The Sources of Internal Dominance

An internally dominant leader is a figure with unquestioned, preeminent power in his or her party—who, in common parlance, stands “head and shoulders” above other elites. When a leader dominates, no one can challenge that leader for presidential nominations, vie for internal control, or advocate expulsion without being marginalized.

Internal power has multiple, non–mutually exclusive sources. One source of internal power, to be sure, is external appeal itself, which gives leaders control over elites’ electoral fortunes and their access to patronage. As already suggested, if a leader has unrivaled external appeal, lower elites know they will lose votes or patronage opportunities if the leader defects. Thus they have an incentive to accommodate, support, and side with the leader. Popular leaders sometimes use their external clout as leverage to negotiate concessions internally (e.g., Lula) (Hunter Reference Hunter2010).

Yet, crucially, electoral indispensability does not, by itself, make a leader internally dominant. There are two reasons. First, major factions are not always driven primarily by electoral incentives (e.g., Peru’s IU, Venezuela’s LCR, Mexico’s early PRD). Insofar as party members are ideological, not pragmatic, popular leaders do not derive leverage from their external appeal. Second, popular leaders may lack additional assets that are critical for maximizing internal power. What are these additional assets?

One is crossfactional ties. A leader with strong relationships across factions, who can negotiate constructively with all players, may be “indispensable” for brokerage and mediation—and hence for cohesion (Ansell and Fish Reference Ansell and Fish1999). Leaders who disengage from internal affairs or who lack or fail to maintain constructive, working relationships with the heads of all major factions cannot serve as crossfactional brokers and mediators (e.g., Barrantes, LCR’s Andrés Velásquez, Óscar Berger of Guatemala’s PAN).

Forging crossfactional links takes skill and, importantly, time. Thus, in new parties, a leader with strong preexisting crossfactional ties may be critical. Here we find variation. Some leaders have strong preexisting crossfactional ties because they led their parties’ founding struggles and closely collaborated with future party elites and feeder organizations (e.g., Lula). Other leaders have weak preexisting crossfactional ties. Indeed, figures may be tapped as leaders precisely because they are relative outsiders and do not tilt the internal balance of power in any faction’s favor (e.g., Barrantes).

Another source of internal power is moral authority. Moral authority is the respect and credibility that party leaders have internally, typically due to their backgrounds. Here, too, we find variation. Some leaders are deeply respected by their base. Certain leaders even possess a mystical quality, or are considered central to the party’s identity or the embodiment of its higher cause. Such stature can follow from pedigree (e.g., Cárdenas); class status (Lula); heroism (Charles de Gaulle of France’s Republicans); publicized experiences of hardship (Nelson Mandela of South Africa’s African National Congress, ANC); or protagonism in founding struggles (Lula, Robert Mugabe of the Zimbabwe African National Union, ZANU).Footnote ⁴ Moral authority can also be rooted in personal charisma (Weber Reference Weber1965; Panebianco Reference Panebianco1988). By contrast, some individuals become party leaders despite lacking moral authority. This may occur when, as described above, new parties select a relative outsider as their leader (Barrantes).

A final source of internal power is ideological representativeness of the base. To be sure, base-level attitudes are heterogeneous, and party leaders often have significant ideological autonomy. Still, ideological representativeness matters. Party leaders have more internal appeal and support when their stances align with those of most members. When leaders’ stances deviate from those prevailing among the rank and file, they become more vulnerable to internal challenges (e.g., Barrantes) and fatal impasses (LCR’s Andrés Velásquez).

In summary, internal power does not come from external appeal alone but also from crossfactional ties, moral authority, and ideological representativeness. These factors may or may not be independent of each other.Footnote ⁵ Regardless, they generate internal power independently and will be treated as having roughly equal weight. The more of these sources, and of each source, that leaders possess, the more internally powerful they will be. Consequently, new party leaders, even externally appealing ones, vary in internal power. While some are internally dominant, others are not (see figure 1).

Figure 1 Sources of Internal Power

Although this argument focuses on the role of leaders, it is primarily structuralist, not voluntarist. To be sure, internal dominance is not wholly static; external events and a leader’s contingent decisions can lead to short-term change in crossfactional ties, moral authority, and ideological representativeness. Nevertheless, the starting endowments largely determine the parameters and likelihood of such change: it is easier to maintain preexisting crossfactional ties than to build them from scratch; it is easier to establish moral authority if one has the right pedigree or background. Seldom does internal dominance follow, mainly, from effort or savvy.Footnote ⁶ It tends to be based primarily on objective endowments: electoral clout, preexisting ties, ancestry, a background of leadership, or heroism. Cárdenas, for example, was not a once-in-a-generation leader like Lula, but his endowments enabled him to dominate internal PRD affairs, as we will see.

