More often than not, contemporary works on political parties start by referring to Schattschneider’s now-famous dictum concerning democracy’s need for political parties. At the same time, many authors have identified parties that, in democratic contexts, fail in various ways to fulfill the function of democratic representation. Mainstream political science has defined a political party as a group of candidates who compete in elections (Reference DownsDowns 1957 and Reference SchlesingerSchlesinger 1994, among many others). This minimal definition has important analytical implications. When analyzing electoral politics, we run the risk of looking for parties – and thus, finding them – without realizing that what we have found, empirically, is only weakly related to democratic representation. In this introduction to the edited volume we present a thick definition of political parties to provide a conceptual framework for classifying different diminished subtypes of political parties in democratic regimes. The volume builds upon the rich literature concerning political parties that highlights the ways in which many party organizations are failing to fulfill their representational role in contemporary democracies. The empirical chapters that follow this introduction apply our conceptual framework to analyze seventeen parties in twelve Latin American countries.
Minimalist definitions of political party (i.e., Schlesinger’s 1994) seem disconnected from reality, that is, the proliferation of electoral vehicles that do not function as parties. The sole attribute of the minimalist definition of a political party is not theoretically linked to a central aspect of democracy, namely the representation of social interests and values. As Reference KitscheltKitschelt (2000) claims, parties “in the institutional sense” can be defined as in the minimalist definition. However, parties in the “functional sense” are those that “solve problems of collective action and of collective choice” (848). The conventional minimalist definition of political party fails to capture two main attributes of parties: horizontal coordination of ambitious politicians and vertical interest aggregation. However, the party politics literature has emphasized the horizontal coordination of ambitious politicians (Reference AldrichAldrich 1995)Footnote 1 while the vertical aggregation of collective interests has been problematized in the political sociology literature (Reference Lipset, Rokkan, Lipset and RokkanLipset and Rokkan 1967; Reference SchwartzSchwartz 1990). Vertical interest aggregation is also related to parties’ expressive function (Reference SartoriSartori 1976).
The mainstream definition of political party assigns the same analytical category (political party) to very different empirical objects. This approach does not distinguish between different kinds of political parties. Recent empirical research conflates political organizations that a thicker theoretical perspective would consider dissimilar entities that have different effects on the democratic process. As Reference SartoriSartori (1976) stresses, the minimalist definition does not suffice to adequately differentiate the various kinds of political organization. The minimalist definition of political party also lacks predictive or explanatory capacity. In this edited volume, we seek to analyze Latin America’s recent party trajectories as an empirical reference for exploring a new conceptual framework for studying political parties, one that includes diminished subtypes. Although we draw our empirical examples from Latin America, our framework is applicable to any region.
There is a recent body of research that has sought to unpack the black box of party organizations (Reference AnriaAnria 2018; Reference Bolleyer and RuthBolleyer and Ruth 2018; Reference Calvo and MurilloCalvo and Murillo 2019; Reference CyrCyr 2017; Reference Levitsky, Loxton, Van Dyck and JorgeLevitsky et al. 2016; Reference LunaLuna 2014; Reference MadridMadrid 2012; Reference Piñeiro Rodríguez and RosenblattPérez Bentancur, Piñeiro Rodríguez, and Rosenblatt 2020; Reference RosenblattRosenblatt 2018; Reference Vommaro and MorresiVommaro and Morresi 2015). Notwithstanding this renewed interest in the study of party organizations in Latin America, there remains a significant lack of theorized mechanisms and attributes of the concept of political party that connect parties to democratic representation. In her Annual Review article, Reference StokesStokes (1999) claims that it remains unsettled whether parties are good for democracy or instead a necessary evil (244). The author rightly notes that this relationship heavily depends on the definition of democracy: “Do parties reveal and aggregate voters’ preferences such that governments are responsive to citizens? Or do parties form oligopolies of competitors with interests and preferences at odds with those of voters?” (Reference StokesStokes 1999, 248–249).
The literature has identified various pitfalls party organizations encounter in various contexts and thus has highlighted the fact that many parties do not fulfill the expectation of contributing to democratic representation. However, the weak conceptualization of diminished political party subtypes lessens the analytical value of the study of parties. These problems of conceptualization neglect an important way in which political parties differ not simply in degree but in kind.Footnote 2 Moreover, the literature tends to conflate the age of a party with its degree of consolidation qua political party. An electoral vehicle might emerge as a political party and over time lose its ability to either coordinate horizontally or to vertically aggregate interests. Conversely, an electoral vehicle might gain those capacities over time. The minimalist conceptualization implies a static view that omits consideration of the changes organizations undergo over time. While the literature on democratic regimes has developed the notion of diminished subtypes of democracy (Reference Collier and LevitskyCollier and Levitsky 1997; Reference GoertzGoertz 2006), there exists no such parallel in the party politics literature. In this introductory chapter we suggest a new typology of political parties that combines the two main attributes mentioned here: horizontal coordination of ambitious politicians, and vertical aggregation to electorally mobilize collective interests and to intermediate and channel collective demands – for example, by simplifying and clarifying political preferences for the citizens.
Our work is an attempt to remedy the lack of conceptualization of diminished subtypes in the political parties’ literature. This helps to clarify analytical differences between failed parties that other authors have already described (and even explained) but have not yet conceptualized. In so doing, we revise the concept of political party in relation to its contributions to democratic accountability. On that basis, we propose a typology of political parties that includes diminished subtypes – with each type having different implications for democratic accountability – and we propose analytical strategies to empirically distinguish between them. The ultimate goal of our framework is to highlight how not all electoral vehicles – not even those with stable labels – are theoretically equivalent and thus do not contribute equally to democratic representation. While the absence of stable parties hinders democratic representation, the presence of stable electoral vehicles cannot fully guarantee the smooth operation of representation. Thus, our theoretical and conceptual contribution has concrete analytical consequences that reshape the debate concerning political parties.
