof observa- tion,including43 units,or periods,of cohabitation. 4

Table1liststhecountriesthat are

includedinthe study,the periodswhen theyhavebeenclassedas electoraldemocraciesand whetherthey haveapresident-parliamentar yorpremier-presidentialsubtypeofsemi-

EXPLAININGTHEONSET OFCOHABITA TION5

POLITICALSTUDIES:2011

Table1:Semi-presidentialElectoralDemocracies, 1989-2008Inclusive CountryYears includedindatasetSubtypeof semi-presidentialism

Armenia1999-2002President-parliamentary

Austria1989-President-parliamentary

Bulgaria1992-Premier -presidential

CapeVerde 1991-Premier-presidential

CentralAfricanRep. 1991-2000,2005-7President-parliamentary

Congo-Brazzaville1992-6Premier -presidential

Croatia1991-President-parliamentary (to2000)

Premier-presidential(from2001)

Finland1989-Premier -presidential

France1989-Premier -presidential

Guinea-Bissau1994-2002, 2005-President-parliamentary

Haiti1995-9,2006- Premier-presidential

Iceland1989-President-parliamentary

Ireland1989-Premier -presidential

Lithuania1993-Premier -presidential

Macedonia1992-Premier -presidential

Madagascar1993-Premier -presidential(to1995)

President-parliamentary(from1996)

Mali1992-Premier -presidential

Mauritania2007-8President-parliamentary

Moldova1995-2000Premier -presidential

Mongolia1992-Premier -presidential

Montenegro2006- Premier-presidential

Mozambique1994-President-parliamentary

Namibia1990-President-parliamentary

Niger1993-6,1999- Premier-presidential

Peru1989-91,2001- President-parliamentary

Poland1991-Premier -presidential

Portugal1989-Premier -presidential

Romania1992-Premier -presidential

Russia1993-2003President-parliamentary

SãoTomé ePríncipe1991-President-parliamentary(to 2002)

Premier-presidential(from2003)

Senegal2000-President-parliamentary

Serbia2006-Premier -presidential

Slovakia1999-Premier -presidential

Slovenia1992-Premier -presidential

SriLanka1989- President-parliamentary

Taiwan1996-President-parliamentary

Timor-Leste2002-Premier-presidential

Turkey2007-Premier-presidential

Ukraine1994-President-parliamentary (to2006)

Premier-presidential(from2007)

Sourceforelectoral democracies:FreedomHouse, http://www.freedomhouse.org/template.cfm?page=439[accessed 9February2010].

6ROBERTELGIEAND IAINMcMENAMIN

©2011The Authors.P oliticalStudies©2011PoliticalStudiesAssociation

POLITICALSTUDIES:2011

presidentialism(seeH5below). Table 2liststhe periodsofcohabitationinthesecountries. Table3listsperiodsofnon-par tisanpresidencies inthesecountr ies.Thereare 44unitsof observationwithnon-partisanpresidents. Wepr esentrelevantdescr iptivestatisticshypoth- esisby hypothesis.

Explainingthe OnsetofCoha bitation

Underwha tConditionsisCohabitationLikel ytoOccur?

Thefirst conditionistheabsenceof amajority forthepr esident'sparty intheleg islature . Whilew ewouldhesitatetoraise thisoranyconditionto thatofa true 'sociological law', wecancertainlytr eatit asadefactonecessar yconditionforcohabitation.Bydefinition, undersemi-presidentialism thegovernmentis responsible tothelegislature.Ifthepresident ispartisan andifthepresident'spar tyhasa majorityin theleg islature,thenitis difficultto imaginethecircumstancesunder whichthepr esident'spartywouldnot ber epresentedin thegov ernment.Anyattempttoinstallagovernment withoutthepresident'spar tywould resultinitbeing votedout ofofficeb ythepr esidentialmajority.Socohabitation will, inevitably,beassociatedwithsituationswherethepresident'sparty fails toenjoy thesupport ofanabsolute majorityin theleg islature. H1.Cohabitationwill occurwhenthe president'spar tyhasless than50 percentof theseats intheleg islature.