How Externally Appealing, Internally Dominant Leaders Prevent Schisms

Externally appealing, internally dominant leaders generate cohesion in new parties. One key difference between new and institutionalized parties is that very few new parties have minimally efficient, routinized decisionmaking procedures. Although some parties eventually establish such procedures, new parties must do so from scratch and avoid alienating important factions, making the process of institution building delicate and gradual. Many new parties simply lack formal procedures in key areas (Venezuela’s LCR; Mexico’s early PRD). Some make decisions by elite consensus (Peru’s IU, LCR). Almost none have consolidated internal democracies. Thus, new parties typically lack formal mechanisms for aggregating member preferences and cannot, by institutional means, resolve important conflicts or prevent obstruction.

Externally appealing, internally dominant leaders solve these problems. First, they anchor dominant factions (Panebianco Reference Panebianco1988), which control party machinery and facilitate collective decisionmaking (Lula, Cárdenas). Second, they influence internal debates (e.g., on program, alliances), often in their own favor. Morally authoritative leaders may persuade members to moderate or make concessions for electoral purposes (Lula). Internally dominant leaders can use party candidacies and posts as leverage in disputes (Cárdenas). The inability to influence internal debate in these ways may create incentives to defect (Barrantes, LCR’s Velásquez).

Third, an internally dominant leader can play a substituting role as preference aggregator, central decisionmaker, and final arbiter. Where internal democracy is weak and factions lack horizontal links, a new party leader with strong crossfactional ties can collect a wide range of views and factor them into party-level decisions (Cárdenas). Internally dominant leaders have significant leeway to take decisions and actions for their parties. Often, their word is effectively law, meaning that they can impose party lines and arbitrate conflicts (e.g., Víctor Raúl Haya de la Torre of Peru’s APRA, ]uan Perón of Argentina’s Peronists, Roberto D’Aubuisson of El Salvador’s ARENA, as well as Cárdenas and Lula). Morally authoritative leaders are unlikely to be considered sellouts or traitors if they unilaterally break with party orthodoxy or precedent (e.g., Lula). If denied such freedom of action and maneuvering room, leaders are more liable to defect (Barrantes) or to reach impasses with vital rival factions (LCR’s Velásquez). In short, just as electorally indispensable leaders substitute for brands, internally dominant leaders—in the above ways—substitute for institutions of preference aggregation, collective decisionmaking, and conflict resolution.

Fourth, the presence of externally appealing, internally dominant leaders prevents crippling nomination battles. Presidential candidate selection is a winner-take-all choice with singular stakes. Naturally, an externally appealing, internally dominant leader can secure a party’s nomination with limited resistance. By contrast, internally nondominant leaders—even electorally indispensable ones—may face serious resistance in pursuit of nominations (Barrantes). Debilitating schism might result, as the aspiring nominee who is not selected, or fears not being selected, could defect (again, Barrantes).

A final point requires mention. The argument just laid out is probabilistic. Success or failure in party building is always the product of multiple causes or determinative variables. This article identifies one potential cause: the presence or absence of an externally appealing, internally dominant leader during the formative period. New parties can rise and take root without such leaders (e.g., Mexico’s PRI; El Salvador’s FMLN) or fail to take root despite the presence of such leaders (e.g., Carlos “Chacho” Álvarez of Argentina’s FREPASO). The presence of an externally appealing, internally dominant leader, then, is neither necessary nor sufficient for successful party building. Instead, it increases the likelihood of successful party building—at least in new parties that, like the PT, PRD, and IU, emerge from social mobilization.Footnote ⁷ This article will argue, however, that the role of the leader was indeed outcome-determinative for the PT, PRD, and IU.

Barrantes and Peru’s IU

IU was a socialist coalition founded in 1980, months after Peru’s transition to democracy. During the 1980s, with Alfonso Barrantes as lead candidate, IU rose to prominence, becoming one of Peru’s three leading parties, alongside APRA and AP. IU depended on Barrantes for its popularity. According to IU leader Henry Pease, “without Barrantes, IU was nothing” (Pease Reference Pease2010, Reference Pease2011). Among IU leaders, Barrantes was uniquely popular with floating, lower-income voters, a decisive constituency in the 1980s. These voters supported redistribution but identified with no party—especially no left party (Roberts Reference Roberts1998, 186, 248; Herrera Reference Herrera2002, 82, 186). Yet many supported Barrantes, who softened and humanized the left (Panfichi Reference Panfichi2010 and Tanaka Reference Tanaka2011). Due to his popular appeal, Barrantes also attracted pragmatic left voters who regarded an IU with Barrantes as electorally viable.

