Endogenizing legislative candidate selection procedures in nascent democracies: evidence from Spain...

This article was downloaded by: [Wake Forest University]On: 05 September 2011, At: 09:06Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH,UK

DemocratizationPublication details, including instructions for authorsand subscription information:http://www.tandfonline.com/loi/fdem20

Endogenizing legislativecandidate selection proceduresin nascent democracies:evidence from Spain and ChileBonnie N. Field a & Peter M. Siavelis ba Department of Global Studies, Waltham, MA, USAb Department of Political Science, Winston-Salem, NC,USA

Available online: 25 May 2011

To cite this article: Bonnie N. Field & Peter M. Siavelis (2011): Endogenizing legislativecandidate selection procedures in nascent democracies: evidence from Spain and Chile,Democratization, 18:3, 797-822

To link to this article: http://dx.doi.org/10.1080/13510347.2011.563120

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions

This article may be used for research, teaching and private study purposes.Any substantial or systematic reproduction, re-distribution, re-selling, loan,sub-licensing, systematic supply or distribution in any form to anyone isexpressly forbidden.

The publisher does not give any warranty express or implied or make anyrepresentation that the contents will be complete or accurate or up todate. The accuracy of any instructions, formulae and drug doses should beindependently verified with primary sources. The publisher shall not be liablefor any loss, actions, claims, proceedings, demand or costs or damages

http://www.tandfonline.com/loi/fdem20

http://dx.doi.org/10.1080/13510347.2011.563120

http://www.tandfonline.com/page/terms-and-conditions

http://www.tandfonline.com/page/terms-and-conditions

whatsoever or howsoever caused arising directly or indirectly in connectionwith or arising out of the use of this material.

Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

Endogenizing legislative candidate selection procedures innascent democracies: evidence from Spain and Chile

Bonnie N. Fielda∗ and Peter M. Siavelisb

aDepartment of Global Studies, Waltham, MA, USA; bDepartment of Political Science,Winston-Salem, NC, USA

(Received 8 March 2010; final version received 22 October 2010)

This article contributes to filling a gap in the resurgent literature on legislativecandidate selection procedures by analysing the adoption of such proceduresin nascent democracies. We contend that within transitional systems distinctcontexts constrain choice and bargaining for candidate selection proceduresin different ways, and condition the adoption of legislative candidateselection procedures by parties. In particular we posit that the relative levelsof uncertainty about the installation and continuance of democracy, strategiccomplexity of the electoral system, and party leadership autonomy, createincentives for the adoption of more or less inclusive candidate selectionprocedures. We evaluate our propositions based on evidence from therelevant political parties in Spain and Chile.

Keywords: candidate selection; political parties; democratization; Spain;Chile

Introduction

In recent years, there has been a resurgence of interest in candidate selection pro-cesses. Candidate selection refers to ‘the predominantly extralegal process bywhich a political party decides which of the persons legally eligible to hold an elec-tive public office will be designated on the ballot and in election communicationsas its recommended and supported candidate or list of candidates’.1 Scholars haverecognized that in all electoral environments, the candidate selection process dra-matically narrows the electorate’s choice, with profound effects on representation.2

Those who select candidates for public office are essentially gatekeepers of democ-racy, and in certain contexts the candidate selection process can largely determinewho will be elected. Therefore, if we are interested in representation, both in newand longstanding democracies, we should study candidate selection dynamics asclosely as elections. This includes understanding how candidates are selected aswell as the determinants and effects of distinct candidate selection processes.

ISSN 1351-0347 print/ISSN 1743-890X online# 2011 Taylor & Francis

DOI: 10.1080/13510347.2011.563120

http://www.informaworld.com

∗Corresponding author. Email: [email protected]

DemocratizationVol. 18, No. 3, June 2011, 797–822

Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

The newer candidate selection (from here on CS) literature provides improvedtypologies and essential empirical information regarding parties’ CS processes.Both the newer and classic literatures on parties elucidate some of the importanteffects of CS procedures.3 However, there are fewer studies of the origins of dis-tinct CS processes, particularly for legislative candidates. The most commonapproach treats selection processes as tangential by-products of other institutionalprocesses,4 and very little of it has comparative explanatory traction, particularly innascent democracies. This article seeks to shed light on why parties in nascentdemocracies adopt the candidate selection procedures that they do.

We contend that candidate selection processes are not simply tangential by-pro-ducts of other institutional arrangements. Rather, within transitional systems, distinctcontexts differently constrain choice and bargaining, and condition party adoption oflegislative CS procedures. In particular we posit that the relative levels of uncertaintyabout the installation and continuance of democracy, strategic complexity of the elec-toral system, and party leadership autonomy, create incentives for the adoption ofmore or less inclusive candidate selection procedures. We evaluate our propositionsbased on evidence from the relevant political parties in Spain and Chile.

We focus on the selection of candidates for national legislative offices. Morespecifically, we analyse the inclusiveness of the selectorate, defined as the bodythat chooses the candidates. We employ Rahat and Hazan’s typology in which theselectorate, on a spectrum from inclusive to exclusive, may be the electorate as awhole, party members, selected party agency, nonselected party agency, selectedparty leader(s), and nonselected party leader(s).5 Therefore, the procedures may runfrom party primaries of various types on the inclusive end, selection by a conventionof intermediate-level party activists in the mid-range, and selection by leadership onthe exclusive end. A serious problem confronted when studying CS is that aspirantsoften face multiple selectorates on their path to candidacy. That is, CS is often a multi-stage process, with varying selectorates and distinct degrees of inclusiveness at differ-ent stages.6 On which or how many should we concentrate? In this study, we ask,‘Who ultimately decides?’ The answer may or may not correspond with formalparty procedures; therefore any characterization of CS procedures must be basedon a party specific understanding of the de facto and informal processes.

This article continues, first, with a critical discussion of the explanations com-monly advanced to explain the use of distinct CS procedures. Secondly, it outlinesthe methodological approach used in our study. Thirdly, it develops the key pro-posed variables, namely regime uncertainty, strategic complexity of the electionsystem, and party leadership autonomy in more depth for the parties in Spainand Chile. Finally, we review our findings and elaborate why it is important tobetter understand candidate selection in democratizing countries.

Explaining CS procedures: casting doubt on alternative explanations

In the limited literature that mentions the origins of CS procedures, arguments takea few general forms, none of which we find fully convincing. We outline our

798 B.N. Field and P.M. Siavelis

Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

critique through an initial examination of two countries which we will later analysein much greater depth: Chile and Spain.

Perhaps the most common explanation of the varying degrees of inclusivenessof the selectorate argues that the incentives associated with different electionsystems drive parties to employ different CS procedures.7 The conventionalwisdom is that the increased importance of personality and candidate character-istics in single-member or low-magnitude systems (with magnitude defined asthe number of seats per district) provides incentives to adopt primaries ratherthan more exclusive procedures.8 On the other hand proportional representation(PR) systems, multimember districts and electoral systems with higher districtmagnitudes should tend towards exclusive selection, especially in closed listsystems, given that a candidate’s personal characteristics are less important andpolicy and programme should dominate campaigns.9

A cursory comparison of Chile and Spain shows, however, that the electoralsystem format alone is insufficient to account for different selection procedures.Chile uses a two-member district system to elect deputies across 60 districts,which tends toward majoritarianism. Given this small magnitude and the use ofopen lists, CS processes that highlight individual candidate differences like pri-maries should be the norm. In comparison, Spain’s Congress of Deputies iselected using a closed-list PR system with an average district magnitude of 6.7.Extant theory would predict that this type of election system is likely to tendtowards less inclusive CS procedures. Nonetheless, exclusive and party dominatedprocedures have been the norm in both countries.

Similarly, parliamentary systems are said to encourage party and leadershipselection of legislative candidates given the centrality of generating the party dis-cipline necessary to sustain parliamentary government, while presidential systemsare more apt to rely on more inclusive procedures because discipline is less centralto the maintenance of governments.10 Once again, presidential Chile’s and parlia-mentary Spain’s similarly exclusive CS procedures fail to provide evidence thatregime structure alone can account for the emergence of different selectionprocedures.

Neither do arguments based on the geographic dispersion of power hold muchtraction. Multilevel systems with important regional governments have been tied tomore inclusive selection procedures because parties are also likely to follow geo-graphic decentralization to mirror state institutions, arguably allowing for moreinclusive procedures at the sub-national level. Multilevel systems have also beentied to decentralized selection procedures because local party branches will haveincentives to offer candidates with appeal to local electorates.11 Once again,Chile’s extremely unitary state and Spain’s decentralized state vary completelyon these counts. Indeed, in both systems we find local input into the CSprocess.12 Yet, ultimately CS remains a very exclusive process with central partyleaders making the final decision in both countries.

Additional explanations of different modes of selection are based on aspectsof internal party organization.13 Parties with large membership bases are

Democratization 799

Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

hypothesized to tend toward more inclusion. Again, a rudimentary review of ourcases disaffirms a strong connection between these variables and differentmodes of selection. At the party level, party organizations during the dictatorshipsand transitions varied substantially. Chile’s opposition parties faced brutal repres-sion, yet with the return of democracy emerged with the same strength, levels ofsocial penetration, and competitive physiognomy that characterized them beforethe dictatorship. In contrast, the Spanish parties all but disappeared, with thepartial exception of the Partido Comunista de Espana (PCE), over the near 40years of authoritarian rule. Therefore, parties with both large and small member-ship bases relied on exclusive modes of selection.14

Here we have discounted some of the most common and potentially powerfulexplanations for the emergence of different kinds of CS procedures. In the next sec-tions, we argue that the interaction of regime uncertainty, strategic complexity ofthe election system, and party leadership autonomy help explain degrees of inclu-siveness of the selectorate. First, however, we provide a discussion of the method-ology we employ.