Table2:PeriodsofCohabitation inSemi-presidential

ElectoralDemocracies,1989-2008 Inclusive

CountryPeriodof cohabitationtoend 2008

Austria2004-7

Bulgaria1995-7,2001-2, 2002-5

Finland1991-4,1994-5, 2007-

France1993-5,1997-2002

Iceland1996-9,1999-2000, 2000-3,2003-4,2004-7

Ireland1990-2,1997

Lithuania1996-8,2003-4

Macedonia2002-4,2006-8, 2008-

Mongolia1993-6, 1997-2000

Niger1995-6

Poland1991-3,1993-5, 1997-2000,2000-1,2007-

Portugal1987-91,1991, 1991-5,2002-5,2006-

Romania2007-8

SãoTomé 1994-6,1996-9,1999-2001,2004-5

Serbia2006-7

Slovakia2004-6

Slovenia2004-6

SriLanka2001-4

EXPLAININGTHEONSET OFCOHABITA TION7

POLITICALSTUDIES: 2011

Unsurprisingly,theempiricalevidencesupportsthishypothesis.W ecalculatedthepercent- ageofseats heldby thepresident's partyin eachoftheunitsof observation andwefound thatall oftheper iodsofcohabitation occurred whenthepresident'sparty failedto enjoy majoritysupportin thelegislature.F orexample, inPor tugaltheSocialistpartywonthe parliamentaryelectioninFebr uary2005, returning 52.6percentofthetotalnumberof deputiestothe legislature andforming asingle-partygovernment.At thetime, the incumbentpresident wasJorgeSampaio, alsorepresentingthe Socialistparty.Sother ewas nocohabitation.Ho we ver,inJanuary2006AníbalCavacoSilva,representing thecentre- rightSocialDemocraticparty ,won thepresidential election.TheSocialDemocratsenjoyed thesupport of32.6percentofdeputies intheleg islature thathadbeen electedthepr evious year.WiththeSocialistsstillenjoyinganabsolutemajority there, thegov ernment didnot changeanda periodof cohabitationbegan.This exampleclearlyshowsthat theonsetof cohabitationwas associatedwiththeshiftfrom asituationwher ethe president'spar ty enjoyedmajoritysupportin thelegislaturetoone wherethepresident's partyw asina minority. Whilecohabitation isalmostinconcei vable insituationswher ethepresident'spartyenjoys majoritysupportinthe legislature,suchan outcomeisr elativ elyuncommon.Infact,inour datasetthe president'spar tyenjoy edthesupportof anabsolutemajorityinthe legislature inonly40 of218 observations (18.3percent) intotaland in40ofthe175observations (22.9percent) thatremain whenwe excludetheobser vationswherethe presidentw as non-partisan,andwhere, bydefinition, thepresident'spartycouldnotenjo ymajority support.Therefore, whiletheabsenceofanabsolutemajorityforthepr esident'sparty in Table3:UnitsofNon-partisan Presidenciesin Semi-presidentialElectoralDemocracies,

1989-2008Inclusive

CountryNon-partisanpresidents toend2008

CentralAfricanRep. 2005-7(president-parliamentary) Croatia2000-3,2003-5, 2005-7,2007- (premier-presidential) Iceland1988-91,1991-2, 1992-5,1995-6(president-parliamentary) Lithuania1998-2000,2000-3, 2004,2004-8(premier -presidential) Macedonia1992-4,1994-8, 1998-9(premier-presidential)

Mali2002-7,2007- (premier-presidential)

Mauritania2007(president-parliamentary)

Moldova1995-6,1996-8, 1998-2000(premier -presidential) Russia1993-5,1995-6, 1996-9,1999-2000, 2000-3(president-parliamentary) SãoTomé 1991-4,1994-6(president-parliamentary)

Slovakia2006-(premier -presidential)

Slovenia1992-6,1996-7, 1997-2000,2000-2, 2006-7(premier-presidential) Ukraine1994-8,1998-9, 1999-2002,2002-4(president-parliamentary)