Barrantes’s main internal power source was electoral indispensability. During the 1980s, no left figure emerged who rivaled him in popularity. Even in the late 1980s, after his image of electoral prowess had begun to decline due to successive losses, his status as IU’s most electable elite remained undisputed. His moderate supporters emphasized that he was the left candidate most likely to win the 1990 presidential election, and perhaps the only one who could win (Herrera Reference Herrera2002, 379–80, 411; Gonzales Reference Gonzales and Adrianzén2011, 40). His radical opponents, until the end, conceded his unrivaled popularity (Cameron Reference Cameron1994, 80).

Yet despite his singular external appeal, Barrantes was not internally dominant (Tanaka Reference Tanaka1998, 139). In part, this was because his electoral clout did not matter much to IU radicals, who constituted most of the leadership and active base. IU radicals were not concerned, fundamentally, with maximizing IU’s vote share. They did not consider electoral failure—their own, much less IU’s—existentially threatening (Cameron Reference Cameron1994, 78). They were wary of governing, and hence of winning executive posts; this became especially true in the late 1980s, when Peru’s massive economic and security crises brought the country to the brink of state collapse (Roberts Reference Roberts1998, 224, 230, 252–53).

Even more damaging for Barrantes, however, was that his internal power rested almost exclusively on electoral leverage. He lacked moral authority and strong crossfactional ties, and his moderate views put him at odds with most of IU’s rank and file.

These limitations can be traced to the circumstances of his selection as IU leader. Given the sectarianism of IU parties, coalition leaders, at IU’s inception, were unwilling to cede IU’s presidency and lead candidacy to rivals. In addition to having electoral potential, Barrantes did not belong to an IU constituent party and was perceived as neutral between them (Roberts Reference Roberts1998, 248; Herrera Reference Herrera2002, 82). Yet neutral and independent leaders, almost by definition, lack the background and preexisting network ties characteristic of internally dominant leaders. So it was with Barrantes. His external origins made him acceptable but also limited his internal power.

When IU was born, Barrantes had never held office or built or led a party. He had not participated prominently in the mobilizations fomented by President Velasco (1968–75) or in the movement to topple President Morales Bermúdez (1975–80). In short, he was not an IU founder, unlike figures such as Jorge del Prado (Peruvian Communist Party, PCP) and Javier Díez Canseco (Mariateguista Unified Party, PUM). According to one pro-Barrantes IU elite, he “did not found IU but was called to preside over it” (Marcial Rubio, paraphrased in Tuesta Reference Tuesta1987). Barrantes thus entered IU without moral authority. In December 1981, Horacio Zevallos, a radical IU legislator, publicly stated, “Barrantes does not represent any of the organized political sectors, nor does he represent the masses. He is a novice lawyer, and we have made him, a substitute, a center forward in the leadership of the left” (quoted in Herrera Reference Herrera2002, 119). After Barrantes resigned as IU president in 1987, analyst Fernando Tuesta criticized his pre-IU record:

IU members offer similar assessments retrospectively. Gonzales (Reference Gonzales and Adrianzén2011) writes,

In the pithy formulation of an IU moderate, “Barrantes was accepted as a candidate but questioned as a leader” (Panfichi Reference Panfichi2010).

Barrantes also lacked strong crossfactional ties. Naturally, as a relative outsider, he lacked a large network of preexisting relationships with top IU leaders. Moreover, as IU president, he did not act as a crossfactional broker or arbiter. He “tended to be an aloof leader who was disengaged from the internal affairs of IU coalition” (Roberts Reference Roberts1998, 248). Moreover, by the mid-1980s, he advocated the expulsion of the PUM and Union of the Revolutionary Left (UNIR), IU’s two strongest parties, which together constituted the vast bulk of its radical bloc (Cameron Reference Cameron1994, chap. 5; Herrera Reference Herrera2002). In mid-1987, Barrantes resigned as IU president, abandoning the formal pretense of standing above faction or representing the whole coalition (Herrera Reference Herrera2002, 370).

For most of IU’s existence, the role of crossfactional broker was occupied by the neutral bloc, a group of moderates anchored by the Peruvian Communist Party and independent left Christians (Cameron Reference Cameron1994, chap. 5; Herrera Reference Herrera2002). During the second half of the 1980s, neutral bloc leaders, especially Henry Pease, sought to fuse the radical and moderate sectors of IU into a party. They opposed expulsions and divisions. In the late 1980s, they regularly met with both Barrantes and radical leaders in a futile effort to keep the coalition united (Herrera Reference Herrera2002). Far from standing above faction, then, Barrantes led one of the two factions between which the neutral bloc mediated.