Parties and Democracy: A Necessary Reassessment
What is the theoretical and empirical relationship between political parties and democracy? If democracy is simply the competition between groups of people for votes and access to government (i.e., a vision that some associate with Schumpeter’s vision of democratic competition), then defining a political party as a group of individuals who compete in elections to access office and receive a handful of votes – the minimal definition of “political party” employed in mainstream postwar political science (c.f. Reference DownsDowns 1957; Reference SartoriSartori 1976; Reference SchlesingerSchlesinger 1994) – would suffice to ensure a positive relationship between parties and democracy. This implies functions that are necessary for democracy, such as the recruitment and nomination of candidates that fosters elite-level socialization. Thus, if electoral competition, in and of itself, automatically engenders the representation of citizens’ preferences, the type of party is irrelevant. As agents in such competition, parties are automatically functional to democratic representation.Footnote 3
If, however, one proceeds from Reference DahlDahl’s (1971) definition of polyarchy, the competition for votes does not necessarily lead to representation of citizens’ preferences. Dahl’s perspective requires that, for citizens to have equal influence in politics, certain conditions and guarantees must exist; competition among groups does not suffice for there to be a positive relationship between parties and democracy. Not all electoral vehicles that compete in elections are functional to interest representation. The types of electoral vehicles that compete in elections determine how democracy works. A party system can exist without representing or distorting citizens’ preferences (Reference GilensGilens 2012). Only under very specific (and unrealistic) conditions, as in the Downsian perfect information competition model, can it be the case that any group that competes for votes represents citizens’ preferences. Yet, as Downs stressed, democracy does not function in these conditions and representation does not automatically derive from the existence of competition. In practice, in different democracies, electoral vehicles might or might not function as channels for citizen representation. Thus, according to Dahl’s logic, some electoral vehicles facilitate democratic representation, while other vehicles are less sensitive to citizens’ demands and interests and so channel them less effectively. This complex relationship between electoral vehicles and citizen representation has been studied extensively in the party politics literature (as will be discussed).
Democratic representation in modern societies can be analyzed as a principal-agent relationship (Reference MichelsMichels 1999 (1911)). Different types of electoral vehicles structure the principal-agent relationship differently, with some being unable to structure it at all, given their detachment from their principals. The latter occurs in contexts where citizens can vote for a given electoral vehicle without having the ability to monitor the vehicle’s actions in the aftermath. The inability to hold electoral vehicles accountable can derive from exogenous factors; that is, it may be contingent on socioeconomic conditions – poverty, inequality, or economic crises – or institutional settings, such as more autocratic contexts (Reference Kitschelt and WilkinsonKitschelt and Wilkinson 2007; Reference LunaLuna 2014; Reference Taylor-RobinsonTaylor-Robinson 2010). Here, however, we are interested in analyzing whether party organizations channel the principals’ preferences. We claim that there are endogenous constraints that relate to the specific characteristics of each political party.
The literature has systematically argued that there exists a much more nuanced relationship between existing parties (and party systems) and democratic representation (Reference HickenHicken 2009; Reference Kitschelt and WilkinsonKitschelt and Wilkinson 2007; Reference Lawson and MerklLawson and Merkl 1988; Reference LevitskyLevitsky 2003; Reference LunaLuna 2014; Reference Luna and ZechmeisterLuna and Zechmeister 2005; Reference MainwaringMainwaring 2018; Reference Mainwaring and ScullyMainwaring and Scully 1995; Reference Piñeiro Rodríguez and RosenblattPiñeiro Rodríguez and Rosenblatt 2020; Roberts 2014b). The party politics literature has extensively considered the exogenous conditions that determine levels of representation. Developing societies, where the structural conditions for channeling citizens’ preferences are unfavorable, have a wide variety of electoral vehicles with differing capacities to channel citizens’ preferences (Reference BartoliniBartolini 2000; Reference KitscheltKitschelt 1994; Reference Kitschelt, Hawkins, Luna, Rosas and ZechmeisterKitschelt et al. 2010; Reference LunaLuna 2014; Reference Mainwaring and ZocoMainwaring and Zoco 2007; Reference Samuels and ShugartSamuels and Shugart 2010; Reference StollStoll 2013; Reference Taylor-RobinsonTaylor-Robinson 2010). Yet, even developed societies, with more favorable exogenous conditions, have also witnessed the emergence of various types of political organizations that seek to perform the political representation function, and not all succeed in doing so.
The literature on party politics in developing countries in general, and in Latin America in particular, has identified various kinds of agents that compete in elections but do not contribute to democratic representation. However, this literature has not provided a conceptual discussion that theorizes the existence of diminished political party subtypes (with some exceptions, e.g., Reference MustilloMustillo 2007). While there exists abundant empirical evidence concerning the various failures of different party organizations in modern democracies and several theoretical arguments regarding the causes and effects of such failings, there remains a lacuna in the conceptualization of the type of parties that function as channels of democratic representation. This lack of theoretical debate concerning diminished party subtypes derives from the minimalist definition of political party. There has been little discussion in the literature as to whether this minimalist definition is useful for differentiating the various ways an agent can compete for power in a democratic process. While the minimalist definition is efficacious for encompassing different electoral vehicles, it obscures the debate about which vehicles contribute to the functioning of democracy. This is especially critical because the minimalist definition of political party works better in dialogue with a definition of democracy that privileges electoral competition as the main attribute of the regime, but it does not fit a more demanding perspective, such as Dahl’s. When electoral competition does not suffice as a defining attribute of democracy, the minimalist definition of political party makes it difficult to articulate a clear-cut relationship between parties and democracy. The minimalist definition grants the label “party” to electoral vehicles that compete in elections but do not hold the status of party.
In fact, for much of the twentieth century, the relationship in Latin America between parties and democracy was problematized in terms of the acceptance of electoral competition: the movement-parties and the “illiberal” parties did not support democracy. However, in the twenty-first century, parties accept democratic competition, but they do a poor job of fulfilling their representation function. In several countries – for example, Bolivia, Peru, Venezuela, and Argentina – many of the traditional parties have been weakened or have disappeared. Their social bases were transformed or became more heterogeneous (e.g., weakening of the industrial working class, crisis of the farming sector, emergence of new middle classes and pauperization of others, emergence and consolidation of an informal sectors). New electoral vehicles emerged in turbulent times around electorally successful leaders (e.g., Alberto Fujimori in Peru, Mauricio Macri in Argentina, or Hugo Chávez in Venezuela), who, in some cases, exited from traditional parties (e.g., Álvaro Uribe in Colombia).
