JOURNAL OF

DEMOCRACYWiuter 1990 97.50

Tiananmen and BeyondWuer Kaixi Merle Goldman

Fang Lizhi m ThomasGold

The Struggle Against NoriegaRoberto Eisenmann

Uncertaintiesof a DemocraticAgeLeszekKolakowski

The Perils of PresidentialismJuan Linz

The Crumbling of the Soviet Blocbeck Kurod Jdnos Kis

Vladimir Bukovsky Vilém Predan

everything in the indefinite future andto the firm belief that this is whateach of us deserves.But these hopesare bound to end up in bitterdisappointment,at least for the overwhelmingmajority of the world’speople.

Overpopulation,shrinking resourcesof agriculturalland andof water,and ecologicalcatastropheswill certainly compel mankind in the nearfuture to devote more and more effort and money to repairing thedamagesalreadyinflicted on our environmentandto warding off furthercalamities.This will not only leadto growing restrictionsimposedon ourfreedomof movementandproperty rights. It will result, above all, in adampeningof our hopes for “more and more and, indeed, in thedemandthat we recognizethat we have enough, or even that we mustmanagewith less, limit our wants, and accepta more modest life. Theamount of frustration, irrational rage, and aggressivenessthat theseimperativesaregoing to causewill be enormous,andwill affect the poorand the rich alike. For the degreeof frustration dependsnot on theabsolute level of satisfaction, but on the distance betweenthis absolutelevel and our subjective needs, and our needs can and do expandindefinitely alongthe endlessspiral of greed. It is hardto predictwhatideological expressionor other channelsthis frustrationmight find, butin order to tame it and to preventsociety from plunging into chaosorfalling prey to a lawless tyranny, it is likely that many undemocraticrestrictionswill be needed.

Widespreadmisery is fertile ground for the successfuldemagogyoftotalitarianmovementsand for the temptationto “solve socialproblemsby means of military dictatorship. We have seen this many times,especiallyin Latin Americaandin Africa. If the relatively rich countrieswill be compelledto lower human expectations—evenwithout causingreal misery—it can only addto the danger.

This is not to say that the causeof freedomis lost; we have enoughevidenceto concludethat peopleneednot only security but freedomaswell. But we must also never forget that freedomis always vulnerableandits causenever safe.

THE PERILSOF PRESIDENTIALISM

Juan J. Linz

Juan J. Linz, Sterling Professorof Political andSocial Scienceat YaleUniversiC is widely known for his contributions to the study ofauthorita,~ja,

2ism and totalitarianism, political pa/ties and elites, and

democraticbreakdownsand transitions to democracy In 1987 he wasawardedSpain’s Principe de Asturias ~ in the social sciences.Thefollowing essay is basedon a paper he presentedin May 1989 at aconferencein Washington,D.C. organizedby the Latin AmericanStudiesProgram of Georgetown UniversiG with support from the FordFoundation An annotated revised, and expandedye/-sianof this essay(includinga discussionof sem1~residentialsystems)will appearunderthetitle “Presjdentjaljsm andParl/amentarj’sn2 DoesIt Makea Difference?”in a publication basedon the conferencebeing editedby the author andProfessorArturo Valenzuelaof GeorgetownUniversiG.

A~more of the world’s nations turn to democracy, interest inalternative constitutional forms and arrangementshas expandedwellbeyond academic circles. In countries as dissimilar as Chile, SouthKorea, Brazil, Turkey, and Argentina, policymakers and constitutionalexpertshavevigorously debatedthe relative merits of different types ofdemocraticregimes.Some countries,like Sri Lanka, haveswitchedfromparliamentary to presidential constitutions. On the other hand. LatinAmericansin particularhave found themselvesgreatly impressedby thesuccessfultransition from authoritarianismto democracythat occurredinthe l970s in Spain, a transition to which the parliamentary form ofgovernmentchosenby that country greatly contributed.

Nor is the Spanishcase the only one in which parliamentarismhasgiven evidenceof its worth. Indeed, the vast majority of the stabledemocracies in the world today are parliamentary regimes, whereexecutive power is generatedby legislative majorities and dependsonsuch majorities for survival.

By contrast,the only presidentialdemocracywith a long history of

52 10(11710/ of Den 001001 11(01? 1 LII:

constitutionalcontinuity is the United States.The constitutionsof FinlandandFranceare hybrids ratherthan true presidentialsystems,and in thecase of the FrenchFifth Republic, the jury is still out. Aside from theUnited States,only Chile has manageda centuryanda half of relativelyundisturbedconstitutionalcontinuity under presidentialgovernment—butChilean democracybroke down in the l970s.

Parliamentaryregimes, of course, can also be unstable, especiallyunderconditionsof bitter ethnic conflict, as recentAfrican history attests.Yet the experiencesof India and of some English-speakingcountries inthe Caribbean show that even in greatly divided societies, periodicparliamentarycrises need not turn into full-blown regime crises and thatthe ousting of a prime minister and cabinet need not spell the end ofdemocracyitself.

The burdenof this essayis that the superiorhistorical performanceofparliamentary democraciesis no accident. A careful comparison ofparliamentarism as such with presidentialism as such leads to theconclusion that, on balance, the former is more conducive to stabledemocracythan the latter. This conclusionapplies especiallyto nationswith deep political cleavagesand numerouspolitical parties; for suchcountries, parliamentarismgenerally offers a better hope of preseningdemocracy.

Parliamentary vs. Presidential Systems

A parliamentary regime in the strict sense is one in which the onlydemocratically legitimate institution is parliament; in such a regime, thegovernment’s authority is completely dependent upon parliamentaryconfidence.Although the growing personalizationof party leadershipinsome parliamentaryregimes has made prime ministers seem more andmore like presidents,it remainstrue that barring dissolutionof parliamentand a call for new elections, premiers cannot appeal directly to thepeopleover theheadsof their representatives.Parliamentarysystemsmayinclude presidentswho are elected by direct popular vote, but theyusually lack the ability to competeseriously for power with the primeminister.