8ROBERTELGIEAND IAINMcMENAMIN

©2011The Authors.P oliticalStudies©2011PoliticalStudiesAssociation

POLITICALSTUDIES:2011

thelegislatur eisanecessaryconditionforcohabitation,w estillneed toidentify the circumstancesinwhichcohabitationoccur swhenthe president'spar tydoes notenjoy majoritysupportther e. Inthiscontext, webeg inbynotingthatcohabitationcanoccuras aresultofthr eetypesof electoralsituation: whensynchronisedpresidentialand legislativeelectionsr eturnopposing majorities;whenapr esidentialelectionr eturnsa candidatewhoisinoppositionto the incumbentlegislati vemajority(asinthepreviouslycited Portugueseexample);andwhen alegislati veelectionreturnsamajoritythatisopposed tothe incumbentpresident(as happenedinFrance in1986,1993 and1997).The onlyothersituation underwhich we wouldobserve cohabitationisifitweretooccur midway through anelectoralper iod.We hypothesisethatthislatterscenar ioishighly unlikely. Cohabitationcould occurifthe president'spartywer etowithdrawunilaterallyfr omthegovernment.How ever ,allelse beingequal,the president'spar tyisunlik elytodosobecause itwouldleave thepresident totallyalonewithin theexecuti ve. Cohabitationcouldalso occuriftheincumbentgov- ernmentwasvoted downbytheleg islatureandanew governmentw asformedthat excludedthe president'spar ty.Howev er,ifthelegislatureweretovotedownagovernment thatincludedthe president'spar ty,it islikelythatthe presidentwouldstillbeabletoavoid cohabitation.Byvirtueofhis orherpopularlegitimacy, thepresident may beina position toinsiston hisorher partybeing includedinthe new gover nment,ortorequire a non-partytechnicalgov ernment.Alter natively,thepresidentmaydecideto dissolvethe legislatureratherthanacceptanunexpectedper iodofcohabitation andpotential lossof influence.Sowhilecohabitation canoccuroutside anelectoralcontext, itisv eryunlik ely todoso .

H2.Cohabitationonly occursafter elections.

Ourdataset stronglysuppor tsthish ypothesis.Onlytwoofthe 43casesofcohabitation beganmidwa ythroughanelectoralperiod:inRomania fromApril2007 untiltheelection ofDecember2008 andinSão Tomée Príncipefrom March 2004untilthepresident'sparty rejoinedthegover nmentinJ une2005. 5

TheRomaniancase illustratestheexceptional

circumstancesthatneedtocombine toleadto cohabitationoccurr ingoutside anelectoral context.Follo wingthe2004legislativeelection, PresidentBa ˘sescu'sDemocraticLiberal party(PD-L)waspar tofthe governingcoalitionled byPrimeMinister Popescu-Ta˘riceanu oftheNational Liberalparty (PNL)thatincluded thePNL,the PD-Landtwoother thatinApr il2007Pr imeMinisterPopescu-Ta˘riceanuproposedacabinetreshufflethat excludedthePD-L ministersfr omthene wgovernment (StanandZaharia,2008,p .1120). Acoupleof weekslater ,parliamentv otedtoimpeachPresidentBa ˘sescu,accusinghimof authoritarianism.InMayareferendumw asheldtovalidatethe impeachmentvote .Turnout wasrelatively low,butthevotewasov erturned.Boththepresidentand thegovernment remainedinofficeandcohabitation continueduntil thelegislati ve electioninDecember

2008,but thefactthattheonset ofcohabitationmidw aythroughanelectoral period

coincidedwithparliament votingto impeachapr esidentandarefer endumonthe president'ssurvival inofficeshowsthatthisscenarioislikely tobeassociated withonlyan