Qualitative case studies and most different systems design

Despite pointed methodological debates in political science during the last twodecades, widespread agreement has emerged that the choice of methods mustmatch the nature and goals of inquiry. The remedial state of the CS literature isa reflection of the inherent difficulties of finding the appropriate method to studya phenomenon so informal and intentionally veiled that Gallagher and Marshcall it the ‘secret garden of politics’.15 What, then, is the best method to jumpstart theory development for such a process? We propose that research must necess-arily be incremental and that concrete methods must be aimed at wading throughthe informality and incentives that political elites have to veil this critical process.

Therefore, this study employs qualitative case study methodology and wedefine a case as a relevant political party that affects party tactics or competitionat the national level by having ‘coalition’ or ‘blackmail’ potential.16 Qualitativecomparative case studies are most appropriate for the initial stages of researchon the origins of CS methods for many reasons. First, theory on the determinantsof CS procedures is not advanced enough to employ solely a ‘testing’ logic.17 Weneed a method that can best advance theory development and rigorous incrementaltesting. Case studies are particularly good for hypothesis generation, theory devel-opment, and identifying missing variables,18 especially those that are unexpectedor counterintuitive. Secondly, variables may interact in distinct ways to produce thesame outcome, and qualitative case research is better suited to grasping the distinc-tions, which might be missed in statistical analyses.19 Thirdly, case studies permiton-going clarification of key concepts20 and better achieve measurement validity.We pay particular attention to operationalizing the concepts appropriately for dis-tinct environments, the importance of which would be missed in statistical analysesthat employ common indicators.21


Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

Research for this article is based on the CS procedures employed by the rel-evant political parties in Chile and Spain. As such, while stylistically this articleis organized by country, the unit of analysis is the political party. This nonrandomcase selection approximates the ‘most different system design’22 and Mill’s‘method of agreement’.23 We identify a common outcome to be explained andseek to identify common traits that can account for it. We also rely on whatMunck calls selecting ‘sharply contrasting cases’, which ‘may permit strongertests of hypotheses through focus on diverse contexts. High variability specificallyon rival explanations may yield more leverage in test of theory’. 24 In accordancewith this approach, this study maximizes variation particularly on rival institution-ally-based factors. As seen above, we selected two countries, Chile and Spain,which dramatically differ with regard to constitutional structure, electoral laws,and the territorial organization of the state. The claim here is that since these exist-ing potential explanatory variables vary across cases, they are not sufficient toexplain the outcome. Moreover, having similarly exclusive selectorates helps ushone in on the importance of alternative explanatory variables. Though the speci-fics vary by party, the parties in both countries faced uncertainty about the stabilityof the nascent democracy, strategic complexity of the electoral system and exhib-ited party leadership autonomy – our proposed explanatory factors.

Nevertheless, the authors are aware of the potential limitations of this design,25

and their causal inferences are appropriately tempered. These limitations includebut are not limited to: (1) nonrandom case selection is subject to ‘cherry picking’whereby the investigators chose cases that fit their favoured theoretical variables;(2) homogeneity of variables (for example, presence of x and y) does not aloneestablish a causal link between the variables; (3) unidentified independent variablescould account for the outcome, y; (4) the lack of variation of the dependent variablecan lead to serious problems of causal inference; and (5) the possibility that distinctand multiple combinations of causes produce the outcome.

Here we simply make the case for the plausibility of the causal variables ident-ified. Given the nature of CS, only through an incremental approach can we begin togenerate valid explanations of why parties adopt different CS procedures. Ourevaluation is based on examining media reports, secondary sources, party docu-ments and numerous interviews with party officials, in the spirit of process-tracing analysis: ‘The process-tracing method attempts to identify the interveningcausal process – the causal chain and causal mechanisms – between an independentvariable (or variables) and the outcome of the dependent variable’.26 The authors arealso cognizant of the potential problems associated with interview-based researchmethods; therefore using multiple sources and contrasting interviewee statementsalso allow us to assess the reliability of direct personal testimonies.

Candidate selection procedures in post-authoritarian Spain and Chile

In general terms, we are interested in the CS procedures that emerge just prior tothe founding elections, defined as elections that are likely to occur under

Democratization 801

Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

reasonably fair conditions,27 and the consolidation of the democratic regime –when democracy has become ‘the only game in town’.28 In Spain, this corre-sponds to the period from 1977 to 1982. Negotiations between the incumbentauthoritarian government and the opposition produced a democratically electedgovernment under reasonably free and fair conditions in June 1977, approxi-mately 19 months after the death of dictator Francisco Franco in November1975. The elections were won by a new centre-right electoral alliance – theUnion de Centro Democratico (UCD) – organized around the last reformist pre-sident of the authoritarian regime, Adolfo Suarez, who subsequently led a UCDminority government as prime minister. The Partido Socialista Obrero Espanol(PSOE) became the largest opposition party, followed by the PCE and theright-wing Alianza Popular (AP). Other relevant parties were regional alliancesand parties that attained significant representation in the national parliament:Pacte Democratic per Catalunya (PDC) and later Convergencia i Unio (CiU),and the Partido Nacionalista Vasco (PNV). Between 1977 and 1982, the newdemocracy was challenged by unreformed and potentially hostile military andsecurity forces, exacerbated by ETA (Basque independence organization) andright-wing terrorism. However, by 1982, Spain had negotiated a new, consensualconstitution (1978), defeated a coup attempt (1981) and experienced an alterna-tion of power to the PSOE (1982) – a centre-left party unassociated with theauthoritarian past. Therefore, Spanish democracy is considered consolidatedafter 1982.29

In turn, Chile’s lengthy democratic transition began with the 1980 plebis-cite. Termed an opening by the military government, the plebiscite representedan attempt to provide a purportedly ‘legal’ formula to prop up a fatigued regimeand an institutional framework to keep it in power. Despite the plebiscite’squestionable democratic propriety, voters simultaneously approved a newconstitution and a timeline for transition that kept former dictator AugustoPinochet as president until another plebiscite would be held on his continuedrule in 1988, followed by presidential elections in 1989. Two multiparty alli-ances contested these and subsequent elections. The centre-left Concertacionalliance comprised the Partido Democrata Cristiano (PDC), the Partido Socia-lista (PS), the Partido por la Democracia (PPD) and the smaller PartidoRadical (PR) and Partido Social Democrata (PSD). The latter two partiesmerged in 1994 to form the Partido Radical Social Democrata (PRSD). Onthe right, the Alianza alliance was composed of two major parties, RenovacionNacional (RN) and the Union Democrata Independiente (UDI), with the excep-tion of the 1993 election when the Union de Centro Centro (UCC) also joinedthe Alianza. The Chilean transition ended with the 2005 constitutional reforms,which eliminated most of the constitution’s non-democratic elements. Afterthese reforms, Chile finally met the basic standards of democracy set outby Linz and Stepan, because ‘the executive, legislative and judicial powergenerated by the new democracy’ no longer had to ‘share power with otherbodies de jure’.30


Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

Spain

All Spanish parties employed CS procedures at the exclusive end of the spectrum.Focusing here on the two largest parties, the UCD and PSOE, CS in the founding1977 elections was largely a leadership affair, with the parties struggling to findcandidates in some districts. De jure, by 1979, sub-national party-leadershipbodies proposed candidate lists that could be subsequently altered by the nationalparty-leadership bodies in Madrid. There were no party conventions to choosecandidates and no membership ballots; yet in this process rank-and-file membersof sub-national party branches could voice their preferences on candidates totheir sub-national party organizations. The PSOE’s CS procedures remained inplace until 1998, and, in essence, the UCD’s procedures remained in effect untilthe party disbanded in 1983. De facto, CS in all parties was largely controlledby a few party leaders or leadership bodies.31 National party organizations didnot just have formal veto power, but could (and did) replace proposed candidateswith candidates of their own choosing, including ‘parachuting’ politicians intodistricts with which they had no ties.

The UCD’s procedures were less inclusive than those used by the PSOE. Fewernational level actors were included, and where sub-national party actors wereinvolved in CS they were more likely to be small groups of local notables thanregional-party organizations representing a base. For the 1977 UCD alliancelists, Suarez selected Leopoldo Calvo-Sotelo to handle negotiations with the con-stituent parties’ leaders; Calvo-Sotelo in turn required party leaders to sign a docu-ment granting him the power of final arbitration of candidate lists, which theydid.32 In effect, a few unelected alliance leaders ultimately decided the compositionof candidate lists. After the UCD became a formal party with an elected executive,the 1978 party statutes gave the party president formal control of ‘electoralmatters’, which according to Hopkin allowed Suarez to control the lists.33 Accord-ing to our interviews, the process was largely controlled by the party’s executivecommittee, and Suarez’s voice within the executive as party president and asprime minister was decisive. Furthermore, the influence of the original constituentparty barons34 and some of Suarez’s own allies was reduced. A similar processgoverned the 1982 elections: the party’s executive committee dominated theprocess, but, with internal divisions leading to the departure of important segmentsof the party, including Suarez himself, and declining electoral prospects, there werefewer aspirants clamouring for candidacies. Nonetheless CS in 1979 and 1982 wasultimately carried out by a selected party agency, with a few key actors playingpowerful roles.