In presidential systemsan executive‘~ ith considerableconstitutionalpowers—generallyincluding full control of the composition of thecabinet and administration—isdirectly electedby the people for a fixedterm and is independentof parliamentaryvotesof confidence. He is notonly the holder of executivepower but also the symbolic head of stateand can be removed between elections only by the drastic step ofimpeachment.In practice, as the history of the United States shows.presidentialsystemsmay be more or less dependenton the cooperationof the legislature; the balancebetweenexecutive and legislativepowerin such systemscan thus vary considerably.

Two things aboutpresidentialgovernmentstand out. The first is thepresident’sstrongclaim to democratic,evenplebiscitarian,legitimacy; thesecondis his fixed term in office. Both of thesestatementsstand in needof qualification. Some presidentsgain office with a smallerproportionof the popular vote than many premiers who head minorit? cabinets,although voters may see the latter as more weakly legitimated. Tomention just one example, Salvador Allende’s election as presidentofChile in 1970—he had a 36.2-percent plurality obtained by aheterogeneouscoalition—certainly put him in a position very differentfrom that in which Adolfo Suárezof Spain found himself in 1979 whenhebecameprime minister after receiving35.1 percentof the vote. As wewill see, Allende received a six-year mandate for controlling thegovernmenteven with much less than a majorit) of the popular vote,while Suárez, with a plurality of roughly the same size, found itnecessaryto work with other parties to sustain a minority government.Following British political thinker Walter Bagehot,we might say that apresidentialsystem endows the incumbentwith both the “ceremonial’functions of a head of state and the “effective” functions of a chiefexecutive, thus creating an aura, a self-image, and a set of popularexpectationswhich are all quite different from those associated ith a

prime minister, no matter how popular he may be.But what is most striking is that in a presidential system, the

legislators,especiallywhen they representcohesive,disciplined partiesthat offer clear ideological and political alternatives, can also claimdemocratic legitimacy. This claim is thrown into high relief when amajority of the legislaturerepresentsa political option opposedto theone the president represents.Under such circumstances,who has thestronger claim to speak on behalf of the people: the presidentor thelegislative majority that opposeshis policies? Since both derive their

power from the votes of the people in a free competition amongwell-defined alternatives,a conflict is always possible and at times mayerupt dramatically. Theme is no democratic principle on the basis ofwhich it can be resolved, and the mechanismsthe constitution mightprovide are likely to prove too complicatedand aridly legalistic to be ofmuch force in the eyes of the electorate.It is thereforeno accidentthatin somesuchsituations in the past,the armedforceswere often temptedto intervene as a mediating power. One might argue that the UnitedStates hassuccessfullyrenderedsuchconflicts “normal” and thusdefusedthem. To explain how Americanpolitical institutions and practiceshaveachievedthis result would exceedthe scope of this essay,but it is worthnoting that the uniquely diffuse character of American politicalparties—which,ironically, exasperatesmany Americanpolitical scientistsand leads them to call for responsible, ideologically disciplinedparties—hassomething to do with it. Unfortunately, the American caseseemsto be an exception; the developmentof modernpolitical parties,

JUG?? I Li,,:

particularly in socially and ideologically polarized countries,generallyexacerbates,rather than moderates,conflicts betweenthe legislative andthe executive.

The secondoutstandingfeatureof presidentialsystems—thepresident’srelatively fixed term in office—is also not without drawbacks.It breaksthe political process into discontinuous, rigidly demarcatedperiods,leaving no room for the continuous readjustmentsthat events maydemand. The duration of the president’s mandate becomesa crucialfactor in the calculationsof all political actors,a fact which (as we shallsee) is fraught with important consequences.Consider, for instance, theprovisionsfor successionin caseof the president’s deathor incapacity:in somecases,the automaticsuccessormay have beenelectedseparatelyand may representa political orientationdifferent from the president’s:in other cases,he may have been imposed by the president as hisrunning mate without any consideration of his ability to exerciseexecutivepoweror maintain popular support.Brazilian history providesus with examplesof the first situation,while Maria EstelaMartinez dePerOn’s successionof hem husband in Argentina illustrates the second,II is a paradoxof presidential government that while it leads to thepersonalizationof power, its legal mechanismsmay also lead, in theeventof a suddenmidterm succession,to the rise of someonewhom theordinary electoralprocesswould neverhave madethe chief of state.

Paradoxes of Presidentialism

Presidential constitutions paradoxically incorporate contradictoryprinciples and assumptions.On the one hand, such systemsset out tocreatea strong, stableexecutivewith enoughplebiscitarian legitimationto stand fast against the array of particular interests representedin thelegislature. In the Rousseauianconceptionof democracyimplied by theidea of ‘the people,” for whom the presidentis supposedto speak,theseinterests lack legitimacy: so does the Anglo-American notion thatdemocracy naturally involves a jostle—or even sometimes a melee—ofinterests.Interestgroup conflict then bids fair to manifest itself in areasother than the strictly political. On the other hand, presidentialconstitutions also reflect profound suspicion of the personalizationofpower: memoriesand fears of kings and caudillosdo not dissipateeasily.Foremostamongthe constitutionalbulwarks againstpotentially arbitrarypower is the prohibition on reelection.Other provisions like legislativeadvice-and-consentpowersover presidentialappointments,impeachmentmechanisms, judicial independence,and institutions such as theContralorIa of Chile also reflcct this suspicion. Indeed. politicalintervention by the armedforcesacting as a podermaderadarmay evenbe seen in certainpolitical cultures as a useful check on overweeningexecutives.One could explore in depth the contradictionsbetweenthe

constitutionaltexts and political practicesof Latin American presidentialregimes;any studentof the region’s history could cite many examples.

It would be useful to explore the way in which the fundamentalcontradictionbetweenthe desire for a strongand stableexecutiveand thelatent suspicionof that samepresidentialpoweraffects political decisionmaking. the style of leadership,the political practices,and the rhetoricof both presidents and their opponents in presidential systems. Itintroduces a dimensionof conflict that cannot be explained wholly bysocioeconomic.political, or ideological circumstances.Even if one wereto acceptthe debatablenotion that Hispanicsocietiesare inherentlyproneto personalisma, there can be little doubt that in some cases thistendencyreceivesreinforcementfrom institutional arrangements.