EXPLAININGTHEONSET OFCOHABITA TION9

POLITICALSTUDIES:2011

exceptionalpoliticalsituation. Overall,our findingsprovideoverwhelmingsupportforthe propositionthatcohabitationislik elytofollo wanelection. 6 Giventhatcohabitationishighlyunlik elytooccur outsideanelectoral context,the questionarises astowhetherparticulartypes ofelectionar emor eorless likelytoleadto cohabitation.Inthisregard, wecan hypothesisethatcohabitationisunlikelytooccurwhen presidentialandparliamentaryelections aresynchr onised.Forcohabitationtooccur under suchconditions,ther ewouldhavetobe considerablesplit-ticketvoting. Whilepresidential electionsar enecessarilymajoritarian, thelegislativeelection wouldhavetoretur na coherentmajoritythatw asactivelyopposedto thepr esident.Thisispossible, butunlikely. Asnotedpr eviously ,inFrancethelengthofthepresident'smandatewas changeddelib- eratelytominimise thefutur elikelihood ofcohabitation.By contrast,in2003thepresi- dent'sterm inRomaniawasincr easedfrom fourtofiv eyears.Pre viously,presidentialand parliamentaryelectionshadbeenheld simultaneouslywith theformer actingas a'trail- blazer'intermsofgov ernmentselection(Ta˘na˘sescu,2008,p.69).By'desynchronising'the elections,the2003 refor mwas notdesignedspecificallytoencouragecohabitation,butit waspassedwithavie wtor educingthepr e-eminenceofthe presidentinthepolitical process,whichisconsistentwith thegeneraleffect ofcohabitation. H3.Cohabitationwill notoccurwhen presidentialand legislativ eelectionsar esynchr onised. Again,descriptivestatisticsaresufficienttoprovidestrong support forthishypothesis.In50 ofour218 observations (22.9percent) presidentialandlegislativeelections were held withinthree monthsofeachother.F orexample, inMozambique ,Namibiaand Romania (until2008)the constitutionspecifiedsim ultaneouselections.In othercountries, simulta- neous(orquasi-sim ultaneous)electionsw eresimplyafunction oftheelectoral timetable, aswiththe above caseofSlo veniain1992.However ,none ofthese50 observationsresulted incohabitation.In otherwor ds,thee videncestrongly suggeststhatcohabitationisover- whelminglylikely tooccuraftera'mid-term' election,thatis tosay following alegislati ve electionthatoccur satsome pointduringapresident's term orfollowing apresidential electionthatoccur satsome pointduringalegislatur e'ster m. Wehavealr eadyshownthatcohabitationisunlikelye vertobeassociated withthesituation wherethepresident'spar tyenjo ysanabsolutemajorityin thelegislature.By thesamelogic , evenwhenthepresident'sparty doesnotenjo ymajority support,wew ouldexpect cohabitationtobe lesslikely whenthepr esident'sparty hasmoresupport intheleg islature thanwhenthe president's partyis onlyasmallforce. True,ifther ew asatw o-partysystem, thenthepr esident'spar tycouldwinaver ylargeminorityofseats inthe legislature and cohabitationcouldstill occurbecausethe oppositionwould hav eamajor ity.Ho wev er,all elsebeingequal, ifthe president'spar tyhasa very largeminorityofseats,then itishighly likelythatthepresident's partywill beanindispensab lepartofthe coalition-building process,thusensuringthatcohabitationis avoided.Bycontrast,if thepresident headsa party withlittlesuppor tinthe legislature,then,e ven thoughthepr esidentmayhaveconsiderable personallegitimacyb yvirtueofbeingdir ectlyelected,theparliamentaryar ithmeticmay meanthatthe president'spar tyisnot requiredfortheconstr uctionofastable legislativ e majority.

10ROBERTELGIEAND IAINMcMENAMIN

©2011The Authors.Political Studies©2011PoliticalStudiesAssociation

POLITICALSTUDIES: 2011

H4.Whenthe president'spar tydoesnot havemajority supportinthelegislatur e,cohabitation islesslik elyto occurastheshareof theseatsheld bythe president'spar tyincr eases. Totestthishypothesis, weperfor malo gisticregressionofcohabitationaspredicted bythe shareoftheseatsheld bythe president'spar tyinthe legislatur ewhentheseatshare is50 percentor less.Thesample alsoexcludesnon-par tisanpresidents. Themeanpr esidential sharewasalmost 31percent,withastandar ddeviation of13,a minimumof zero anda maximum,byconstruction, of50.Unsurprisingly, asTable4 shows,w efindthat this variableisquitepowerful,explainingo ver11 percentof thevariationinthissample.Thequotesdbs_dbs21.pdfusesText_27

[PDF] cohabitation rules in germany

[PDF] cohabitation visa germany

[PDF] cohesion markers

[PDF] cohesion pdf

[PDF] cohesive devices examples pdf

[PDF] cohesive devices exercises pdf

[PDF] cohesive devices exercises with answers

[PDF] cohesive devices in narratives

[PDF] cohesive devices list pdf

[PDF] cohesive devices pdf ielts advantage

[PDF] cohesive devices worksheet pdf

[PDF] cold war causes

[PDF] cold war conflicts in asia

[PDF] cold war containment timeline

[PDF] cold war map