Furthermore, Barrantes, a moderate, did not ideologically represent the mostly radical IU base. Barrantes had a tense relationship with radicals throughout IU’s existence, and radical critiques of Barrantes intensified in the late 1980s. First, radicals objected to the amicable relationship that developed in the mid-1980s between Barrantes and Alán García (APRA), Peru’s president from 1985 to 1990. They argued that García sought to use Barrantes to tame and divide IU by marginalizing its radical wing. This view, which became widespread among the base, fueled two pivotal conflicts between Barrantes and IU activists in 1986 and 1987 that helped to precipitate Barrantes’s resignation as IU president in mid-1987 (Herrera Reference Herrera2002, 360).

Second, as Peru was reeling from economic crisis and insurgent violence in the late 1980s, Barrantes and other moderates argued that the left should pursue its agenda exclusively through democratic institutions, collaborate with APRA and the army to preserve Peruvian democracy, and prioritize victory in the 1990 presidential election. By contrast, most radicals believed that Peru had entered a revolutionary situation, that elections had become secondary, that armed struggle had to remain on the table, and that there could be no alliance with APRA or the army.

In short, although Barrantes was electorally indispensable, his moral authority was limited, his crossfactional ties were not strong, and he did not ideologically align with most IU members. Thus he was not internally dominant. In the late 1980s, his arguments and entreaties for moderation had little to no effect on IU radicals; throughout 1989, for example, PUM categorically refused to reject armed struggle (Herrera Reference Herrera2002, 460–62, 475). And crucially, Barrantes failed to impose himself as IU’s 1990 presidential nominee. In the lead-up to the election, IU radicals, who held a majority on IU’s national executive committee, opposed Barrantes’s nomination and demanded a closed primary in which they could run a rival nominee. Barrantes believed that he might lose an internal primary, as the radical parties, given their size and mobilizing capacity, held an advantage over moderates (Tanaka Reference Tanaka1998, 139). Barrantes calculated that his most likely route to victory was to contest the first round on a non-IU ticket. If he made it to the second round (a plausible goal in late 1989), a broad center-left coalition including much of IU would probably coalesce around him.

By this rationale, Barrantes defected from IU—and killed it. In the 1990 election, both Barrantes and IU performed extremely poorly. IU candidate Henry Pease won only 8 percent of the vote. Barrantes, newly divorced from IU and competing against its label and machines, did even worse, garnering a mere 5 percent. Shortly afterward, Barrantes left politics, while IU splintered and collapsed.

Cárdenas and Mexico’s PRD

The PRD was born in 1989 under the authoritarian rule of the Institutional Revolutionary Party (PRI). It grew out of a movement to elect PRI defector Cuauhtémoc Cárdenas president in 1988. After Carlos Salinas (PRI) defeated Cárdenas in a close and questionable election, the pro-Cárdenas movement strengthened, giving rise to the PRD.

Cárdenas was electorally indispensable. Writes Borjas, “without his presence, [the PRD] was nothing” (2003, 508). The early PRD lacked media access and could not build support quickly (Lawson Reference Lawson2002), but Cárdenas did not need media or campaigns to attract mass support. He was the son of General Lázaro Cárdenas, revolutionary general, ex-president, and left hero. Because of his last name, Cárdenas “communicated just by virtue of existing. The campaign was carried out every day in school” (Lajous Reference Lajous2011). Cárdenas’s coattails also generated cohesion, as “there were few external incentives for the party’s intermediate leaders to split” (Rodríguez Reference Rodríguez2010, 257).

In addition to attracting voters, Cárdenas, importantly, attracted PRD members. Indeed, it was Cárdenas, not the PRD, who supplied the “higher cause” motivating most PRD activists. He had “an almost monopolistic capacity to produce collective goods” for the rank and file (Rodríguez Reference Rodríguez2010, 255), whose loyalty “was fundamentally directed to [Cárdenas] and only in the second place to the party itself” (Borjas Reference Borjas2003, 450–51).

This loyalty followed from Cárdenas’s moral authority. According to one party member, Cárdenas had “the moral quality to be everyone’s leader” (Hidalgo Reference Hidalgo2011). Cárdenas’s moral authority came in part from his political record. As PRI governor of Michoacán (1980–86), Cárdenas defied the neoliberalism of President Miguel de la Madrid. He then defected from the PRI, was probably defrauded of victory in the 1988 presidential election, then refused to be co-opted, rejecting President Salinas’s offer of Mexico City’s regency and becoming an intransigent opponent of the government.