Where the UCD was able to ignore a largely nonexistent base, the PSOE lea-dership could not completely ignore sub-national party organizations. In the firstelections, the elected PSOE executive played the determinative role in CS; yet,the process was modified in 1979. Each provincial branch presented a list ofnames twice as long as the number of seats, with final selection to be approvedand possibly modified by the national party organization.35 However, our

Democratization 803

Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

interviews clearly indicate that the national party leadership attempted to influencethese proposals. There was a party List Committee, as there was in the UCD, whichreviewed these proposals. Significantly, party executive members had a majorityon the committee.36 In contrast to the UCD changes at the national level had tobe approved by the PSOE’s Federal Committee. The Federal Committee is abroader, more representative body, composed of members of regional party federa-tions and of the Executive Committee. Therefore, the PSOE moved from selectionby selected party leaders to selection by a selected party agency.

In interviews, party officials consistently cited one or two individuals who ulti-mately controlled CS: for the PSOE, Alfonso Guerra; and for the UCD, AdolfoSuarez, Leopoldo Calvo-Sotelo, Rafael Arias Salgado, or Landelino Lavilla,depending on the election. Yet the key actors quickly claimed that their ‘power’was exaggerated as many voices were heard, and that while they were statutorilypermitted to alter candidacies, this was difficult in practice. Therefore, that leadersultimately decided does not mean that other party organs or individuals were notconsulted; or that proposals were not influential. PSOE party leaders all mentionedthe role of provincial and later regional-party bodies; UCD party leaders referencedfaction leaders and key regional figures. Ultimately, however, CS was formally in thehands of selected party agencies, though a broader one in the case of the PSOE;informally the process was greatly influenced by a few key figures.

Chile

Focusing on CS in the major parties of the Concertacion and Alianza, in Chile, too,the CS procedures were at the exclusive end of the spectrum. In the first post-authoritarian elections of 1989 every party relied on either party elites or a bodycomposed of party elites to choose its candidates, and then entered into alliancenegotiations to determine final lists, often trading away slates at the alliancelevel. The process thus conformed to Rahat and Hazan’s category of nonselectedparty agency at both the party and the alliance level.

Subsequent elections saw more of a multilevel game that, while slightly moreinclusive, still clearly tended toward the exclusive end of the scale. Party leadersusually consulted with sub-national party bodies for initial candidate proposals,though local actors often simply ratified recommendations coming from Santiago,and incumbents were understood to have a right of re-nomination. Both in theAlianza and Concertacion, each party had its own process to designate initialslates of candidates by party executive committees (with few exceptions discussedbelow) before alliance negotiations, representing a move from nonselected partyagency selection toward the selected party agency designation of candidates.

While observers have pointed to the use of primaries in the Concertacion in sub-sequent elections, primaries rarely happened and when they did party leaders oftenoverrode the results. The PDC and PPD held a few primaries relying only on partymilitants for the 1993 and 1997 elections. In the very limited cases in which theywere used for the 2001 elections, the PPD continued to rely on closed primaries


Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

and the PDC was the only party to use open primaries.37 However, primaries werefew and far between. One principal involved in the process noted that the use of‘internal primaries is really the exception. I can only remember three or fourcases’.38 They were usually held only where an incumbent was not running orwhere there were challenges to a ‘natural’ candidate.39 More importantly in allcases parties accepted that primaries could be overruled during alliance-level nego-tiations. On the centre-left every party leader interviewed acknowledged willing-ness to overrule primary results to satisfy the interest of the alliance, or, as notedby one interviewee, to satisfy the preferences of the party leadership hiddenbehind the guise of alliance interests. In this sense, it would be a mistake to charac-terize CS in Concertacion as inclusive simply because a few primaries were held. Inthe end, the combination of automatic incumbent re-nomination and veto points atboth the party and alliance level mean that CS fits best within the categories ofselected party agency and at times nonselected alliance negotiators, because theseagencies still wielded ultimate control.

The calculus on the right for overturning the decisions of its constituent partieswas distinct, but the outcome in terms of ‘who decides’ remained the same. Withonly two parties at the negotiating table (with the exception of the 1993 electionswhen the UCC participated, and indeed complicated the selection process), the RNand the UDI simply decided to divide candidacies, precluding the necessity oftrading away party designated candidacies for concessions elsewhere or to accom-modate small parties. However, all significant actors on the right who were inter-viewed unapologetically noted that elites named candidates. Thus, CS in theAlianza too is best characterized as controlled by selected party agencies. Unlikethe parties of the Concertacion, neither the UDI nor RN relied on significantpopular input, nor was there even pretence that primaries might be desirable.Interestingly, then, in both Spain and Chile the CS procedures during the initialpost-authoritarian elections can be categorized at the exclusive end of thespectrum. The following sections will explore the three explanatory factors thathave been discussed above, regime uncertainty, the strategic complexity of theelectoral system, and party leadership autonomy.

Regime uncertainty

Regime uncertainty refers to the degree to which significant political actors believethat the installation and continuance of democracy are endangered. The propositionis that moments of political uncertainty provide an external constraint on the choiceof CS procedures.40 The logic underlying this proposition is that high uncertainty islikely to encourage a more exclusive selectorate because it provides party leaderswith an incentive and a credible argument vis-a-vis midlevel activists and the partyrank-and-file to seek and attain control of who will represent the party – the stakesare too high to leave to the vagaries of a more inclusive process. Uncertainty shouldaffect democratic and nondemocratic actors as party leaders would want to controlthe direction of change. There may still be challenges to leadership control of CS,

Democratization 805

Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

but we posit that uncertainty is more conducive to the leadership attaining andmaintaining control. While this cannot be tested here, we also posit that transitionswith greater certainty are not as constrained and do not prescribe particular CSprocedures. Therefore, certainty would support, but does not guarantee, greaterinclusiveness of CS procedures.

Actor calculations of regime uncertainty vary across parties, countries and overtime as transitions unfold. In addition to key actors’ perceptions of uncertainty, weassess regime uncertainty based on three conditions: first, the attitudes of the lastauthoritarian ruling elites toward democracy.41 Valenzuela categorizes casesaccording to whether the last authoritarian elites favoured full democratization(for example, Spain 1976), preferred a liberalized authoritarian regime butwould accept democratization (for example, Argentina 1983), or opposed demo-cratization (for example, Chile 1990). Secondly, the regime preference(s) andstrength of the military and paramilitary forces affect certainty. The military is ina unique position to impose its political preferences, and therefore its attitudetoward democratization affects actor calculations. The military’s regime preferencemay or may not correspond with the attitude of the last authoritarian regimedepending on whether the non-democratic regime was composed of and led by ahierarchical military, a nonhierarchical military, civilian leadership (including amonopoly party) or sultanistic leadership.42 Finally, the strength of non-democratic(and non-military) collective actors is important. For example, significant insur-gency movements, terrorist organizations, political parties or other collectiveactors that employ undemocratic means and/or advocate undemocratic ends willaffect actors’ perceptions of the stability of democracy.

In terms of our two country studies, both had highly uncertain transitions.In Spain, the last incumbent authoritarian government, under the guidance ofKing Juan Carlos, favoured full democratization and initiated the transitionprocess; however, transition initiators were civilian not military leaders. Themilitary was a key pillar of the Francoist regime, and its acceptance of democracywas seriously in question. Furthermore, the transition left the military largelyuntouched, with true military reform only occurring when the Socialistsassumed power after 1982. The Franco regime was, however, a civilianized author-itarian regime,43 suggesting a greater tendency of the military to respond to civiliangovernment, compared to the more militarized Chilean regime. However, this wastempered by the fact that Spanish armed-forces ministers were military leaders.Finally, nondemocratic and nonmilitary actors, particularly ETA’s use of terrorismplayed into the hands of those segments of the military reticent about the demo-cratic transition. Several coup plots were discovered and, on 23 February 1981,a coup attempt interrupted a session of the Congress of Deputies, holding theentire political class hostage. While the coup attempt was defeated with the helpof King Juan Carlos, this outcome could not have been predicted in advance.