Confronting that emerging reality, several scholars turned their attention to causal factors and theories about party building, failure, and success including Reference AnriaAnria (2018); Reference CyrCyr (2017); Reference HunterHunter (2010); Reference LevitskyLevitsky (2001; Reference Levitsky2003); Reference Levitsky, Loxton, Van Dyck and JorgeLevitsky et al. (2016); Reference LupuLupu (2016); Reference MadridMadrid (2012); Reference TavitsTavits (2005; Reference Tavits2008; Reference Tavits2013); Reference SamuelsSamuels (2004; Reference Samuels2006); and Reference Vommaro and MorresiVommaro and Morresi (2015). However, the resurgence of party politics research in the last decade has not been adequately matched by a conceptual reanalysis of the empirical objects that we label as political parties. To address this gap in the literature, we reanalyze the concept of political party and its diminished subtypes, by adding or subtracting attributes to its definition. Specifically, we propose to distinguish between diminished subtypes by adding to the current mainstream minimalist definition two dimensions: horizontal coordination and vertical aggregation.
Conceptualization, Operationalization, and Measurement
Following Reference GoertzGoertz (2006), our conceptual analysis assumes the existence of specific links or associations between the existence of parties and democracy. Electoral vehicles that exhibit both dimensions (horizontal coordination and vertical aggregation) positively influence democratic representation. Political organizations that exhibit high levels of both dimensions reduce transaction and informational costs for citizens, who are the principals in the representation relationship.
An electoral vehicle is an association of candidates, that is, office-seekers, whose members compete in elections under the same label. Although the coalition seeks to win office, not all electoral vehicles fulfill the two basic functions necessary for a political party to be an effective means of democratic representation. A political party is, then, an electoral vehicle subtype, a more intense and less extended concept (Reference SartoriSartori 1970): it coordinates the activities of ambitious politicians (during campaigns and between elections) and vertically aggregates collective interests. “Electoral vehicle” is a more general concept than “political party,” which occupies a lower level of abstraction (Reference SartoriSartori 1970). More specifically, political parties want to access office and promote policies (Reference StromStrom 1990). Parties seek to win state power and impose an allocation of resources through policies and state institutions. This is achieved by crafting social coalitions, which involves coordination during campaigns and between elections.
Parties can accomplish the two functions in very different ways and with very different organizational forms (Reference Gunther and DiamondGunther and Diamond 2003). The literature has extensively documented different types of parties in different historical and geographical settings (i.e., with an evolutionary logic), including cadre and mass-based parties (Reference DuvergerDuverger 1954), catch-all parties (Reference Kirchheimer, LaPalombara and WeinerKirchheimer 1966), professional-electoral parties, and cartel parties (Reference Katz and MairKatz and Mair 1995), among others. As opposed to these typologies, our conceptualization is independent of organizational form and assumes that different organizational arrangements can fulfill both conditions. Moreover, our framework does not imply that the linkages between the party and its constituency must necessarily be programmatic. In this vein, our idea of interest aggregation is broad. Because clientelistic politics can represent groups, it is possible to aggregate collective interests in a clientelistic manner. The horizontal coordination can be based on party members’ adherence to shared rules or on a personalistic leadership. In this regard, very different parties, at different periods, such as the Radical Party in the early twentieth century, and the Unión Demócrata Independiente (Independent Democratic Union, UDI) in Chile, the Partido dos Trabalhadores (Workers’ Party, PT) in Brazil, and the Partido Conservador (Conservative Party, PC) in Colombia (see Wills-Otero, Ortega, and Sarmiento this volume) throughout the twentieth century (until 1991), differ in their organizational structure and in their linkages with voters, though all accomplished the two defining functions.
Our concept of political party comprises five levels. The basic level constitutes the concept of political party itself. The secondary level introduces its main attributes. We identify two necessary and sufficient conditions that qualify an electoral vehicle as a political party in terms of democratic representation: the horizontal coordination of ambitious politicians and vertical interest aggregation. Figure 1.1 presents the structure of the concept of political party and its attributes (indicators will be presented in subsequent figures). Horizontal coordination denotes the role of parties in facilitating the coordination of ambitious politicians during campaigns and between electoral cycles. Vertical interest aggregation denotes the role of parties in the electoral mobilization and intermediation (or channeling) of collective interests and demands between elections. There is low substitutability between these two main attributes. They are separately necessary and are jointly sufficient conditions; thus, they interact, and both need to be present to warrant labeling a given electoral vehicle as a political party.
These two dimensions (horizontal coordination and vertical interest aggregation) are functional to the idea of democratic representation. Horizontal coordination implies that political parties solve collective action problems of ambitious politicians, and this benefits democratic representation by helping stabilize electoral vehicles. Many electoral vehicles can support horizontal coordination between politicians; yet this function can be achieved without considering any societal preferences. This occurs, for example, in political systems where the competition between parties is stable but does not incorporate citizen preferences and thus alienates important portions of the electorate, as Reference Luna and AltmanLuna and Altman (2011) show for the Chilean case. Therefore, electoral vehicles should also perform vertical interest aggregation to function as a channel for democratic representation. Conversely, electoral vehicles that aggregate collective interests but do not support horizontal coordination tend to be fragmented, undisciplined, and unstable organizations.
At the third level, following Reference AldrichAldrich (1995), we stipulate that horizontal coordination implies coordination during electoral campaigns and between elections (i.e., in Congress and in office). During campaigns, a political party is an electoral vehicle capable of monopolizing the candidate selection process, monopolizing the electoral coordination strategy (i.e., deciding the number of candidates that will compete in each district), and providing a common electoral label. These three capabilities are necessary and sufficient attributes for coordination during elections and entail the existence of a minimum common platform. In political parties, thus, candidates must be personally or collectively validated. These attributes enable parties to propose a uniform and coherent electoral offer. This coordination can be achieved in very different ways; for example, the candidate selection process can be centralized or decentralized, and can be carried out through open primaries or by a commission (Reference Hazan and RahatHazan and Rahat 2010; Reference Rahat and HazanRahat and Hazan 2001; Reference Siavelis and MorgensternSiavelis and Morgenstern 2008a). The crucial point is that a political party has the ability to coordinate action to avoid electoral losses. Between elections, a political party coordinates activity in Congress and in local governments. A political party establishes formal and informal obstacles to prevent its leaders from proposing contradictory public policies at different levels of government, and generates incentives to favor a certain amount of discipline among their legislators regarding whether to support or oppose given policies. Coordination both during and between elections is necessary and sufficient; that is, there is low substitutability between the two instances of horizontal coordination.