Perhaps the best way to summarizethe basic differences betweenpresidential and parliamentary systems is to say that whileparliamentarismimparts flexibility to thepolitical process,presidentialismmakes it rather rigid. Proponentsof presidentialismmight reply that thisrigidity is an advantage, for it guards against the uncertainty andinstability so characteristicof parliamentarypolitics. Underparliamenta~government,after all, myriad actors—parties,their leaders,even rank-and-file legislators—may at any time between elections adopt basicchanges, cause realignments, and, above all, make or break primeministers.But while the need for authority and predictability would seemto favor presidentialism,there are unexpecteddevelopments—rangingfrom the deathof the incumbentto seriouserrors in judgmentcommittedunder the pressureof unruly circumstances—thatmake presidential ruleless predictableand often weaker than that of a prime minister. Thelatter can always seek to shore up his legitimacy and authority, eitherthrough a vote of confidenceor the dissolution of parliament and theensuing new elections. Moreover, a prime minister can be changedwithout necessarilycreating a regime crisis.

Considerationsof this sort loom especially large during periods ofregime tran~itionand consolidation,when the rigidities of a presidentialconstitution must seeminauspiciousindeedcomparedto the prospectofadaptability that parliamentarismoffers.

Zero-sum Elections

The precedingdiscussionhas focusedprincipally on the institutionaldimensions of the problem; the consideration of constitutionalprovisions—somewritten, someunwritten—hasdominatedthe analysis.In addition, however, one must attend to the ways in which politicalcompetition is structuredin systemsof direct presidentialelections; thestyles of leadershipin suchsystems;the relationsbetweenthe president,the political elites, and society at large; and the ways in which poweris

exercised and conflicts are resolved. It is a fair assumption that

lila!? I LI/il

institutional arrangementsboth directly and indirectly shape the entirepolitical process, or “way of ruling.’ Once we have described thedifferencesbetv

1eenparliamentaryand presidential formsof government

that result from their differing institutional arrangements,we shall beready to ask which of the two forms offers the best prospect forcreating,consolidating,and maintaining democracy.

Presidentialismis ineluctablyproblematicbecauseit operatesaccordingto the rule of “winner-take-all—-an arrangementthat tends to makedemocraticpolitics a zero-sumgame, with all the potential for conflictsuch gamesportend. Although parliamenta~elections can produceanabsolutemajority for a single party, they more often give representationto a number of parties. Power-sharingand coalition-forming are fairlycommon, and incumbentsare accordingly attentive to the demandsandinterests of even the smaller parties. These parties in turn retainexpectationsof sharing in powerand, therefore,of having a stake in the

system as a whole. By contrast, the conviction that he possessesindependent authority and a popular mandate is likely to imbue apresidentwith a senseof power and mission, even if the plurality thatelectedhim is a slenderone. Given suchassumptionsabouthis standingand role, he will find the inevitable opposition to his policies far moreirksome and demoralizing than would a prime minister, who knowshimself to be but the spokesmanfor a temporarygoverning coalitionrather than the voice of the nation or the tribune of the people.

Absent the support of an absolute and cohesive majority, aparliamentary system inevitably includes elements that becomeinstitutionalized in what has been called “consociational democracy.”Presidentialregimes may incorporate consociational elements as well,perhaps as part of the unwritten constitution. When democracy wasreestablishedunder adverse circumstancesin Venezuelaand Colombia,for example, the written constitutions may have called for presidentialgosernment, but the leaders of the major parties quickly turned toconsociationalagreementsto softentheharsh,winner-take-allimplicationsof presidentialelections.

The danger that zero-sumpresidentialelectionspose is compoundedby the rigidity of the president’sfixed term in office. Winners and losersare sharply defined for the entire period of the presidentialmandate.There is no hope for shifts in alliances,expansionof the government’sbase of support through national-unity or emergency grand coalitions,new elections in responseto major new events,and so on. Instead,thelosers must wait at least four or five years without any accesstoexecutive po’aer and patronage. The zero-sum game in presidentialregimes raises the stakes of presidential elections and inevitablyexacerbatestheir attendanttension and polarization.

On thc other hand, presidential elections do offer the indisputableadvantageof allowing the people to choosetheir chiefexecutiveopenly,

directly, and for a predictablespan rather than leaving that decision tothebackstagemaneuveringof thepoliticians. But this advantagecan onlybe present if a clear mandate results. If there is no required minimumplurality and several candidatescompete in a single round, the margin_____________________________bet’aeen the victor and the runner-up may

be too thin to support any claim that adecisive plebiscite has taken place. Topreclude this. electoral laws sometimesplace a lower limit on the size of thewinning plurality or createsomemechanismfor choosing among the candidatesif noneattains the minimum number of votesneeded to win; such proceduresneed notnecessarily award the office to thecandidate with the most votes. Morecommon are run-off provisionsthat set tipa confrontation between the two majorcandidates,with possibilities for polari7ationthat have already been mentioned.One of

the possible consequencesof two-candidateracesin multiparty systemsis that broad coalitions are likely to be formed (whether in run-offs orin preelection maneuvering) in which extremist parties gain undueinfluence. If significant numbersof voters identify strongly with suchparties. one or more of them can plausibly claim to represent thedecisive electoral bloc in a close contest and may make demandsaccordingly. Unless a strong candidateof the centerrallies widespreadsupport against the extremes,a presidentialelection can fragment andpolarize the electorate.