Key party actors across the political spectrum felt the uncertainty. Rafael AriasSalgado, former UCD Secretary General, pointed to constant latent uncertaintybecause many generals thought democratization was ‘foolish’ (un disparate).44


Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

According the Jose Marıa (Txiki) Benegas, former PSOE Organizational Secretary,the conditions were ripe for a military response, including rampant terrorism:‘imagine what a year with a 100 [deaths] is; that is every three days a death,plus the funerals, the burials, the families. . .of course. . .and many militarydeaths’.45 Santiago Carrillo, former PCE Secretary General, noted:

Spanish democracy at that moment was walking along the edge of an abyss. It wasknown that the army was conspiring and that in block the army was hostile to demo-cratic development; it was the army that had won the civil war with Franco and morethan a Spanish army it was a Francoist army. . .We were always worried that, in spiteof being in the European democratic context, at any moment there could have been amilitary coup.46

In Chile, the Pinochet regime allowed a controlled transition despite itsunfavourable attitude toward democracy and backed a ‘distinctly authoritarian’constitutional design.47 Indeed, the hindsight of success blurs the real challengesto democracy posed by the military. While in many countries former dictatorsboard a plane for Miami or Paris when their regimes come to an end, Pinochetpacked his desk at La Moneda presidential palace and moved across the street tothe Ministry of Defense. Several tense conflicts followed. The ‘ejercicio deenlace’ of 1990 was one of the most serious, with Pinochet putting troops onhigh alert in response to incipient investigations into human rights abuses and rum-blings of trying the military. Allegations of conflicts of interests in the businessdealings of Pinochet’s son combined with military accusations that the Ministryof Defense was holding up military approved promotions led to the ‘boinazo’ of1993, where troops in full military gear surrounded the Ministry of Defense. YetChile had few nonmilitary, nondemocratic collective actors.48

Uncertainty was clearly perceived by the key party actors in Chile too. The firstGeneral Secretary of Presidency explicitly recognized in 1989 the military’s vetopower in a then confidential internal memo circulated among high-level govern-ment officials, noting: ‘The fear of a military regression, and the understandingof the risk of such an event occurring, will be directly determined by the level ofconflict that exists between political parties.’49 In our interview, the first presidentof the nascent democracy, Patricio Aylwin (1990–1994), underscored that hisgravest political problems were the continuation of militarism and the possibilityof Pinochet becoming a political persona equal to the President. Commenting onthe boinazo, he noted, Pinochet ‘always thought we were going to fail and thatthe country was going to call him back to replace me’.50 Similarly tensemoments followed Pinochet’s 1998 London arrest on charges of human rightsabuses. Though by the time of his mandate (2000–2006), President RicardoLagos thought uncertainty had largely been eliminated.51

Interviews provide evidence that uncertainty indeed informed candidate selec-tion, particularly in those parties that opposed the prior authoritarian regime. In theSpanish case, interviewees stated that party leaders sought to select candidates thatwere more moderate in the case of the PSOE when it was clear that the party could

Democratization 807

Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

win the 1982 elections and feared a military reaction52; less apt to push immedi-ately for outright independence from the Spanish state in the case of the PNV53;and that would follow the party executive’s dictates regarding the transitionprocess in the case of the PCE.54 In Chile documentary evidence55 and interviewswith party leaders and presidents underscored that party squabbling and lack ofunity in the form of arguments over spoils and candidacies could provide the mili-tary just the rationale it needed to justify intervention.56 In terms of ideology, theConcertacion’s task was eased by the marginalization of the Communist Party in aseparate alliance on the left, but complicated by the strategic complexity of theelectoral system (discussed below) where distinct parties and party factions hadto be accommodated to underwrite coalitional unity.

Strategic complexity of the election system

While the type of legislative election system cannot simply be tied to a particularpattern of CS, election system dynamics are not irrelevant. We posit that how thepolitical and party system contexts interact with the election system matters, andparticularly strategically complex electoral systems are more likely to elicitexclusive CS procedures to facilitate successful strategic coordination57; whileenvironments with few strategic coordination problems facilitate, but do notguarantee more inclusive CS procedures.

Strategic complexity operates at the systemic and party levels, and depends,not on the electoral system per se, but on its interaction with the individual partyand the party system. Building on Cox58 we define strategically complex electoralsystems as those in which parties must solve coordination problems in order to win,where winning for one candidate depends on the decisions and actions of otherpotential or actual allied candidates or their parties. While this can be said of allelectoral systems, certain situations provide higher probability of a failure of stra-tegic coordination, and we regard these as more strategically complex. Cox alsonotes that if a high number of agents care about ‘current and future policy out-comes, rather than just current seats’ that the ‘probability of strategic failureincreases’.59 For example, if alliances or pacts of exclusion are necessary for elec-toral success or if the electoral system compels deal-making at various electorallevels for the distribution of limited electoral spoils the greater the strategic com-plexity. Electoral rules that produce uncertain outcomes due to thresholds, complexrules, or unpredictability in the translation of votes to seats also produce greaterstrategic complexity. Finally, strategic complexity also varies for individualparties depending on size, negotiating capacity, and the competitive dynamicproduced by the election system. For example, majority parties expecting easyelectoral victory face less strategic complexity than a small party that needs tonegotiate candidacies to ensure representation in an alliance.

While tempting to assert that the higher number of candidates per district inlarge magnitude PR systems makes them more strategically complex, we arguethat such systems may actually provide fewer coordination problems. A


Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

multiplicity of parties can run a full complement of candidates without significantlyreducing their chances of winning seats. Victory depends on voters’ preferencesand not the outcome of successful negotiations over spoils. Small-magnitudesystems introduce relatively more strategic complexity where multiple partiesare forced to negotiate elaborate pre-electoral alliances to ensure the representationof many parties. In these cases, we argue that exclusive CS to solve coordinationproblems is likely to be the norm. Therefore, our contention is that few parties inboth small- and large-magnitude electoral systems and multiple parties in large-magnitude systems are less strategically complex than multiple parties in small-magnitude systems. List type may also be important. Closed-list systems maytend towards more strategic complexity than open list systems because selectingthe candidate often amounts to selecting the representatives, providing more incen-tives for elites to intervene to ensure the election of their nominees. If more thanone party appears on a single ballot (open or closed), negotiations and elite inter-vention aimed at solving coordination problems are the most likely mechanisms fordetermining list position across districts.

We recognize that the format of executive-legislative relations also matters. Weassert that in multiparty presidential systems electoral alliances are often necessarybefore elections, because presidential candidates seek to combine electoral listswith allied parties to be able to rely on multiparty support in the presidential elec-tion and later in the legislature. On the other hand, two-party presidential systemsmay produce less complexity, because presidents will either simply rely on amajority of their own party, or face a united opposition party with which theymust negotiate. Similarly, we argue that parliamentary systems, or the variousvarieties of semi-presidentialism, may create more or less strategic complexitydepending on how predictable or uncertain government formation is perceivedto be, and the composition of potential majorities.

Looking at our two cases, Chile’s election and multiparty systems presentedparties with an extremely high degree of strategic complexity, reinforcing the ten-dency toward exclusive CS. Since 1989, in each of the country’s 60 electoral districts,both the Concertacion and Alianza have presented a negotiated list of two candidatesthat emerged from negotiations between its member parties. The voter casts a singlevote for her preferred candidate on the two-candidate open list. Votes for bothcandidates on each list are first pooled to determine which list wins, allocatingseats via the d’Hondt method.60 After determining whether a list wins one or twoseats, seats are allocated to individual candidates based on vote shares.

While at first glance the system seems straightforward, the binominal electionsystem has created a good deal of strategic complexity because its operationalcharacteristics compel the most-voted list in each district to double the vote totalof the second place list to win both seats. If it does not, each of the two top-polling lists wins one seat. The system, thus, creates two powerful electoralthresholds. With a pattern of two alliance competition to win one seat, an alliancemust garner over 33.3% of the vote of the two largest finishers in each district, andin order to win two seats, it must win more than 66.6% of said vote.61

Democratization 809

Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

These high thresholds mean that one candidate on each alliance list will belikely to lose, making pairing on individual lists crucial, and distinct strategicdilemmas for different parties. Each party seeks to place its candidate on thesame list either with an extremely weak candidate (who they can handily beat),or an extremely strong candidate (who can carry the list to a rare two-seatvictory). Particularly on the left, negotiations are further complicated becausesmaller parties (in the transitional period these have included at various timesthe PPD, the Greens (PV), the PR and the SD – which later fused to form thestill very small PRSD) want to be placed not just on lists, but on lists wherethey can win. List partners from major parties will trounce candidates fromsmall parties, making small parties demand even weaker list partners. Strongparties, in turn, need weak parties to maintain the presidential coalition.62 Thesecoordination problems and the need to honour the electoral bargains provide astrong incentive for exclusive CS to manage the complexity and for each partyto negotiate itself out of its peculiar strategic dilemma. Former Chilean PresidentLagos noted that the binomial system ‘obligates centralized negotiation tochoose candidates, because it is not like the American system where in everydistrict you just have to choose one candidate and every party can simply doso’. After describing the horse trading involved and cases where primaries wereoverruled by alliances he continued: ‘that is the perverse part of the system. . .itobliges centralized control, because if you don’t have it, how are you going toengage in negotiations?’63

The dynamic of alliance formation on the right reinforces these points. TheRN’s principal negotiator, Roberto Palumbo, acknowledged that CS in theAlianza was less complicated because there was less horse trading with only twoparties involved.64 Therefore initial candidate designations are less frequentlytraded away by nonselected alliance negotiators. Nonetheless, in 1993 the insur-gent UCC agreed to join the Alianza after polling 7% in the previous municipalelection and was offered a generous 21 slates – however in all but three cases,UCC candidates would be paired with major parties. The party’s leader, FranciscoJavier Errazuriz responded with a request for 11 districts where UCC candidateswould compete only against each other, well aware that if forced to compete onthe same lists with major parties of the right that the UCC would lose. Summingup the strategic complexity faced by small parties paired with larger ones, Errazurizargued, ‘Don’t give me slates I am asking for seats’.65

Our argument concerning strategic complexity is reinforced by the evolution ofinternal party policies regarding candidate selection as the transition unfolded. Inparticular, Chilean parties did move to institute some limited primaries for presi-dential candidate selection. However, this push for more ‘democratic’ forms ofcandidate selection was limited in the case of legislative elections, largely due tothe extraordinary complexity faced by individual parties in negotiating to ensuretheir continued representation in congress.