Also at the third level, the electoral mobilization of collective interests and the intermediation and channeling of collective demands are the two attributes that compose vertical interest aggregation. Both are necessary and sufficient attributes of the vertical dimension and, thus, there is low substitutability between them. To serve as agents for democratic representation, political parties need to aggregate preferences during campaigns (by mobilizing collective interests) and between elections (providing a channel for articulating collective interests). Parties must be valid options for citizens and collective actors (classes, movements, social groups) in democratic elections and must channel citizens’ and collective actors’ demands between elections. Voters must know that by voting for a particular label they are voting for a certain type of bias in public policies and especially in distributive policies. This dimension highlights the crucial role of vertical accountability in contemporary democracies (Reference AdamsAdams 2001; Reference DownsDowns 1957; Reference Przeworski, Stokes and ManinPrzeworski, Stokes, and Manin 1999) and both attributes, the electoral mobilization of collective interests and the intermediation and channeling of collective demands, are needed to promote what Reference DahlDahl (1971) considered an essential attribute of democracy: “the continuing responsiveness of the government to the preferences of its citizens” (1).
Figure 1.2 presents the complete conceptual tree for one of the two secondary-level attributes of a political party: horizontal coordination. It shows the two necessary and sufficient second-level attributes (coordination must occur both in elections and between elections) and it introduces a set of indicators. The figure also specifies the relationship between dimensions (or attributes) at each level and their indicators. During electoral campaigns, a party must monopolize the process of candidate selection and the electoral coordination strategy and candidates must use the common party label. We introduce two indicators, each necessary and both jointly sufficient, to determine the presence of the party’s monopoly control of the candidate selection process: (1) a party authorizes candidate nomination at all levels and (2) prospective candidates accept nomination processes and the results of those processes. Parties must enforce horizontal coordination among ambitious politicians throughout a candidate selection process. This implies that the party has the power to define who can run under the party’s label. Also, all prospective candidates should respect the results of the candidate selection process; for example, there should be no defections by those who were not selected. This is not related to how open or closed the rules are.
The indicator of the party’s monopoly control of the electoral coordination strategy is that the party considers the restrictions of the electoral system and enforces electoral coordination among candidates. More specifically, the party must control the number of candidates to avoid a situation that might affect the party candidates’ joint probability of accessing office. On some occasions, candidates have more influence in the selection processes than does the party. When this happens, candidates might end up failing to coordinate and, thus, may hinder the party’s electoral performance.
Finally, the indicators for the use of a common label are: (1) candidates use the same campaign logo; or, (2) candidates use party emblems or colors; or, (3) candidates use the party’s propaganda (i.e., campaign literature). In this case, there is substitutability between the different indicators as each is functionally equivalent to the other (i.e., each one captures different ways to observe the use of a common label).
Between elections, a party must coordinate in Congress and in the different local-level governments, including in local-level legislative bodies. The indicator for horizontal coordination in Congress is the observation of significant party discipline. The indicator for coordination in local-level governments is the observation of a general consistency of public policies across different units; that is, in general terms, a party must have a similar policy orientation throughout the country and while voting in Congress. This coordination distinguishes parties from electoral vehicles that only coordinate different autonomous agents for the election (national or local). An environmentalist party, for example, should consistently promote a “green” agenda in all the governmental institutions in which it has representatives. Similarly, labor-based parties oppose deregulatory labor reforms even in times of policy convergence promoting economic liberalization and state retrenchment (Reference MurilloMurillo 2001).
Figure 1.3 presents the complete operationalization of vertical interest aggregation. The figure shows the two necessary and sufficient attributes of vertical interest aggregation: a party electorally mobilizes collective interests and it intermediates and channels collective demands. A party mobilizes collective interests when its electoral platform includes general demands of one or several of the party’s constituencies or when the party has a stable core constituency. A party might not have developed a core constituency (or it might have lost it), but its electoral platform has unequivocal references to a clear constituency. These parties have a platform that is oriented toward formal workers but many times those workers do not vote for these parties. The family resemblance structure in this case (i.e., complete substitutability between the indicators) helps to capture these situations.
The intermediation and channeling of collective demands has two indicators: the existence of formal or informal ties with civil society organizations and the observation that party decisions are constrained by its core constituency. Both are necessary and sufficient, that is, there is low substitutability between them. Also, the attribute “existence of formal or informal ties with civil society organizations” itself has three indicators: the existence of dual membership (elites or grassroots), the existence of formal ties between the party and civil society organizations, or the existence of informal ties between the two. We allow complete substitutability between the three indicators, because each represents a different path to the same result. For example, the Movimiento al Socialismo (Movement toward Socialism, MAS) and the Frente Amplio (Broad Front, FA) both have strong ties with social movements. However, the two parties build their ties in dissimilar ways. Social movements are organically part of the MAS (Reference AnriaAnria 2018 and Anria this volume) whereas the FA’s ties with social movements are informal and FA members often have dual membership in both the party and social movements (Reference Piñeiro Rodríguez and RosenblattPérez Bentancur, Piñeiro Rodríguez, and Rosenblatt 2020 and Pérez Bentancur, Piñeiro Rodríguez, and Rosenblatt this volume).Footnote 4
To measure each indicator, we propose using a five-point scale where values on the scale indicate the degree to which a particular condition is satisfied, with the scale values 1–5 corresponding to 0 percent, 25 percent, 50 percent, 75 percent, and 100 percent fulfillment of a given condition, respectively. For example, when a party has rules for nominating candidates, but half of the time prospective candidates do not comply with the rules, the case should receive a score of “3” on the indicator “Prospective candidates accept nomination processes and results,” indicating 50 percent fulfillment of the condition. If there is no rule at all and candidates can nominate themselves, the case should receive a score of “1” on this indicator, corresponding to 0 percent fulfillment of the condition. Each indicator is normalized on a scale from 0 to 1. The overall index is computed by the aggregation rule reflecting the conceptual structure at each level. The overall index varies from 0 to 1, where “0” signifies that the case lacks any and all characteristics of a political party and “1” signifies that it exhibits all of them.