In countries ‘a here the preponderanceof voters is centrist, agreesonthe exclusion of extremists, and expects both rightist and leftistcandidates to differ only within a larger, moderate consensus, thedivisiveness latent in presidentialcompetition is not a seriousproblem.With an overwhelmingly moderate electorate, anyone who makesalliancesor takes positionsthat seem to incline him to the extremesisunlikely to win, as both Barry Goldwater and George McGoverndiscovered to their chagrin. But societies beset by grave social andeconomicproblems,divided aboutrecentauthoritarianregimesthat onceenjoyed significant popular support, and in ‘a hich well disciplinedextremist partieshave considerableelectoralappeal.do not fit the modelpresentedby the United States. In a polarized society with a volatileelectorate,no seriouscandidatein a single-roundelection can afford toignore partieswith which he would otherwisenever collaborate.

A two-round election can avoid some of these problems, for thepreliminary round shows the extremistpartiesthe limits of their strengthand allows the two major candidatesto reckonjust which alliancesthey

“In a polarizedsociet~jwith avolatile electorate,no seriouscandidate in asingle-roundelection can affordto ignore partieswith which hewould otherwisenever collaborate.”

must make to sUn. This reducesthe degreeof uncertaintyand promotesmore rational decisions on the part of both voters and candidates.Ineffect, the presidential system may thus reproducesomething like thenegotiationsthat ‘form a government”in parliamentaryregimes.But the

potential for polarization remains, as does the difficulty of isolatingextremist factions that a significant portion of the voters and elitesintensely dislike.

The Spanish Example

For illustration of the foregoing analysis, considerthe caseof Spainin 1977, the year of the first free election after the death of FranciscoFranca.The parliamentary elections held that year allowed transitionalprime minister Adolfo Suhrezto remain in office. His moderateUniondel Centro Democratico(UCD) emergedas the leadingparty with 34.9percent of the vote and 167 seats in the 350-seat legislature. The

Socialist Party (PSOE), led by Felipe Gonzalez, obtained 29.4 percentand 118 seats,followed by the CommunistParty (PCE) with 9.3 percentand 20 seats, and the rightist Alianza Popular (AP), led by ManuelFraga, with 8.4 percentand 16 seats.

Theseresults clearly show that if insteadof parliamentaryelections,a presidentialcontesthad beenheld, no party would have had more thana plurality. Candidateswould have been forced to form coalitions to havea chanceof winning in a first or secondround. Prior to the election,however, there was no real record of the distribution of the electorate’spreferences.In this uncertainatmosphere,forming coalitions would haveproven difficult. Certainlythe front-runnerswould have foundthemselvesforced to build unnecessarilylarge winning coalitions.

Assuming that the democraticoppositionto Francawould haveunitedbehind a single candidatelike Felipe Gonzáles(something that was farfrom certain at the time), and given both the expectationsabout thestrength of the Communists and the ten percentof the electoratetheyactually represented, he would never have been able to run asindependentlyas he did in his campaign for a seat in parliament. Apopular-frontmentality would havedominatedthecampaignandprobablysubmerged the distinct identities that the different parties, from theextremists on the left to the Christian Democratsand the moderateregional parties in thecenter,were able to maintain in most districts. Theproblem would have beeneven more acute for the center-rightistswhohad supportedreforms,especiallythe i’eforma pactadathat effectively putan end to the authoritarianregime. It is by no meanscertain that AdolfoSuarez, despite the great popularity he gained during the transitionprocess,could or would have united all those to the right of theSocialistParty. At that juncture many Christian Democrat~,including those whowould later run on the UCD ticket in 1979, would not have beenwilling

to abandon the political allies they had made during the years ofoppositionto Franca:on the other hand, it would have beendifficult forSukrez to appearwith the supportof the rightist AP, since it appearedto representthe “continuist” (i.e., Francoist)alternative.For its part. theAP would probably not have supported a candidate like Suáres whofavored legalization of the CommunistParty.

Excluding the possibility that the candidateof the right would havebeen Fraga (who later became the acceptedleader of the opposition).Suárezwould still have been hard-pressedto maintain throughout thecampaign his distinctive position as an alternative to any thought ofcontinuity with the Francaregime. Indeed, the UCD directed its 1977campaignas much against the AP on the right as againstthe Socialists

on the left. Moreover,given the uncertaintyabout the AP’s strength and the fear andloathing it provoked on the left, muchleftist campaigningalso targetedFraga.Thishad the effect of reducing polarization.especiallybetweenlongtime democrats,onthe one hand,and newcomersto democraticpolitics (who comprised important segmentsof both the UCD’s leadershipand its rankand file), on the other. Inevitably, thecandidate of the right and centerrightwould have focusedhis attackson the left-democratic candidate’s “dangerous”

supporters. especially the Communists and the parties representingBasque and Catalan nationalism. In replying to these attacks thecandidateof the left and center-left would certainly have pointed to thecontinuity between his opponent’s policies and those of Franca, theputative presenceof unreconstructedFrancoistsin the rightist camp,andthe scarcityof centristdemocratsin the right wing coalition.

There can be np doubt that in the Spain of 1977, a presidentialelection would have been far more divisive than the parliamentaryelections that actually occurred. Had Suárezrejected an understandingwith Fraga and his AP or had Fraga—misled by his own inflatedexpectationsabout the AP’s chancesof becomingthe majority’ party ina two-party system—rejectedanyalliancewith the Suaristas,the outcomemost likely would have been a plurality for a candidate to the left ofboth Suárezand Fraga. A presidentwith popular backing, even withouta legislative majority on his side, would have felt himself justified inseeking both to draft a constitution and to push through political andsocial changesfar more radical than those the Socialist Prime MinisterFelipe Gonzalez pursuedafter his victory in 1982. It is important torecall that Gonzalezundertook his initiatises when Spain had already’experiencedfive yearsof successfuldemocraticrule, and only after both

“There can be nodoubt that in theSpainof 1977, cipresidentialelectionwouldhavebeenfarmore divisive thanthe parliamentaryelections...”

60mu,,, ii of Deii?ocra(1 Jim,, I Li,,: (1

a party congressthat saw the defeatof the PSOE’sutopian left wing anda campaign aimed at winning over the centrist majority of Spanishvoters. Spanish politics since Franca has clearly felt the moderatinginfluence of parliamentarism; without it, the transition to populargovernment and the consolidation of democraticrule would probablyhave taken a far different—and much rougher—course.