Spanish parties, in turn, faced significantly less strategic complexity than theirChilean counterparts; yet, strategic complexity was present in varying degrees,


Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

ranging from moderate to high depending on the election year and party. TheSpanish Congress of Deputies has 350 members that are directly elected from50 multi-member district lists (provinces) and two single-member districts. Listsare closed and seats are allocated using the d’Hondt method of PR with aminimum threshold of 3% of the valid votes at the district level. The law guaran-teed a minimum of two seats per district, and there was an average district magni-tude of 6.7. The electoral system was designed to prevent fragmentation, producedisproportional effects and over-represent rural, presumably conservative districts.According to Gunther, the electoral system produces similarly disproportionalresults as single-member district systems.66

There are several aspects of the Spanish party and electoral systems that pro-duced strategic complexity: first, the relatively small district magnitude. Therewere 33 districts with fewer than 7 seats; 15 between 7 and 10 seats; and 4 withgreater than 10 seats.67 Only 19 (36.5%) districts were large enough (≥ sevenseats) to produce reasonably proportional results. Sixty-three and a half percenthave fewer than seven seats, which encourages cross-party electoral coordinationwhere a single large party does not exist. Secondly, when it became evident thatcompetitive parliamentary elections would occur in 1977, the political sceneexploded with political parties, and it was not clear which formations wouldsucceed. Polls conducted during the campaign found that 40% of respondentswere undecided.68 The incentives of the electoral law, the sheer number ofparties, and uncertainty created strong incentives for party mergers and alliancesto maximize their possibilities or minimize their losses.

At the party level, strategic complexity varied moderately. Party mergers andalliances were most pronounced on the right where no significant historic partiesemerged. Centrist and centre-right political forces allied in the UCD and moreright-wing forces allied in AP. Esteban and Lopez Guerra surmise that alliancepolitics required that list formation be left to the party leaders.69 In the UCD,list-makers had to accommodate the distinct parties and political families (Chris-tian democrats, liberals, azules,70 and social democrats), similar to the alliances inChile; and mold candidate lists to make the party appear more moderate to theelectorate.71 The greater number of parties on the centre-right and right compet-ing for votes further heightened the degree of strategic complexity. While thePSOE benefited from its status as a historic party, and did not widely employ elec-toral alliances, multiple socialist and centre-left formations emerged, includingthe Partido Socialista Popular (PSP), PSOE (historico), and several regionalsocialist parties. In subsequent elections (1979 and 1982), the party reachedagreements with many of these formations, including the social democratsfrom the UCD, to join the PSOE. PSOE leaders intervened in CS to fulfil theseagreements.72 Party leaders determined not only the number of seats grantedon candidate lists but also the placement on the closed lists. While strategic com-plexity likely encouraged exclusive selection in all parties, the greater strategiccomplexity facing the UCD helps explain its moderately more exclusiveprocedures.

Democratization 811

Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

Party leadership autonomy

While control of CS can be a powerful tool to influence the types of candidates andultimately legislators, party leaderships vary in their ability to determine whichselection procedures are adopted. Party leadership autonomy refers to the abilityof the central party leadership to make decisions relatively independently ofother party and non-party affiliated actors without systematic vertical checks.Accountability means that the party leadership’s decisions are more constrainedby other actors or internal party processes.73 Our proposition is that high party lea-dership autonomy by definition allows party leaders greater latitude to choosewhich procedures are employed to select candidates. Yet, while likely to be mostclosely associated with exclusive forms of CS party leadership autonomy is com-patible with a variety of CS procedures because the leaders’ primary goals canvary.74 For example, leaders may see inclusive CS as an effective strategy toattract members and/or to improve electoral prospects,75 even when they are notforced by internal pressure to do so. Gallagher demonstrates that party organiz-ational weakness is compatible with primaries, as occurred in the United States,or with party leadership selection, as in India.76 Yet, in highly constrained tran-sitions, such as the Chilean and Spanish ones, leaders are likely to prefer moreexclusive CS procedures. Leadership accountability, on the other hand, encouragesthe inclusion of powerful party actors because they are in a better position tocompel their place in the process; the precise form of inclusion will depend onthe internal dynamics of the party.77 However, this does not necessarily encouragethe inclusion of rank-and-file party members or the electorate. While accountableleaders may still manage to exact control of CS, they must face and convince theother significant actors that they should have this control.

A few additional notes are in order here. First, while the three new explanatoryfactors have been presented as analytically distinct, they clearly interact.Uncertainty and strategic complexity can heighten party leadership autonomy.However, they are not synonymous. Leadership autonomy can stem from avariety of sources, and can exist without strategic complexity and uncertainty. Sec-ondly, while we cannot fully test this here, the degree of party leadership autonomyis likely to have the greatest independent effect on the choice of CS procedures whenthe constraints of uncertainty and strategic complexity are low or point in opposingdirections. Thirdly, party leadership autonomy is extremely difficult to measure intransitional environments where decision-making processes may not be formalizedor well-institutionalized, and where behavioural norms may have little to do withthe formal procedures. Therefore, we must take into consideration a variety ofcharacteristics that could affect leadership autonomy. We focus on the distributionof resources within the parties and state institutions, the degree to which the centralparty leadership was disputed, and leaders’ perceptions of their autonomy.

Using the above guidelines, the Spanish and Chilean parties exhibited highleadership autonomy that was reinforced by uncertainty and strategic complexity.While we focus on the two largest parties in Spain, the UCD and PSOE, all


Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

Spanish party leaderships were unconstrained by strong party organizations as eitherthe parties were entirely new or were severely weakened by decades of authoritarianrule. Resources were also concentrated in the central party office and in nationallevel political institutions. Party- and election-financing laws, designed by theparty leaders themselves, provided state funding to the central party headquartersfrom the outset of the transition. Also, though it was clear that decentralizationwould occur in the heavily centralized political system, at least to Catalonia, theBasque Country and Galicia, actual (and substantial) decentralization did notoccur until years later.78 The first local democratic elections took place in 1979(after the second national parliamentary elections). Regional autonomy charterswere approved between 1979 and 1983. A few regional elections occurred in1980 (after two parliamentary elections), yet most were not held until 1983 (afterthe third parliamentary elections). Therefore power and resources were concentratedin the national level and in the central party offices at the outset of democracy.

The leaderships of the parties were relatively undisputed at the outset of the tran-sition. Felipe Gonzalez and his team had taken control of the PSOE from the leader-ship in exile in 1974, and had won recognition from the Socialist International.Those disenchanted with this outcome formed alternative socialist parties, includ-ing the PSOE (historico). Though the PSOE did not win a general election until1982, Gonzalez’s popularity was undeniable as he helped deliver the unexpectedlyhigh electoral success in the founding 1977 elections. The degree to which the lea-dership was indispensable is illustrated by Gonzalez’s stepping down as GeneralSecretary in 1979 when the party Congress voted against removing the referenceto Marxism from the party’s statutes. There was no viable alternative. Gonzalez ulti-mately returned and the statutes were modified. These events further consolidatedthe leadership. In the case of the UCD, while there were a number of party and fac-tional leaders who wanted to lead the UCD; it was clear that only with Suarez wouldthe nascent alliance have any chance of winning the elections. Notably, when theUCD began to divide internally beginning in 1981, Suarez sought to concentrateCS for the 1982 elections exclusively in his hands. At that point his leadershipwas clearly in question, and his demand was rejected. He abandoned the party in1981 to form his own, at which time the UCD entered terminal illness and ulti-mately disbanded in 1983.

Party leaders themselves perceived substantial decision-making autonomy.Regarding leadership autonomy at the outset of democracy, Alfonso Guerra(former PSOE Vice-Secretary General) tellingly stated: ‘From a statutory per-spective, very little; from the perspective of moral authority, a lot, a lot. That isto say that if a local party body said that this is what we needed to do, very diffi-cult to derail that, very difficult to derail from an organic/statutory perspective.Now, if Felipe Gonzalez or I said, look this isn’t the way, it was very, verydifficult that they didn’t pay attention’.79 Similar sentiment was expressed bythe leaders of the UCD, PCE, AP, and Convergencia Democratica de Catalunya(CDC).80 In the words of Miquel Roca, CDC General Secretary at the time: ‘theparty didn’t have many power centres’.81 Spain is therefore a case of a highly

Democratization 813

Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

constrained new democracy – high uncertainty and moderately high strategiccomplexity. In this context, autonomous party leaders ultimately selected legis-lative candidates.

While Chilean political parties had more elaborate party organizations andfirmer roots in society,82 in some cases due to their survival during authoritarianrule83 and in part due to the need to organize for the plebiscite on Pinochet’scontinued rule,84 the party leadership still enjoyed ample autonomy. As Angellstates: ‘One feature that has remained constant is the highly centralized nature ofthe parties, in which real decision-making is limited to a small elite’.85 Whilethe parties may have enjoyed high levels of identification among party activistsin early years of the transition, the organizational mechanisms through whichthese activists could exercise control over leaders had yet to develop. In addition,the dictatorship weakened organizations like trade unions that had traditionallyacted as a check on parties of the centre and left. As the democratic transitionadvanced these groups also surrendered influence in the interests of securing thedemocratic transition.86

At the outset of the transition party leaders also had independent control overparty financing and regularly negotiated policy deals with the opposition and alliedparties. Furthermore, there were few resources available to potential challengers tothe party leadership. Chile is a unitary system; therefore, regional elites armed withsignificant state resources did not exist. Even elections of local governments,which have limited functions,87 did not occur until 1992. Regional and localparty branches were therefore weak vis-a-vis the national party office.