Consistent with our conceptualization of political party, we aggregated the component indices as follows. When there is complete substitutability between the indicators of an attribute, we used the maximum value. For example, the attribute “Existence of formal or informal ties with civil society organizations” has three indicators that we consider functionally equivalent measures of the attribute observed in different contexts, that is, each indicator captures a different way to fulfill the attribute (see Figure 1.3). Therefore, in a given case, the degree of fulfillment of the attribute will be determined by the highest value of the three indicators. In cases where the relationship between indicators or attributes, at different levels, is one of necessity and sufficiency, we use the geometric mean.Footnote 5 This aggregation rule allows for low substitutability. A low level of one indicator is partially compensated for by a high level of another indicator. Nonetheless, it emphasizes the necessary and sufficient conceptual structure and implies lower levels of compensation than does using the average or the maximum (Reference GoertzGoertz 2006). Using the geometric mean mitigates the loss of additional information associated with using the minimum, and thus captures the multi-dimensionality of the concept. For example, vertical interest aggregation has two dimensions: “Electorally mobilizes collective interests” and “Intermediation and channeling of collective demands.” If a case has a score of 2 on the former dimension, representing a 0.25 degree of fulfillment, and a score of 4 on the latter dimension, representing a 0.75 degree of fulfillment, the case will have an aggregate score of 0.43Footnote 6 for vertical interest aggregation.
We asked the authors of each case study in this edited volume to categorize their cases according to our conceptual scheme.Footnote 7 In the online appendix we include the codebook and the value of each indicator for each case.Footnote 8 We considered the following cases: Propuesta Republicana, (Republican Proposal, PRO, Argentina), Partido Justicialista (Justicialist Party, Argentina), MAS (Bolivia), the Partido por la Democracia (Party for Democracy, PPD, Chile), the Partido Liberal (Liberal Party, PL, Colombia), PC (Colombia), Partido Acción Ciudadana (Citizen Action Party, PAC, Costa Rica), Liberación Nacional (National Liberation Party, PLN, Costa Rica), Movimiento Alianza PAIS (PAIS Alliance Movement, AP, Ecuador), Unidad Nacional de la Esperanza (National Unity of Hope, UNE, Guatemala), Partido de la Revolución Democrática (Party of the Democratic Revolution, PRD, Mexico), Movimiento Regeneración Nacional (National Regeneration Movement, MORENA, Mexico), Partido Colorado (Colorado Party, Paraguay), Partido Liberal Radical Auténtico (Authentic Radical Liberal Party, PLRA, Paraguay), Fuerza Popular (Popular Strength, FP, Peru), the FA (Uruguay), Primero Justicia (Justice First, PJ, Venezuela) and Voluntad Popular (Popular Will, VP, Venezuela).Footnote 9
Table 1.1 shows each party’s score on the two dimensions of the political party concept as well as on the overall party index. The scores vary across almost the entire range of the measure, showing that it is sensitive to differences between cases. Overall, the cases exhibit higher ratings on the horizontal coordination dimension than on the vertical interest aggregation dimension. The former is an easier property to achieve because a party’s basic raison d’etre is to solve collective action problems for politicians. However, the different cases show variance in both dimensions and this variance is independent. These results show that each dimension captures different aspects of the concept and are not redundant.
|Party||Horizontal coordination||Vertical interest aggregation||Party index|
|Colorado Party (Paraguay)||0.45||0.87||0.62|
|PAC (Costa Rica)||0.59||0.57||0.58|
|Justicialist Party (Argentina)||0.35||0.93||0.57|
|PLN (Costa Rica)||0.87||0.68||0.77|
Typology of Political Parties and Diminished Subtypes
To capture the existence of political organizations that lack one or more of the necessary dimensions in our conception of political party, we develop a typology of electoral vehicles: political parties and diminished subtypes. While the literature has analyzed the effects of the existence of independent candidates, flash parties, etc., it has been relatively silent on diminished subtypes, in which one of the two attributes of the political party concept is absent (Reference Collier and LevitskyCollier and Levitsky 1997; Reference GoertzGoertz 2006). Thus, these diminished subtypes are not subsets of a more general category of political party. On the contrary, these are theoretically possible variant forms of electoral vehicle, that is, political party diminished subtypes. Diminished subtypes are neither more nor less abstract than the concept of political party (Reference GoertzGoertz 2006; Reference SartoriSartori 1970). The absence of one or more attributes does not imply greater abstraction or greater extension; rather, it indicates a diminished subtype. Thus, the different types in our taxonomy occupy the same level of abstraction, but diminished subtypes are cases that lack one or more of the attributes of a political party.
We identify the various possible electoral vehicles to understand the different types of political organizations and groups that compete in elections in contemporary democracies and their effects on democratic representation. If we treat the two attributes identified in our definition of political parties as binary variables that can be either present or absent, we create a 2 × 2 conceptual space, which yields four different types of political organization, as shown in Figure 1.4.
|Vertical interest aggregation||Yes||Uncoordinated party||Political party|
In our framework, the political party denotes an electoral vehicle that accomplishes two essential functions: it coordinates ambitious politicians and aggregates collective interests vertically. This category encompasses long-standing parties such as the PLN in Costa Rica and the Partido Acción Nacional (National Action Party, PAN) in México; more recently established parties such as the FA in Uruguay, the PT in Brazil, the PRD in Mexico, and the UDI in Chile; and new parties like the PRO in Argentina, the MAS in Bolivia, and VP in Venezuela. These examples illustrate that the two attributes, horizontal coordination and vertical interest aggregation, can be fulfilled with different organizational structures. The PT and the FA resemble mass organic parties, while the PAN, the PRO, and the UDI resemble cadre and professional electoral parties. Also, the age of a party, an indicator commonly used to assess a party’s stability, does not define its capacity to fulfill the functions associated with a political party, as we define it. For example, a political organization can be vibrant at the time of its origin, showing robust horizontal coordination and vertical aggregation of interests, such as the PRO in Argentina, but lose one or both of those attributes over time as a consequence of endogenous or exogenous crises, such as the Partido Socialista de Chile (Chilean Socialist Party, PSCh). Studies of adaptation and party collapse provide accounts of this phenomenon (Reference LevitskyLevitsky 2003; Reference LupuLupu 2016), while recent works have analyzed the factors that determine political organizations’ degree of vibrancy over time (Reference RosenblattRosenblatt 2018).