Let me now add a moderatingnote of my own. I am not suggestingthat the polari7ationwhich often springs from presidentialelectionsis aninevitableconcomitantof presidentialgovernment.If thepublic consensushasers reliably around the middle of the political spectrum and if thelimited weight of the fringe parties is in evidence,no candidatewill haveany’ incentive to coalescewith the extremists.They may run for office,but they will do so in isolation and largely’ as a rhetorical exercise.Under these conditions of moderation and preexisting consensus,presidential campaignsare unlikely to prove dangerously divisive. Theproblem is that in countries caught up in the arduous experienceofestablishingandconsolidating democracy,such happy circumstancesareseldom present.They certainly do not exist when there is a polarized

multiparty system including extremist parties.

The Style of Presidential Politics

Since we have thus far focused mostly’ on the implications ofpresidentialismfor the electoral process,one might reasonablyobservethat while the election is one thing, the victor’s term in office is another:once he has won, can he not sethimself to healing the woundsinflictedduring the campaignand restoring the unity of the nation? Can he notoffer to his defeatedopponents—butnot to the extremistelementsof hisown coalition—a role in his administration and thus make himselfpresident of all the people? Such policies are of coursepossible, butmust depend on the personality and political style of the new presidentand, to a lesserextent, his major antagonists.Before the election no onecan be sure that the new incumbent will make conciliatory moves;certainly the processof political mobilization in a plebiscitariancampaignis not conducive to such a turn of events. The new’ presidentmustconsider whether gestuiresdesigned to conciliate his recent opponentsmight weaken him unduly, especially if he risks provoking his moreextreme allies into abandoning him completely. There is also thepossibility that the opposition could refuse to reciprocate hismagnanimity, thuis causing the whole strategy to backfire. The publicrejectionof anolive branchpublicly profferedcould hardenpositions onboth sides and lead to more. rather than less. antagonism andpolarization.

Someof presidentialism’smost notableeffectson the style of politicsresult from the characteristicsof the presidential office itself. Among

these characteristicsare not only the great powers associatedwith thepresidencybut also the limits imposedon it—particularly thoserequuringcooperationwith the legislative branch, a requirement that becomesespecially salient when that branch is dominated by opponentsof thepresident’sparty. Above all, however,thereare the timo constraintsthata fixed term or number of possible terms imposes on the incumbent.The office of president is by nature two-dimensional and. in a sense,ambiguous: on the one hand,the president is the headof state and therepresentativeof the entire nation; on the other hand, he stands for aclearly partisanpolitical option. If he stands at the head of a multipartycoalition, he may even representan option within an option as he dealswith other membersof the winning electoral alliance.

The presidentmay find it difficult to combinehis role as the headofwhat Bagehatcalled the “deferential’ or symbolic aspectof the polity (arole that Bagehot thought the British monarch played perfectly’ andwhich, in republicanparliamentaryconstitutions,has been successfullyfilled by presidentssuch as SandraPertini of Italy and Theodor Ileussof West Germany) with his role as an effective chief executive andpartisan leaderfighting to promote his party and its program. It is notalways easyto be simultaneously’ the president,say,of all Chileans andof the workers; it is hard to be both the elegant and couirtly master ofLa Moneda(theChileanpresident’sofficial residence)and the demagogicorator of the mass rallies at the soccer stadium. Many’ voters and keyelitesare likely to think that playing the secondrole meansbetraying thefirst—for should not the president as head of state stand at leastsomewhatabove party in order to be a symbol of the nation and thestability of its government?A presidential system, as opposed to aconstitutionalmonarchyor a republic with bath a premierand a headofstate, does not allow such a neat differentiation of roles.

Perhapsthemostimportantconsequencesof thedirect relationshipthatexists betweena presidentand the electorateare the sensethe president

may have of being the only’ elected representativeof the whole peopleand the accompanyingrisk that he will tend to conflate his supporterswith ‘the people” as a whole. The plebiscitariancomponentimplicit inthe president’s authority’ is likely to make the obstaclesand oppositionhe encountersseemparticularly annoying. In his frustration he may betemptedto definehis policies as reflectionsof thepopular will and thoseof his opponents as the selfish designs of narrosv interests. Thisidentification of leader with peoplefosters a certainpopulism that maybe a sourceof strength. It may’ also, however, bring on a refuisal toacknowledge the limits of the mandate that even a majority to saynothing of a mereplurality—can claim as democraticjustification for theenactmentof its agenda. The doleful potential for displays of coldindifference,disrespect,oreven downrighthostility toward the oppositionis not to be scanted.

62J,,um,,io/ of De,,?o(-,-a(V Jima,, .1 L/,i: 63

Unlike the ratherOlympian president,the prime minister is normallya memberof parliament who, even as he sits on the governmentbench,remains part of the larger body. He muist at some point meethis fellowlegislators upon terins of rough equality, as the British prime ministerIegularly does during the traditional question time in the House ofCommons.If he headsa coalition or minority governmentor if his partycommandsonly a slim majority of seats,then he can afford preciouislittle in the uvay of detachmentfrom parliamentaryopinion, A president,by contrast, headsan independentbranchof governmentand meetswithmembersof the legislature on his own terms. Especially uncertain inpresidentialregimesis the placeof Oppositionleaders,who may not evenhold public office and in any case have nothing like the quasi-officialstatus that the leadersof the oppositionenjoy in Britain, for example.

The absencein presidential regimes of a monarchor a “presidentofthe republic” who can act symbolically as a moderatingpowerdeprivesthe system of flexibility and of a means of restraining power. Agenerally neutral figure can provide moral ballast in a crisis or act as amoderatorbetweenthe premierand his opponents—whomay include notonly his parliamentary’foes but military leadersas well. A parliamentaryregime hasa speakeror presidingmemberof parliament who can exertsome restraining influence over the parliamentaryantagonists,includingthe prime minister himself, who is after all a member of the chamberover which the speakerpresides.