Before the Chilean transition began, leadership was highly contested on the left.Factions of the Socialist Party vied for control, with much of the leadership structuredecapitated by the death or exile of members. Carlos Altamirano led a faction ofmore moderate socialists, while Cladomiro Almeyda’s faction increasingly alliedwith Chile’s communists to take a more militant line.88 Less traumatized by theregime, the Christian Democratic leadership structure was largely frozen, withlong-standing members like Patricio Aylwin, Gabriel Valdes, Andres Zaldıvar,and former president Eduardo Frei remaining as the largely undisputed, naturalleaders of the party. Dealt a blow by Frei’s death in 1982, the PDC experiencedlimited leadership struggles between its chascon (long-haired or progressive) andguaton (pot-bellied or conservative) factions. While leadership disputes existed,they were based largely on personal followings, and the citizenry viewed thiscoterie of leaders as the natural inheritors of the new democracy.

Pinochet’s strategy of calling a plebiscite in 1988 unintentionally solidified thisleadership by providing strong incentives for unity. The 1985 National Accord fora Full Transition to Democracy unified the centre-left, in its first significant agree-ment, and Pinochet’s decision for a yes or no plebiscite on his continued rule in1988 sealed this unity. Several factions of the moderate Socialist Party joinedthe PDC to form the Concertacion alliance, while the Almeyda Socialist leadershipallied with the Communist Party. Those who gained popularity and support in cam-paigning for, and ultimately winning the ‘no’ vote emerged as the unchallenged


Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

leaders of their parties, including Patricio Aylwin in the Christian Democratic Partyand Ricardo Lagos in the PPD. Leaders themselves noted a high degree of leader-ship autonomy. PDC leader Jorge Navarrete contended that party leaders are ‘com-pletely autonomous for the majority of areas where they must make decisions withvery few exceptions’ and that party culture is ‘very vertical. . .where politics isundertaken by specialists who make decisions between four walls away from theview of citizens’.89 To maintain unified opposition, and later the unity to govern,meant that party elites had to strike deals across parties and between parties andsocial actors. Continuing divisions within the party system necessitated extensivecoordination to navigate the very complex post-authoritarian political dynamic.Party leaders demanded and were able to present a rationale for almost completeautonomy from the rank and file in order to strike and enforce these agreements.

Party leadership autonomy on the right is much less complex. In the firstelections following the return of democracy, leaders on the right had a virtuallyfree hand because they were in most cases authoritarian regime incumbents andparty organizations were virtually non-existent.90 With the return of democracythe historically weakly institutionalized nature of the traditional right (fromwhich the RN grew) left RN leaders subject to little pressure from party activists,while as a new party, UDI leaders self-consciously built an elite dominatedcentralized party organization as a strategy for growth and selecting candidatesthat the leadership felt could win.91 In the Chilean system, the very highdegree of uncertainty and strategic complexity of the electoral system andparty leadership autonomy encouraged a CS process largely in the hands of theparty leaderships.

Conclusion

Despite important advances in the literature on CS procedures in recent years, thereare few studies that focus on the origins or determinants of distinct forms of selec-tion. In this article, we have provided a number of propositions that put the veryorigins of the CS process under the microscope. In doing so we aim to providean alternative, more powerful comparative explanation for choice of CS pro-cedures than those traditionally relied upon in the literature. We contend thatwithin transitional systems distinct transitional environments condition partychoice of CS procedures. In particular we posit that the relative levels of uncer-tainty about the installation and continuance of democracy, strategic complexityof the electoral system, and party leadership autonomy, conspire to create incen-tives for the adoption of more or less inclusive procedures.

In Spain, the parties faced a very uncertain transition process due to the poten-tial for military intervention. While not as complex as in the Chilean case, the sheernumber of parties combined with a PR system, which in practice produces dispro-portional results, and the uncertainty of electoral outcomes created moderate tohigh strategic complexity. In this context, autonomous party leaders or leadershipbodies ultimately controlled CS. Significantly, the UCD faced even greater

Democratization 815

Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

strategic complexity than the PSOE due to its origins as an alliance of multiple pol-itical parties; and it also employed more exclusive CS procedures. Similarly inChile, the significant variables conspired to push the CS process into the handsof elites. Chile is clearly a case of an uncertain transition, and the imperative of alli-ance formation and party consensus-building was central to the maintenance ofdemocracy. In turn, the imperative to accommodate multiple parties within eachalliance and on two-member electoral lists created enormous strategic complexity.Finally, all significant political parties were characterized by leadership little con-strained by vertical checks.

Because the study of the genesis of CS procedures is in its infancy, we see thiswork as a hypothesis generating case study of the type identified years ago by Lij-phart.92 Subsequent research needs to be done in order to substantiate, or refute,in more detail the propositions set out here. In particular, future empirical testsusing parties that employ more inclusive selection procedures will allow us todevelop, test and refine the findings. Additionally, scholars must assess thepath dependency of CS procedures. On the one hand, historic parties in newdemocracies may be more likely to re-adopt CS procedures they used in thepast; however, this cannot explain the procedures adopted by new parties, suchas the UCD in Spain, or why historic parties, such as the Partido Justicialista(PJ) in Argentina depart from prior selection procedures. On the other hand,we need to determine how ‘sticky’ candidate selection procedures are in newdemocracies and in turn what accounts for change. While this is not the focusof the research presented here, two potential implications can be derived: (1)changes over time in the variables presented may account for change in selectionprocedures; or that (2) once the transitional period has terminated, other factorsbecome more important.

In addition to helping to explain the origins of a process so central to represen-tation, we hope to spur additional research on the relationships between candidateselection and democratization processes, about which little is written. Yet there ispotential value in this line of research. There is some face validity to the propositionthat CS procedures may have an indirect effect on the probability of consolidatingdemocracy. O’Donnell and Schmitter claim that the success of minimalists (mod-erates) over maximalists (radicals) in the founding elections is best for installingdemocracy but provide little insight into when and why this occurs.93 Do distinctCS mechanisms foster moderation or radicalism, or more generally affect the typesof candidates and thus legislators? Does this behaviour then affect consolidationprocesses? Also, what effect do distinct selection procedures have on party institu-tionalization? It is plausible that different CS procedures may offer distinct incen-tives for exit from, voice within or loyalty to the party.94 CS processes may alsointeract with distinct modes of transition. For example, Field’s comparative analy-sis of the pacted transitions in Spain, Colombia and Venezuela, demonstrates that,in the absence of formal institutional guarantees of cross-party consensus, disci-plined political parties with CS controlled by the pact-making elites can providethe mutual assurance necessary to make political pacts credible.95 Additionally,


Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

it is critical to study the trade-offs involved in the adoption of distinct CS pro-cedures; for example, the potential representational trade-off between broader par-ticipation in the selection of candidates and the representation of distinct subsectorsof society that have historically been underrepresented, such as women.96 In sum,we have much to learn about how candidate selection affects the democratizationprocess and vice-versa.

AcknowledgementsThe authors thank the anonymous reviewers and the editorial team for comments on pre-vious drafts. Bonnie Field thanks the Fundacion Transicion Espanola and the Program forCultural Cooperation between Spain’s Ministry of Culture and United States Universitiesfor their generous research funding and Paige Roland for research assistance. Peter Siavelisacknowledges the support of the Social, Behavioral and Economic Sciences Research Fundat Wake Forest University.

Notes1. Ranney, ‘Candidate Selection’, 75.2. For a review see Field and Siavelis, ‘Candidate Selection Procedures in Transitional

Polities’.3. For example, Duverger, Political Parties; Gallagher and Marsh, Candidate Selection

in Comparative Perspective; Michels, Political Parties; Norris, Passages to Power;Rahat and Hazan, ‘Candidate Selection Methods’; Siavelis and Morgenstern, Path-ways to Power. However there is little published on the effects of CS on democratiza-tion processes.

4. Gallagher, ‘Introduction’; Gallagher, ‘Conclusion’; Hermens, Democracy orAnarchy?, 51–9; Epstein, Political Parties in Western Democracies, 225–6.

5. Rahat and Hazan, ‘Candidate Selection Methods’.6. Ibid., 300.7. Gallagher, ‘Introduction’; Gallagher, ‘Conclusion’; Hermens, Democracy or

Anarchy?, 51–9; and more recently, Jones, ‘The Recruitment and Selection of Legis-lative Candidates’; Siavelis, ‘The Hidden Logic of Candidate Selection’.

8. Gallagher, ‘Introduction’, 9–10.9. Epstein, Political Parties in Western Democracies, 225–6.

10. Ibid., 219–25.11. Ibid., 209, 229; Gallagher, ‘Introduction’; Gallagher, ‘Conclusion’; Harmel, ‘Environ-

ment and Party Decentralization’.12. Spain began decentralizing in 1980 with no significant change in CS procedures.13. Epstein, Political Parties in Western Democracies, 207, 210, 231; Gallagher, ‘Con-

clusion’; Key, Politics, Parties and Pressure Groups, 377.14. While this departs from the research design, it is clear that the use of exclusive CS pro-

cedures by all of the parties within each country casts doubt on explanations basedsimply on ideological, left-right categorizations.