A political organization can achieve harmonious coordination between its elites (both during campaigns and between elections), without having a consistent capacity to articulate collective interests. We designate this electoral vehicle an unrooted party. This kind of electoral vehicle can contribute to the stability of democratic institutions, but they are weak in terms of channeling the electoral and congressional representation of social groups/interests. In Latin America, there are cases of established political groups that have a high capacity for horizontal coordination among their elites, but have substantially lost (or never developed) stable linkages with any social base. This type of vehicle generally appeals to the “citizen” and espouses a negative vision regarding the representation of different social sectors in the political arena. Usually, they are centrist vehicles but not all centrist vehicles lack a constituency. The clearest example is the Partido Demócrata Cristiano (Christian Democratic Party, DC) in Chile; at the time of its origin, it was a centrist party with a clear constituency.
Unrooted party elites coordinate during campaigns and between elections. These vehicles can coordinate between elections because the agreements between individual leaders are also kept in the parliamentary arena, or because one of these leaders stands as primus inter pares (e.g., by being elected president, prime minister, mayor, or because of the leader’s electoral appeal), and manages to retain coordination mechanisms for incumbents based on the distribution of selective incentives and/or collective incentives associated with the persistence of the vehicle. This type of vehicle fails to build effective channels for aggregating collective interests. These are usually traditional electoral labels, such as the Partido do Movimento Democrático Brasileiro (Party of the Brazilian Democratic Movement, PMDB) in Brazil, activated during election season. However, the reference to a unified electoral list reflects an alliance between individual ambitious political leaders rather than the existence of a political party.
There are electoral vehicles that develop persistent ties with loyal constituencies but lack horizontal coordination mechanisms; they usually lack congressional discipline and they have problems coordinating during elections. Sometimes this lack of coordination implies uncoordinated electoral strategies between different leaders. We label this diminished subtype an uncoordinated party. The Peronists in Argentina, in the absence of strong national leaders, lack congressional discipline and are unable to coordinate in the electoral arena. However, as Reference LevitskyLevitsky (2003) shows, this diminished subtype has informal negotiation channels with mobilized groups, such as trade unions. Also, this type of diminished party subtype is more common in organizations built or developed by regional leaders, linked to local interests, who have difficulty establishing common strategies outside the electoral arena, as happens with traditional parties in Colombia (Reference Wills-OteroWills-Otero 2015).
Ambitious politicians can operate without coordinating political activity, running for office based on enabling electoral rules and/or their prestige or popularity (Reference Levitsky, Zavaleta, Levitsky, Loxton, Brandon and DominguezLevitsky and Zavaleta 2016; Reference ZavaletaZavaleta 2014). This diminished subtype tends to proliferate in the context of a party system crisis, when the cost of entry to the competition is low, as occurred in Argentina during the financial and economic collapse of 2000 and 2001, in Ecuador during the emergence of Rafael Correa in 2006, or in Peru in 1990 when Fujimori won the election with Cambio 90, his electoral vehicle (Reference CyrCyr 2017; Reference Dietz and MyersDietz and Myers 2007; Reference Levitsky, Zavaleta, Levitsky, Loxton, Brandon and DominguezLevitsky and Zavaleta 2016; Reference SeawrightSeawright 2012; Reference ZavaletaZavaleta 2014). This subtype also proliferates in party systems where traditional parties have declined, opening electoral competition to individuals who have access to valuable campaign resources (money, fame, prestige) that render them competitive. In federal systems, and in systems with strong regional identities, this type of electoral vehicle often exists at the subnational level. To a certain extent, the subtype Independents represents the extreme case of stretching the party concept that we want our typology to amend.
Unpacking the different types of electoral vehicles better equips researchers to assess electoral vehicles’ effects on democratic representation. In a recent edited volume, Reference Levitsky, Loxton, Van Dyck and JorgeLevitsky et al. (2016) identify different cases of successful party building. The authors classify successful party building (i.e., parties that “take root”) simply by considering the stability of a party label in successive elections: “We score party-building as successful when a new party wins at least 10 percent of the vote in five or more consecutive national legislative elections” (Reference Levitsky, Loxton, Van Dyck and JorgeLevitsky et al. 2016, 8). Temporal bounds, while easy to measure, neglect to consider how or whether party organizations accomplish both essential functions described previously. In our conceptualization, however, the Renovación Nacional (National Renewal, RN) and the PPD of Chile do not constitute true parties but are instead diminished subtypes. In the former, there is no coordination of activity during the elections, while the “party” represents defined interests – business and rural sectors. It is thus an uncoordinated party within our conceptual framework. The latter (PPD) is a coalition of independent politicians who struggle to accomplish either of the two functions (see Piñeiro Rodríguez, Rosenblatt, and Toro Maureira this volume). Conversely, new parties such as the MAS in Bolivia (see Anria this volume) and the PRO in Argentina (see Vommaro this volume) are, indeed, successful cases of party building. In both cases, horizontal coordination mechanisms are observed and there are vertical representation channels – with social movements or business sectors – that have been robust and persistent over time. Both new political party organizations managed to incorporate collective demands. Thus, for example, for the first time in history, the Bolivian peasantry managed to build its own party (Reference AnriaAnria 2018), while a center-right pro-market party managed to compete for power in Argentina (Reference Vommaro and MorresiVommaro and Morresi 2015).