The Problem of Dual Legitimacy

Given his unavoidable institutional situation, a presidentbids fair tobecomethe focus for whateverexaggeratedexpectationshis supportersmay harbor. They are prone to think that he has more power than hereally has or should have and may sometimes be politically mobilizedagainst any adversarieswho bar his way. The interaction between apopularpresidentand thecrowd acclaiminghim can generatefearamonghis opponentsand a tensepolitical climate. Somethingsimilar might besaid about a presidentwith a military backgroundor close militaryties—wl~icl~are facilitated by the absence of the prominent defenseminister one usually finds undercabinet government.

Ministers in parliamentarysystemsare situiated quite differently fromcabinet officers in presidentialregimes. Especially in casesof coalitionor minority go\ernments,prime ministers are muich closer to being onan equal footing with their fellow ministers than presidentswill ever bewith their cabinet appointees.(One must note, however, that there arecertain ti-ends which may lead to institutions like that ofKanzlerdemok,-atjein Germany, under which the premier is free tochoose his cabinet without parliamentary approval of the individualmimsters.Parliamentarysystems with tightly disciplined parties and a

prime minister who enjoys an absoluitemajority of legislative seatswilltend to grow quite similar to presidential regimes. The tendency topersonalizepower in modern politics, thanks especiallyto the influenceof television, has attenuatednot only the independenceof ministers buitthe degree of collegiality and collective responsibility in cabinetgovernmentsas svell.)

A presidentialcabinet is less likely than its parliamentarycounterpartto contain strong and independent-mindedmembers. The officers of apresident’scabinethold their postspurely’ at thesufferanceof their chief;if dismissed,they are out of public life altogether.A premier’sministers,by contrast, are not his creatures but nonnally his parliamentarycolleaguies; they may go from the cabinet back to their seats inparliamentand question the prime minister in party’ caucusesor duringthe ordinary course of parliamentary’ business just as freely as othermemberscan. A president,moreover,can shield his cabinet membersfrom criticism much more effectively than can a prime minister, whosecabinetmembersareregularlyhauledbeforeparliamentto answerqueriesor even, in extremecases,to face censure.

One need not delve into all the complexitiesof the relations betsveenthe executiveand the legislature in various presidentialregimes to seethat all such systems are based on dual democratic legitimacy: nodemocraticprinciple exists to resolvedisputesbetweenthe execuitiveandthe legislatureabout which of the two actually representsthe will of thepeople.In practice,particularly in thosedevelopingcatintrieswherethereare great regional inequalities in modernization, it is likely that thepolitical and social outlook of the legislature will differ from that heldby the president and his supporters. The territorial principle ofrepresentation, often reinforced by malapportionment or federalinstitutions like a nonproportional upper legislative chamber, tends togive greater legislative weight to small tosvns and rural areas.Circumstanceslike these can give the presidentgroundsto questionthedemocraticcredentialsof his legislativeopponents.He may even chargethat they representnothing but local oligarchies and narrow, selfishclienteles. This may or may not be true, and it may or may not besvorse to cast one’s ballot under the tutelage of local notables, tribalchieftains, landowners,priests,or even bossesthan under that of tradeunions,neighborhoodassociations,or party machines.Whatever the casemay be, modernurban elites will remain inclined to skepticismabatitthedemocraticbonafides of legislatorsfrom rural or provincial districts. Insucha context, a presidentfrustrated by legislativerecalcitrancewill betemptedto mobilize the peopleagainsttheptitative oligarchsand specialinterests, to claim for himself alone truie democratic legitimacy as thetribtine of the people, and to urge on his supporters in massdemonstrationsagainstthe apposition.It is alsoconceivablethat in somecountries the presidentmight representthe more traditional or provincial

04 Iou,oio/ of De,,ioc,-a: limo,, I L(,c

electorates and could use their support against the mare urban andmodern sectorsof society,

Even more ominously, in the absenceof any principled method ofdistinguishingthe true bearerof democraticlegitimacy,the presidentmayuseideologicalformulationsto discredithis foes; institutional rivalry may’thus assumethe character of potentially explosive social and politicalstrife, Institutional tensions that in some societies can be peacefullysettledthroughnegotiationor legal meansmayin other, legs happy landsseektheir resolution in the streets.

The Issue of Stability

Among the oft-cited advantagesof presidentialismis its provision forthe stability of the executive,This feature is said to furnish a welcomecontrast to the tenuousnessof many parliamentarygovernments,withtheir frequentcabinet crisesand changesof prime minister, especiallyin

Ihe muiltiparty democraciesof WesternEurope. Certainly thespectacleofpolitical instability presentedby’ the Third and Fourth FrenchRepublicsand, more recently, by Italy and Portugal has contributed to the lowesteem in which many scholars—especiallyin Latin America—hold

parliamentarism and their consequient preference for presidentialgovernment.But suich invidious comparisonsoverlook the largedegreeof stability that actually characterizesparliamentary governments.Thesuperficial volatility they sometimesexhibit obscuresthe continuity ofparties in power, the enduring characterof coalitions, and the way thatparty leadersand key ministershave of sveatheringcabinetcriseswithoutrelinquishing their posts. In addition, the instability of presidentialcabinetshasbeen ignored by studentsof governmentalstability. It is alsoinsufficiently noted that parliamentary systems,precisely by virtue oftheir surfaceinstability, often avoid deepercrises. A prime minister whobecomesembroiled in scandalor loses the allegiance of his party ormajority coalition and whosecontinuancein office might provoke graveturmoil can be much more easily removed than a corrupt or highlyuinpopularpresident.Unless partisan alignments make the formation of

a democratically legitimate cabinet impossible, parliament shouldeventuallybe able to selecta new prime minister who can form a newgovernment.In some more seriouscases,new elections may be called,although they often do not resolve the problem and can even, as in thecaseof Weimar Germany in the l930s, compoundit.