15. Gallagher and Marsh, Candidate Selection in Comparative Perspective, cover matter.16. Sartori, Parties and Party Systems, 123.17. Ragin, ‘Turning the Tables’.18. George and Bennett, Case Studies and Theory Development; McKeown, ‘Case Studies

and the Limits of the Quantitative Worldview’; Munck, ‘Tools for Qualitative Research’.19. Ragin, ‘Turning the Tables’.20. Ibid.

Democratization 817

Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

21. Munck, ‘Tools for Qualitative Research’.22. Przeworski and Teune, The Logic of Comparative Social Inquiry.23. See George and Bennett, Case Studies and Theory Development, 153–60.24. Munck, ‘Tools for Qualitative Research’.25. George and Bennett, Case Studies and Theory Development, 153–60.26. Ibid., 206.27. O’Donnell and Schmitter, Transitions from Authoritarian Rule, 58.28. Linz and Stepan, Problems of Democratic Transition and Consolidation, 5.29. Ibid., 108; Aguero, Soldiers, Civilians and Democracy; Gunther, Diamandouros, and

Puhl, The Politics of Democratic Consolidation, 21.30. Linz and Stepan, Problems of Democratic Transition and Consolidation, 3; Weeks,

‘The Transition is Dead’.31. Esteban and Lopez Guerra, ‘Electoral Rules and Candidate Selection’; Gillespie, The

Spanish Socialist Party, 343; Tusell Gomez, ‘The Democratic Center and ChristianDemocracy’.

32. Ramırez, Ası se Ganaron las Elecciones, 121.33. Hopkin, Party Formation and Democratic Transition, 89.34. Esteban and Lopez Guerra, ‘Electoral Rules and Candidate Selection’, 68.35. Ibid., 71.36. Interview with List Committee member Joaquin Leguina, May 19, 2009.37. Navia, ‘Legislative Candidate Selection in Chile’.38. Interview with Jorge Navarrete, August 18, 2008.39. Navia, ‘Legislative Candidate Selection in Chile’.40. See Coppedge, Strong Parties and Lame Ducks; Field, ‘Transitions to Democracy and

Internal Party Rules’; Thiebault, ‘France’.41. Valenzuela, ‘Democratic Consolidation in Post-transitional Settings’.42. Linz and Stepan, Problems of Democratic Transition and Consolidation.43. Aguero, Soldiers, Civilians and Democracy, 43–58.44. Interview, May 30, 2008.45. Interview, May 27, 2008.46. Interview, May 29, 2008.47. Aguero, Soldiers, Civilians and Democracy, 36.48. Though groups like the Movimiento de Izquierda Revolucionaria (MIR) and Frente

Patriotico Manuel Rodriguez (FPMR) engaged in violent anti-government activitiesduring the dictatorship, including the attempted assassination of Pinochet by theMIR, neither was a significant actor with the potential to destabilize democracyduring the transition period on the order of ETA in Spain.

49. Boeninger, ‘Gestion de Gobierno’.50. Interview, August 20, 2008.51. Interview, August 18, 2008.52. Interview with Gregorio Peces-Barba (PSOE), May 14, 2009.53. Interview with Marcos Vizcaya (PNV), May 28, 2010.54. Interview with Julio Anguita (PCE), May 13, 2010.55. Boeninger, ‘Gestion de Gobierno’.56. Interview with Patricio Aylwin (PDC), August 20, 2008.57. Siavelis, ‘The Hidden Logic of Candidate Selection’, 436–7.58. Cox, Making Votes Count.59. Ibid., 9.60. Siavelis, The President and Congress, 35; Navia, ‘Legislative Candidate Selection in

Chile’, 96.61. See Navia, ‘Legislative Candidate Selection in Chile’; Siavelis, ‘The Hidden Logic of

Candidate Selection’.


Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

62. Siavelis, ‘The Hidden Logic of Candidate Selection’.63. Interview, August 18, 2008.64. Interview, August 19, 2008.65. Interview with Andres Allamand (RN), July 13, 1998.66. Gunther, ‘Electoral Laws, Party Systems, and Elites’, 836.67. Authors’ calculation based on data from the Ministerio del Interior, Spain, http://www.

elecciones.mir.es/ (accessed January 15, 2010).68. Gunther, Sani, and Shabad, Spain after Franco, 99.69. Esteban and Lopez Guerra, ‘Electoral Rules and Candidate Selection’, 63.70. Azules (blues) refers to those with close ties to Suarez and the prior regime.71. Anonymous interview with a high ranking UCD official, June 5, 2008.72. Interview with Txiki Benegas, May 27, 2008.73. Kitschelt, The Transformation of European Social Democracy.74. Our analysis does not assume party leaders’ vote-maximizing behaviour; rather they

can have multiple goals.75. Adams and Merrill, ‘Candidate and Party Strategies’; Carey and Polga-Hecimovich,

‘Primary Elections and Candidate Strength’.76. Gallagher, ‘Conclusion’.77. Where parties are heavily factionalized, more inclusive CS procedures can be used to

resolve internal disputes. See Kemahlioglu, Weitz-Shapiro, and Hirano, ‘Why Pri-maries in Latin American Presidential Elections’.

78. See, for example, Gunther, Montero, and Botella, Democracy in Modern Spain, 280–334; Moreno, The Federalization of Spain.

79. Interview, June 4, 2008.80. CiU is formed by two parties, CDC and Unio Democratica de Catalunya.81. Interview, May 16, 2008.82. Mainwaring and Scully, Building Democratic Institutions; Scully, ‘Reconstituting

Party Politics in Chile’.83. Garreton, ‘La oposicion polıtica’.84. Scully, ‘Reconstituting Party Politics in Chile’.85. Angell, ‘The Durability of the Party System’, 291.86. Siavelis, The President and Congress in Post-Authoritarian Chile, 81–4.87. Angell, ‘The Durability of the Party System’, 291–2.88. Constable and Valenzuela, A Nation of Enemies.89. Interview, August 18, 2008.90. Interview with Roberto Palumbo, August 19, 2008.91. Navia, ‘Legislative Candidate Selection in Chile’.92. Lijphart, ‘Comparative Politics and the Comparative Method’.93. O’Donnell and Schmitter, Transitions from Authoritarian Rule, 61–4.94. Hirschman, Exit, Voice, and Loyalty.95. Field, ‘Transitions to Democracy and Internal Party Rules’.96. Baldez, ‘Primaries vs. Quotas’.

Notes on contributorsBonnie N. Field is associate professor of global studies at Bentley University. She was aFaculty Fellow at the University of California, Irvine, Visiting Fellow at UCI’s Center forthe Study of Democracy, and Fulbright Senior Researcher in Spain. Her publicationsfocus on democratization, political parties and political institutions in Spain and LatinAmerica. She is editor of Spain’s ‘Second Transition’? The Socialist Government of JoseLuis Rodrıguez Zapatero (Routledge 2011) and co-editor (with Kerstin Hamann) of

Democratization 819

Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

http://www.elecciones.mir.es/

http://www.elecciones.mir.es/

Democracy and Institutional Development: Spain in Comparative Theoretical Perspective(Palgrave-Macmillan 2008).

Peter M. Siavelis is Associate Professor of Political Science and Director of the Latin Amer-ican and Latino Studies Program at Wake Forest University. He has published on LatinAmerican electoral and legislative politics including publications in Comparative Politics,Comparative Political Studies, Party Politics, Latin American Research Review, and LatinAmerican Politics and Society. He is the author of The President and Congress in Post-authoritarian Chile: Institutional Constraints to Democratic Consolidation (Penn StatePress, 2000) and co-editor of Pathways to Power: Political Recruitment and CandidateSelection in Latin America (Penn State Press, 2008), with Scott Morgenstern.

BibliographyAdams, James, and Samuel Merrill III. ‘Candidate and Party Strategies in Two-Stage

Elections Beginning with a Primary’. American Journal of Political Science 52, no. 2(2008): 344–59.

Aguero, Felipe. Soldiers, Civilians and Democracy: Post Franco Spain in ComparativePerspective. Baltimore, MD: Johns Hopkins University Press, 1995.

Angell, Alan. ‘The Durability of the Party System in Chile’. In Party Politics in NewDemocracies, ed. Paul Webb and Stephen White, 275–303. Oxford: OxfordUniversity Press, 2007.

Baldez, Lisa. ‘Primaries vs. Quotas: Gender and Candidate Nomination in Mexico, 2003’.Latin American Politics and Society 49 (2007): 69–96.

Boeninger, Edgardo. ‘Gestion de Gobierno y Proceso de Decisiones Publicas’. An InternalMemo Circulated within the Chilean Presidential Palace, Santiago, 1989.

Carey, John M., and John Polga-Hecimovich. ‘Primary Elections and Candidate Strength inLatin America’. Journal of Politics 68, no. 3 (2006): 530–43.

Constable, Pamela, and Arturo Valenzuela. A Nation of Enemies. New York: W.W. Norton,1991.

Coppedge, Michael. Strong Parties and Lame Ducks: Presidential Partyarchy andFactionalism in Venezuela. Stanford, CA: Stanford University Press, 1994.

Cox, Gary. Making Votes Count. Cambridge: Cambridge University Press, 1997.Duverger, Maurice. Political Parties: Their Organization and Activity in the Modern State,

trans. Barbara North and Robert North, 2nd ed. New York: John Wiley & Sons, Inc.,1959.