Figure 1.5 presents the observed values for the analyzed cases on each of the two dimensions (horizontal coordination and vertical interest aggregation) of the party index. We divide the panel to illustrate the classification of cases into each subtype. The classification follows the description presented here. In the upper right cell of the table, we find parties such as the FA, VP, MAS, and PRO, among others. These parties perform both functions, though to varying degrees. For example, while the MAS and the PRD are rated more highly on vertical interest aggregation, the PRO and the PLN are rated more highly on horizontal coordination. In the bottom left cell, we find Independents, such as the FP and the UNE. The Chilean PPD and the Colombian PL are borderline cases that have characteristics of both unrooted parties and independents. The Colombian PC, the Argentinean Justicialist Party, and the Paraguayan Colorado Party most closely resemble the uncoordinated party type. Finally, the Venezuelan PJ is a typical example of an unrooted party. It exhibits high levels of horizontal coordination but lacks vertical interest aggregation. Finally, the distribution of our cases seems to indicate that organizations rarely exhibit the capacity to vertically aggregate social interests without also exhibiting the capacity for horizontal coordination.
Given our emphasis on the functions parties should fulfill to satisfy horizontal coordination and vertical aggregation, our definition might be seen as a functionalist one. Yet, our argument is not functionalist because electoral vehicles can and often do fail to fulfill one or both functions. Our conceptualization is, thus, adequate to make such normatively and substantively consequential variance visible to those interested in exploring it empirically. Indeed, we identify diminished subtypes precisely in order to characterize cases that fail to fulfill one or both functions. Diminished subtypes, which are empirically pervasive in contemporary democracies, exist without fulfilling the functions our conceptualization assigns to political parties (i.e., diminished subtypes are not different forms of a political party, and they exist despite not fulfilling the functions we use to demarcate diminished subtypes from political parties). When identifying diminished subtypes, we are not moving up or down a conceptual ladder of abstraction; rather, we identify a positive pole (political party) and a negative pole (an electoral vehicle that is not a political party). Thus, for example, the absence of the two necessary dimensions – horizontal coordination and vertical aggregation – in the “independents” diminished subtype does not render “independents” a more abstract notion than that of political party. By the same token, political parties are not conceptualized as a subset of a more abstract notion of independents.
This edited volume includes chapters describing seventeen parties in twelve countries in Latin America. The case studies describe and analyze how different electoral vehicles perform or fail to perform horizontal coordination and vertical interest aggregation. The empirical chapters also offer a variety of cases to illustrate political parties and diminished subtypes. The volume seeks to encompass variation in the two theoretically relevant dimensions (vertical interest aggregation and horizontal coordination). The case selection also aims to include regional, organizational, and ideological variation. Additionally, the chapters provide an overview of the evolution of each case over time. This overview is particularly instructive because it provides a dynamic perspective showing that there is no single inevitable developmental trajectory; parties that persist over time do not necessarily improve their ability to perform horizontal coordination and vertical interest aggregation, nor does this ability necessarily deteriorate over time.
In Chapter 2, Pérez Bentancur, Piñeiro Rodríguez, and Rosenblatt analyze the case of the FA in Uruguay as an unusual organization. Since its foundation, the FA has exhibited a dual structure: the coalition (manifested in the factions of the party) and the movement (comprising a common grassroots structure of Base Committees). The FA fulfills both essential functions required to qualify as a political party: vertical interest aggregation and horizontal coordination. The coalition structure of the FA fulfills the criterion of horizontal coordination, while the grassroots activist structure accomplishes vertical interest aggregation and, in crucial decisions, promotes and ensures horizontal coordination. More critically, the coalition and the grassroots structure influence the most important policy decisions of the party’s parliamentary caucus and, when the FA was in government (2005–20), the decisions of the Executive. In terms of vertical interest aggregation, the FA has developed strong and informal ties with social actors, especially with labor unions, based on leaders’ and grassroots activists’ dual membership.
Chapter 3 analyzes the case of the Argentinean PRO. Vommaro’s analysis highlights the fact that the PRO has been able to perform both functions in the districts where the party was born (mainly in the Ciudad de Buenos Aires) but has had greater difficulty fulfilling them in the rest of the country. This shows that parties can vary across districts in how well they perform the defining attributes. The PRO established horizontal coordination through a division-of-labor mechanism. PRO leaders maintain control of the party’s electoral strategy and candidate selection and allow its allies from traditional parties to capture and distribute resources to their clients. This division of labor also implies that PRO leaders rule and the allied parties mobilize voters, especially from popular sectors and from districts in the countryside. In exchange, allied party leaders receive selective incentives, such as important slots in the party lists, access to public resources, and positions in the public administration. These two groups bring different constituencies to the party. PRO leaders have strong linkages with the upper-middle- and upper-class sectors of Argentinean society. Middle- and working-class sectors are incorporated by politicians from traditional parties who joined the PRO. This produces a logic of segmented representation (Reference LunaLuna 2014).
Chapter 4 develops the case of the MAS. Anria’s chapter shows how a loosely organized structure can still achieve horizontal coordination and vertical interest aggregation. This is based on the interaction between the strong leadership of Evo Morales and the social movements and social organizations that formally integrate the party. As Anria stresses, “[The MAS] operates as a hybrid organization that combines top-down leadership by a dominant personality, weak bureaucratic development, and the bottom-up power of autonomous social mobilization” (Anria this volume). As in the case of the PRO, the MAS is heterogeneous throughout the territory. In some districts, social movements are stronger and have more influence over the party’s decision-making and candidate selection, while in others the party’s structure and its leadership prevail. However, in general, party leaders cannot impose their positions and must negotiate with social actors that are part of the MAS. This “contentious bargaining game” (Anria this volume) occurs through informal channels. This interaction is different from that observed in the case of the FA. In contrast to the FA, where social organizations have informal ties with the party but the party’s decision-making is grounded in formal rules, social organizations are formally incorporated in the MAS but the party’s policy agenda is negotiated through informal channels.
In Chapter 5, Combes analyzes the cases of the PRD and MORENA. This chapter highlights the importance of personalism as a factor that facilitates not only horizontal coordination (as in the case of the AP or the FP) but also vertical interest aggregation (as in the case of the MAS). Candidate selection is highly connected to the complex interaction that both parties have with civil society and social movements. However, the PRD and MORENA did not institutionalize their ties with social movements because they rejected the corporatist model historically observed in the PRI. In both the PRD and MORENA, vertical interest aggregation is based on informal ties linked to the recruitment of activists and candidates from social movements. In the case of the PRD, the party recruited leaders of the mobilizations against neoliberal reforms. Given the territorial concentration of the protests, this strategy limited the territorial expansion of the party. In the case of MORENA, there was an explicit strategy to control recruitment from the center.