The government crises and ministerial changes of parliamentaryregimes are of courseexcludedby the fixed term a presidentenjoys.butthis great stability is bought at the price of similarly great rigidity.Flexibility in the face of constantly changing situations is notpresidentiatism’sstrong suit. Replacing a presidentwho has lost theconfidence of his party or the people is an extremely difficult

proposition. Even when polarization has intensified to the point ofviolence and illegality, a stuibborn incumbentmay remain in office. Bythe time the cumbersomemechanismsprovidedto dislodgehim in fas orof a more able and conciliatory successorhave done their work, it maybe too late. Impeachment is a very uncertain and time-consuimingprocess,especiallycomparedsvith the simple parliamentai’y vote of noconfidence.An embattledpresidentcan use his powers in such a waythat his opponentsmight not be willing to wait until the end of his termto oust him, but thereare no constitutionalways—saveimpeachmentoiresignationtinder pressure—tareplacehim. There are, moreover, risksattachedeven to theseentirely legal methods;the incumbent’ssupportersmay feel cheatedby them and rally behindhim. thus exacerbatingthe

crisis. It is hard to imaginehosv the issue could be resolvedpurely by’the political leaders.with no recourseor threat of recourseto the peopleor to nondemocratic institutions like the courts or—in the worstcase—themilitary. The intenseantagonismsunderlyingsuchcrisescannotremain even partially concealedin the corridors and cloakroomsof thelegislature.What in a parliamentarysystemwould be a governmentcrisiscan becomea full-blown regime crisis in a presidentialsystem.

The same rigidity is apparentohen an incuimbent dies or suffersincapacitationwhile in office. In the latter case,there is a temptationtoconceal the president’s infii’mity until the end of his term. In event ofthe president’s death, resignation, impeachment, or incapacity, thepresidentialconstitution very often assuresan automaticand iminediatesuccessionwith no interregnumor powervacuuim, But the instituitian ofvice-presidentialsuccession.svhich has worked so well in the UnitedStates,may not function so smoothly elsewhere.Particularly at risk arecountries whoseconstitutions,like the United StatesConstitution beforethe passageof the Twelfth Amendmentin 1804, allow presidentialticketsto be split so that the winning presidentialcandidateand the winningvice-presidential candidate may come from different parties. If thedeceasedor outgoingpresidentandhis legal successorare from differentparties. thosewho supportedthe former incumbentmight object that thesuccessordoesnot representtheir choiceandlacks democraticlegitimacy.

Today, of course,fesv constitutions would allosv something like theUnited States’ Jefferson-Burrelection of 1800 to occur. Instead theyrequire that presidentialand sice-presidentialcandidatesbe nominatedtogether, and forbid ticket-splitting in presidential balloting. But theseformal measurescan do nothing to control the criteria far nomination,Thereareundoubtedly’casesw’here thevice-presidenthasbeennominatedmainly to balancethe ticket and thereforerepresentsa discontinuity’ withthe president. Instances where a weak vice~presidentialcandidate isdeliberatelypicked by an incumbentjealousof his awn power, or evenwherethe incumbentchooseshis own wife, are not unknown. Nothingabout the presidential system guaranteesthat the country’s voters or

on Jon, ,ia/ of Denioci’acv mom? I Liii:

political leaders would have selected the vice-presidentto wield thepowersthey were willing to give to the formerpresident.Thecontinuitythat the institution of automatic vice-presidential successionseemstoensure thus might prove more apparentthan real. There remains theobviouis possibility’ of a caretakergovernmentthat can fill in until newelectionstake place, preferablyas soon as possible. Yet it hardly seemslikely that the severecrisis which might have requnredthe successionssouild also provide an auspiciousmomentfor anew presidentialelection,

The Time Factor

Democracyis by definition a governmentpro tempore,a regime insvhich the electorate at reguilar intersals can hold its governorsaccountableand impose a change.The limited time that is allowed toelapse between elections is probably the greatest guarantee againstoverweening power and the last hope for those in the minority. Itsdrasvback,however, is that it constrainsa government’sability to makegood on the promises it madein order to get elected.If thesepromiseswere far-reaching, including major programs of social change, themajority may feel cheatedof their realization by the limited term inoffice imposedon their chosenleader. On the other hand, the power ofa presidentis at once so concentratedand so extensivethat it seemsunsafenot to checkit by limiting the numberof timesany onepresidentcan be reelected.Such provisions can be frustrating, especially if theincuimbentis highly ambitious; attemptsto changethe rule in the nameof continuity have often appearedattractive,

Even if a presidententertainsno inordinate ambitions,his awarenessof the time limits facinghim andthe programto w’hich his nameis tiedcannot help buit affect his political style. Anxiety about policydiscontinuitiesand the characterof possiblesuiccessorsencourageswhatAlbert Hirschman has called “the wish of vouloir conclure.” Thisexaggeratedsenseof urgencyon the part of the presidentmay lead toill-conceived policy initiatives, overly hasty stabs at implementation,unwarrantedangerat the law’ful opposition, anda host of otherevils. Apresidentwho is desperateto build his Brasiliaor implementhis programof nationalization or land reform before he becomes ineligible farreelection is likely to spend money unvvisely or risk polarizing thecountry far the sake of seeing his agendabecome reality. A primeminister svho can expecthis party or governingcoalition to win the nextrouind of electionsis relatively free from suchpressuires.Prime ministershave stayed in office over the courseof several legislatureswithoutrouising any fears of nascentdictatorship,for the possibility of changingthe government without recourse to unconstitutional means alwaysremainedopen.

The fixed term in office andthe limit on reelectionareinstitutions of

unquestionablevalue in presidentialconstitutions.buit they meanthat thepolitical systemmust producea capable andpopuilar leaderevery fouiryears or so, and also that w’hatever “political capital” the outgoingpresidentmay have accumulatedcannotendurebeyond the end of histerm.

All political leaders must worry aboutthe ambitions of second-rankleaders,sometimesbecauseof their jockeying for position in the orderof successionand sometimesbecauiseof their intrigues. The fixed anddefinite date of successionthat a presidentialconstitutionsetscan onlyexacerbatethe incumbent’sconcernson this score.Add to this the desirefor continuity, and it requires no leap of logic to predict that thepresidentwill chooseas his lieutenantand successor-apparentsomeonewho is more likely’ to prove a yes-manthan a leader in his own right.