Epstein, Leon D. Political Parties in Western Democracies. New York: Praeger, 1967.Esteban, Jorge de, and Luis Lopez Guerra. ‘Electoral Rules and Candidate Selection’. In

Spain at the Polls 1977, 1979, and 1982, ed. Howard R. Penniman and EusebioM. Mujal-Leon, 48–72. Durham, NC: Duke University Press, 1985.

Field, Bonnie N. ‘Transitions to Democracy and Internal Party Rules: Spain in ComparativePerspective’. Comparative Politics 39, no. 1 (2006): 83–102.

Field, Bonnie N., and Peter M. Siavelis. ‘Candidate Selection Procedures in TransitionalPolities: A Research Note’. Party Politics 14, no. 5 (2008): 619–38.

Gallagher, Michael. ‘Conclusion’. In Candidate Selection in Comparative Perspective: TheSecret Garden of Politics, ed. Michael Gallagher and Michael Marsh, 236–83. London:Sage Publications, 1988.

Gallagher, Michael. ‘Introduction’. In Candidate Selection in Comparative Perspective: TheSecret Garden of Politics, ed. Michael Gallagher and Michael Marsh, 1–19. London:Sage Publications, 1988.

Gallagher, Michael, and Michael Marsh, eds. Candidate Selection in ComparativePerspective: The Secret Garden of Politics. London: Sage Publications, 1988.


Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

Garreton, Manuel Antonio. ‘La Oposicion Polıtica Partidaria En El Regimen MilitarChileno. Un Proceso De Apredizaje Para La Transicion’. In Muerte Y Resurreccion,ed. Marcelo Cavarozzi and Manuel Antonio Garreton, 395–466. Santiago: FLACSO,1989.

George, Alexander L., and Andrew Bennett. Case Studies and Theory Development in theSocial Sciences. Cambridge, MA: MIT Press, 2005.

Gillespie, Richard. The Spanish Socialist Party: A History of Factionalism. Oxford:Clarendon Press, 1989.

Gunther, Richard. ‘Electoral Laws, Party Systems, and Elites: The Case of Spain’. AmericanPolitical Science Review 83, no. 3 (1989): 835–58.

Gunther, Richard, Giacomo Sani, and Goldie Shabad. Spain after Franco: The Making of aCompetitive Party System. Berkeley: University of California Press, 1988.

Gunther, Richard, P. Nikiforos Diamandouros, and Hans-Jurgen Puhl. The Politics ofDemocratic Consolidation. Baltimore, MD: Johns Hopkins, 1995.

Gunther, Richard, Jose Ramon Montero, and Joan Botella. Democracy in Modern Spain.New Haven, CT: Yale University Press, 2004.

Harmel, Robert. ‘Environment and Party Decentralization: A Cross-National Analysis’.Comparative Political Studies 14, no. 1 (1981): 75–99.

Hermens, F.A. Democracy or Anarchy? A Study of Proportional Representation. New Yorkand London: Johnson Reprint Corporation, 1972.

Hirschman, Albert O. Exit, Voice and Loyalty: Reponses to Decline in Firms, Organizations,and States. Cambridge, MA: Harvard University Press, 1970.

Hopkin, Jonathan. Party Formation and Democratic Transition in Spain: The Creationand Collapse of the Union of the Democratic Centre. New York: St. Martin’s Press, 1999.

Jones, Mark P. ‘The Recruitment and Selection of Legislative Candidates in Argentina’. InPathways to Power, ed. Peter M. Siavelis and Scott Morgenstern, 41–75. UniversityPark: Pennsylvania State University Press, 2008.

Kemahlioglu, Ozge, Rebecca Weitz-Shapiro, and Shigeo Hirano. ‘Why Primaries in LatinAmerican Presidential Elections?’. Journal of Politics 71, no. 1 (2009): 339–52.

Key, V.O. Politics, Parties, & Pressure Groups. New York: Thomas Y. Crowell Company,1964.

Kitschelt, Herbert. The Transformation of European Social Democracy. New York:Cambridge University Press, 1994.

Lijphart, Arend. ‘Comparative Politics and the Comparative Method’. The AmericanPolitical Science Review 65 (1971): 682–93.

Linz, Juan J., and Alfred Stepan. Problems of Democratic Transition and Consolidation:Southern Europe, South America, and Post-Communist Europe. Baltimore, MD:Johns Hopkins University Press, 1996.

Mainwaring, Scott, and Timothy R. Scully, eds. Building Democratic Institutions: PartySystems in Latin America. Stanford, CA: Stanford University Press, 1995.

McKeown, Timothy J. ‘Case Studies and the Limits of the Quantitative Worldview’. InRethinking Social Inquiry: Diverse Tools, Shared Standards, ed. Henry E. Brady andDavid Collier, 139–68. New York: Rowman & Littlefield Publishers, 2004.

Michels, Robert. Political Parties: A Sociological Study of the Oligarchical Tendencies ofModern Democracies, trans. Eden Paul and Cedar Paul. Glencoe, IL: The Free Press, 1949.

Moreno, Luis. The Federalization of Spain. London: Frank Cass, 2001.Munck, Gerardo L. ‘Tools for Qualitative Research’. In Rethinking Social Inquiry: Diverse

Tools, Shared Standards, ed. Henry E. Brady and David Collier, 105–22. New York:Rowman & Littlefield Publishers, 2004.

Navia, Patricio. ‘Legislative Candidate Selection in Chile’.In Pathways to Power, ed. PeterM. Siavelis and Scott Morgenstern, 92–118. University Park: Pennsylvania StateUniversity Press, 2008.

Democratization 821

Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

Norris, Pippa, ed. Passages to Power: Legislative Recruitment in Advanced Democracies.Cambridge: Cambridge University Press, 1997.

O’Donnell, Guillermo, and Philippe C. Schmitter. Transitions from Authoritarian Rule:Tentative Conclusions About Uncertain Democracies. Baltimore, MD: Johns HopkinsUniversity Press, 1986.

Przeworski, Adam, and Henry Teune. The Logic of Comparative Social Inquiry. New York:Wiley-Interscience, 1970.

Ragin, Charles C. ‘Turning the Tables: How Case-Oriented Research Challenges Variable-Oriented Research’. In Rethinking Social Inquiry: Diverse Tools, Shared Standards, ed.Henry E. Brady and David Collier, 123–38. New York: Rowman & LittlefieldPublishers, 2004.

Rahat, Gideon, and Reuven Y. Hazan. ‘Candidate Selection Methods: An AnalyticFramework’. Party Politics 7, no. 3 (2001): 297–322.

Ramırez, Pedro J. Ası se Ganaron las Elecciones. Barcelona: Editorial Planeta, 1977.Ranney, Austin. ‘Candidate Selection’. In Democracy at the Polls: A Comparative Study of

Competitive National Elections, ed. David Butler, Howard R. Penniman, and AustinRanney, 75–106. Washington, DC: American Enterprise Institute for Public PolicyResearch, 1981.

Sartori, Giovanni. Parties and Party Systems: A Framework for Analysis. Cambridge:Cambridge University Press, 1976.

Scully, Timothy R. ‘Reconstituting Party Politics in Chile’. In Building DemocraticInstitutions: Party Systems in Latin America, ed. Scott Mainwaring and TimothyR. Scully, 100–37. Stanford, CA: Stanford University Press, 1995.

Siavelis, Peter. The President and Congress in Post-Authoritarian Chile. University Park:Penn State University Press, 2000.

Siavelis, Peter. ‘The Hidden Logic of Candidate Selection for Chilean ParliamentaryElections’. Comparative Politics 34, no. 4 (2002): 419–38.

Siavelis, Peter M., and Scott Morgenstern, eds. Pathways to Power. University Park: PennState University Press, 2008.

Thiebault, Jean-Louis. ‘France: The Impact of Electoral System Change’. In CandidateSelection in Comparative Perspective: The Secret Garden of Politics, ed. MichaelGallagher and Michael Marsh, 72–93. London: Sage Publications, 1988.

Tusell Gomez, Javier. ‘The Democratic Center and Christian Democracy in the Elections of1977 and 1979’. In Spain at the Polls, 1977, 1979, and 1982, ed. Howard R. Pennimanand Eusebio M. Mujal-Leon, 88–128. Durham: Duke University Press, 1985.

Valenzuela, J. Samuel. ‘Democratic Consolidation in Post-Transitional Settings: Notion,Process, and Facilitating Conditions’. In Issues in Democratic Consolidation: TheNew South American Democracies in Comparative Perspective, ed. ScottMainwaring, Guillermo O’Donnell, and J. Samuel Valenzuela, 57–104. Notre Dame:University of Notre Dame Press, 1992.

Weeks, Gregory. ‘The Transition is Dead, Long Live the Transition: Civil-Military Relationsand the Limits of Consensus’. In The Bachelet Government: Conflict and Consensus inPost-Pinochet Chile, eds. Silvia Borzutzky and Gregory B. Weeks, 67–84. Gainesville,FL: University Press of Florida, 2010.


Dow

nloa

ded

by [

Wak

e Fo

rest

Uni

vers

ity]

at 0

9:06

05

Sept

embe

r 20

11

Endogenizing legislative candidate selection procedures in nascent democracies: evidence from Spain...

Documents

Transcript of Endogenizing legislative candidate selection procedures in nascent democracies: evidence from Spain...