In Chapter 6, Alfaro-Redondo and Gómez-Campos present the cases of the PLN and the PAC in Costa Rica, both classified as parties, though the PAC less clearly so. The PLN is a traditional party that has been experiencing a process of organizational decay, especially in its ability to perform the vertical interest aggregation function. Historically, the PLN has had a developed structure and deep ties with specific constituencies, though this has been eroding since the 1990s. The PAC is a new party that also performs both functions but has so far struggled to reproduce the vertical interest aggregation function beyond its core constituency in the middle- and upper-middle-class sectors of the country’s capital, San José.
Chapter 7 analyzes the Paraguayan Colorado Party and PLRA parties. Abente Brun’s chapter shows how vertical interest aggregation is not necessarily programmatic. In the case of the Colorado Party, vertical interest aggregation is related to the satisfaction of particularistic demands for specific constituencies. The electoral machine of the Colorado Party was consolidated during Stroessner’s authoritarian regime and operated as a tool for vertical interest aggregation. This capacity was reduced and had to adapt as the country underwent a transition to democracy. In contrast, the PLRA had operational difficulties during Stroessner’s regime that limited its organizational development. Since the 1992 constitutional reform, horizontal coordination has been weakened. The power to coordinate horizontally now lies with party factions rather than with the national party directorates. The weakening of the capacity to perform horizontal coordination was aggravated by the introduction of an open list electoral system in the 2019 reform. The Paraguayan cases highlight the role that political regimes and institutional rules play in the fulfillment of both functions.
In Chapter 8, Wills-Otero, Ortega, and Sarmiento, show the secular process of erosion of the Colombian traditional parties, the PC and the PL. These parties have found it increasingly difficult to achieve horizontal coordination and vertical interest aggregation. The extensive history of both parties enables a long-term analysis of their evolution, and reinforces the idea advanced in this framework that a party’s ability to fulfill one or both functions can change over time (this is also analyzed in the Costa Rica and Paraguay chapters). The chapter also shows the positive and negative impacts that various changes in electoral rules have had on the parties’ capacity to perform horizontal coordination, as in Paraguay (see Abente Brun this volume). The electoral changes introduced in the 1991 constitutional reform facilitated the personalization of politics, while the electoral reform of 2003 and the 2005 Law that governs political blocs counterbalanced this tendency. The authors claim that the PL and the PC, in contrast to the Colorado Party in Paraguay, have difficulties promoting a programmatic aggregation of interests because of the combination of the pervasive role of clientelism and the power of departmental or local leaders.
Chapter 9 reviews the cases of the PJ and the VP in Venezuela. Cyr’s analysis introduces the challenges of party building in contexts of democratic erosion and high polarization based on chavismo/anti-chavismo logic. This context prevents opposition parties from proactively building ties with different social sectors based on programmatic stances. However, this context does not have the same effect in Cyr’s two cases. The PJ developed a limited capacity to vertically aggregate interests. By contrast, the VP established bottom-up channels to incorporate interests from below. Paradoxically, the regime dynamics, that is, polarization and increasing authoritarian tendencies, hinder vertical interest aggregation but, at the same time, facilitate horizontal coordination.
In Chapter 10, Conaghan analyzes the case of the AP in Ecuador, which achieved something rare among Ecuadorian electoral vehicles: it ran candidates in every district. However, the AP did not become a political party. Since its inception, the AP was designed to concentrate power around its leader, Rafael Correa. Conaghan’s analysis stresses that the AP performs horizontal coordination, yet this coordination was always imposed from above; compliance with the party’s directives was a condition of obtaining access to governmental positions. Vertical interest aggregation, however, is lacking. The grassroots groups that originally supported Correa disappeared. Correa’s government instead used clientelism, co-optation, and strategically targeted benefits as the main tactics to relate with civil society organizations or sectors.
In Chapter 11, Piñeiro Rodríguez, Rosenblatt, and Toro Maureira analyze the case of the Chilean PPD. The PPD exhibits horizontal coordination in Congress and in elections. However, the coordination occurs between leaders who control different territories. There is no common party organization that coordinates and integrates social interests. The structure of the party resembles a federation, where a small cadre of leaders dominates different territories and exerts control over nominations, programmatic proposals, and segmented linkages with society. The case of the PPD shows that one cannot infer simply on the basis of electoral results and the existence of partisan discipline in Congress (as a measure of horizontal coordination) that an organization has a common structure capable of developing some kind of partisan identity in society to aggregate of interests at the national level.
Chapter 12 reviews the case of the FP in Peru. Vergara and Augusto question whether vertical interest aggregation and horizontal coordination are necessarily functional to democracy and the rule of law. The FP is an example of this dysfunctionality because, in Peru, horizontal coordination and vertical interest aggregation involve the incorporation of illegal interests in the political process. According to Vergara and Augusto, the FP represents these illegal interests in Congress while it neglects the party’s programmatic agenda and does not effectively represent its electoral base. Moreover, the FP’s congressional behavior threatens democracy, in both its republican and liberal dimensions. The FP, like the AP in Ecuador (see Conaghan this volume), depends on the power of the party leadership to co-opt candidates and to enforce discipline. This power, in turn, is determined by the electoral power of the leader in the presidential election. When the leadership seems unlikely to win election or is defeated, internal conflicts and coordination problems arise.
Chapter 13 describes the UNE in Guatemala. Sanchez-Sibony and Lemus classify this case as Independents. The UNE, as with all electoral vehicles in Guatemala, relies on local caudillos to fulfill the requirement of presenting candidates in every district, set forth by the electoral law. Even though the UNE has a good record of legislative discipline, it suffers from large numbers of defections in all electoral cycles. This is the main indicator of the UNE’s inability to carry out horizontal coordination. Sanchez-Sibony and Lemus’s chapter shows how structural conditions seriously hamper horizontal coordination and, more critically, vertical interest aggregation. The fragmentation and feebleness of social organizations in Guatemala limit the possibilities of developing programmatic linkages. Therefore, the UNE substituted clientelism in place of vertical interest aggregation to develop a limited popular constituency.