The inevitable successionalso createsa distinctive kind of tensionbetweenthe ex-presidentand his successor.The new man may feeldriven to assert his independenceand distinguish himself from hispredecessor,eventhough both might belong to the same party. The oldpresident, for his part. having known the unique honor and senseofpower that come with the office, will always find it hard to reconcilehimself to beingout of pow-er for good, with no prospect of returningeven if the new incumbentfails miserably. Partiesand coalitions maypublicly split becauseof such antagonismsand frustrations. They canalso leadto intrigues,as when a still-prominent former presidentworksbehind the scenesto influence the next suiccessionor to uindercut theincumbent’spoliciesor leadershipof the party.

Of coursesimilar problemscan also emergein parliamentarysystemswhen a prominentleaderfinds himself out of office but eagerto return.But parliamentaryregimescanmore easilymitigate such difficuilties fora number of reasons. The acute need to preserve party unity, thedeferenceaccordedprominentparty figures, andthe new premier’skeenawarenessthat he needsthe help of his predecessoreven if the latterdoesnot sit anthe governmentbenchor the samesideof the house—allthese contribute to the maintenanceof concord. Leaders of the samepartymay alternateas premiers;eachknows that the othermay’ be calledupon to replacehim at any time andthat confrontationscan be costly toboth, so they sharepower. A similar logic applies to relationsbetweenleadersof competingpartiesor parliamentarycoalitions.

The time constraintsassociatedwith presidentialism,combinedwiththe zero sum characterof presidentialelections,are likely to rendersuchcontestsmore dramatic and divisive than parliamentaryelections, Thepolitical realignmentsthat in a parliamentary’system may take placebetweenelections and within the halls of the legislaturemust occuirpublicly during election campaignsin presidentialsystems.where theyare a necessarypart of the processof building a vvinning coalition.Under presidentialismtime becomesan intensely important dimension

Jon, ,ia/ of’ De,noc,-ac-m Jua,i 1. Li,i: 6i)

of politics. The paceof politics is very different undera presidential,asopposedto a parliamentary,constitution When presidentialballoting isat hand, deals must be madenot only publicly but decisively_for thesvinning side to i’enege on them before the next campaignwould seemlike a betrayalof thevoters’ trust. Compromises,howevernecessary,thatmight appearunprincipled, opportunistic, or ideologically unsoundaremuch harderto make when they are to be scrutinizedby the voters inan upcomingelection. A presidentialregime leaves much less room fortacit consensusbunl~~~gcoalition-shifting and the making ofcompromiseswhich, though prudent,arehard to defendin public.

Consociational methods of compromise, negotiation, and power-sharing underpresidentialconstitutionshave played major roles in thereturn of democraticgovernment to Colombia, Venezuela,and, morerecently, Brazil, But these methods appeared as necessary’antinomies_deviationsfrom the rules of the systemundertakenin orderto limit the voters’ choicesto what hasbeentermed, rather loosely andpejoratively, democi’adui’a The restorationof democracywill no doubtcontinueto require consociationalstrategiessuch as the formation ofgrand coalitions and the making of many’ pacts; the drawback ofpresidentialisinis that it rigidifies and formalizes them. They becomebinding for a fixed period, during which there is scant opportunity forrevision or renegotiation.Moreover,as the Colombiancase shows,suchaiTangementsrob the electorate of same of its freedom of choice;parliamentarysystems, like that of Spain with its consenso,make itmuchmorelikely that consociationalagreementswill be madeonly afterthe peoplehave spoken.

Parliamentarism and Political Stability

This analysis of presidentialism’s unpromising implications fordemocracyis not meantto imply that no presidentialdemocracycan bestable; on the contrary, the world’s most stabledemocracy_theUnitedStates of America—hasa presidential constitution Nevertheless,onecannothelp tentativelyconcludingthat in many othersocietiesthe oddsthat presidentialismwill help preservedemocracyare far less favorable.

While it is true that parliamentarismprovides a more flexible andadaptableinstitutional context for the establishmentandconsolidationofdemocracy it doesnot follow that just any sort of parliamentaryregimewill do, Indeed, to completethe analysisonewould needto reflect uponthe best type of parliamentaryconstitution and its specific institutionalfeatures.Among these would be a prime-ministerialoffice combiningpower with responsibility, svhich would in turn require strong, well-disciplined political parties. Such features—thereare of course manyothers we lack the spaceto discuss—wouldhelp foster responsibledecisionmaking and stablegovernmentsand would encouragegenuine

party competition without causing undue political fragmentation. Inaddition, every country has unique aspects that one must take intoaccount—traditionsof federalism,ethnicor cultural heterogeneity.and soan. Finally, it almost goes ovithout saying that our analysisestablishesonly probabilities and tendencies, not determinisms. No one canguaranteethat parliamentarysystems will neverexperiencegrave crisisor evenbreakdown.

In the final analysis, all regimes. howesem wisely designed, mustdependfor their preservationupon the suipportof society at large—itsmajor forces, groups,and institutions.They rely, therefore, on a publicconsensuswhichrecognizesaslegitimate authorityonly’ that pow’er whichis acquiredthrough lawful anddemocraticmeans,They dependalso onthe ability of their leadersto govern, to inspire trust, to respectthe limitsof their paw’er, andto reach an adequatedegreeof consensus.Althoughthesb qualities are most needed in a presidentialsystem, it is preciselythere that they are most difficult to achieve, Heavy reliance on thepersonalqualities of a political leader—onthe virtue of a statesman,ifyou will—is a risky course,for one neverknows if such a man can befoundto fill the presidentialoffice. But while no presidentialconstitutioncan guaranteea Washington,a Juhrez,or a Lincoln. no parliamentary’regime can guaranteean Adenaueror a Churchill either. Given suchuinavoidableuncertainty,the aim of this essay hasbeenmerely to helprecovera debateon the role of alternative democratic institutions inbuilding stabledemocraticpolities.