Above all, though, Cárdenas’s moral authority came from his lineage. Cárdenas was the last name “most respected by the political left” (Borjas Reference Borjas2003, 293). As Lázaro Cárdenas’s only son, Cárdenas bore, for leftists, the legacy of the Mexican Revolution. His pedigree gave him a mystical quality among the base (Somuano Reference Somuano2011). Even after retiring from active involvement in the PRD, he was known as the party’s moral leader.

Cárdenas established strong crossfactional ties as PRD leader. At the base level, his “incontestable leadership became a node of organization between the different groups that … did not manage to organize horizontally” (Rodríguez Reference Rodríguez2010, 254, 263; Borjas Reference Borjas2003, 448). At the elite level, he was “the one responsible for building bridges” (Martínez Reference Martínez2005, 99). For national leaders, “the establishment of direct contacts with [Cárdenas] became almost a rule” (Prud’homme Reference Prud’homme1996, 12, n 30). Cárdenas “occupied the center, toward which the distinct corrientes [factions] with their respective leaders converged” (Prud’homme Reference Prud’homme2003, 118). To preserve these relationships, Cárdenas generally avoided taking sides in disputes and, where possible, synthesized opposing positions (Borjas Reference Borjas2003, 509).

In addition, Cárdenas ideologically represented the PRD base. This base, drawn from social movements and the ex-PRI, was predominantly radical (Greene Reference Greene2007). Although inclusive as a leader, Cárdenas consistently adopted radical positions. He prioritized base-level demands, supported the conception of the PRD as a movement, and took uncompromising stances at the expense of the PRD’s broader electoral appeal (e.g., intransigence, opposing NAFTA, meeting with rebel leader Subcomandante Marcos). Cárdenas’s steadfast radicalism cemented members’ loyalty to him (Borjas Reference Borjas2003, 303).

Cárdenas’s electoral indispensability, moral authority, strong crossfactional ties, and ideological representativeness made him internally dominant. He “stood above the different leaders that … coexisted in the PRD” (Borjas Reference Borjas2003, 299). Key party decisions were taken and conflicts adjudicated unilaterally by Cárdenas, often without debate or negotiation (Bruhn Reference Bruhn1997, 190; Borjas Reference Borjas2003, 451) and rarely through formal procedures, which remained weakly institutionalized throughout the 1990s and beyond (Borjas Reference Borjas2003, 445–60; Prud’homme Reference Prud’homme2003, 104, 118; Martínez Reference Martínez2005, 97–101; Mossige Reference Mossige2013). Cárdenas thus played a “substituting role for the lack of institutionalization” (Martínez Reference Martínez2005, 101).

To manage his responsibilities, Cárdenas created an informal dominant coalition. He appointed a network of aides, to whom he delegated and entrusted different tasks, and who spoke for him. Although this nucleus never became a formal faction, it constituted, for Prud’homme (Reference Prud’homme1996) and Rodríguez (Reference Rodríguez2010, 256), the early PRD’s dominant coalition, and its “central factor was Cárdenas’s incontestable leadership” (Rodríguez 2005, 256).

Cárdenas dominated internal affairs in three main areas: candidate selection, the allocation of party offices, and program and tactics. Specifically, he imposed the line of intransigence; lent decisive support to Porfirio Muñoz Ledo and Andrés Manuel López Obrador in their campaigns for the PRD presidency in 1993 and 1996; imposed a radical agenda in the mid-1990s, meeting with Subcomandante Marcos, opposing NAFTA, and terminating negotiations on electoral reform; resoundingly defeated Muñoz Ledo in a closed primary election to determine the PRD’s Mexico City 1997 mayoral nominee; and, with overwhelming internal support, secured the PRD’s presidential candidacy in 1994 and 2000 in the absence of a formal nomination procedure (Borjas Reference Borjas2003).

Cárdenas became less active in the PRD after the mid-1990s, and following his 2000 presidential defeat, retired from politics. Yet by this time, the PRD had an effective brand and was no longer electorally dependent on him. Most Mexicans, by the year 2000, could locate the PRD symbol, an Aztec sun, on the left-right spectrum (Lawson et al. Reference Lawson2001), and partisan voters guaranteed the PRD a high electoral floor. This brand made the PRD electorally viable and discouraged elite defection independently of any leader.

The PRD’s development since the mid-1990s is well known and widely studied (Borjas Reference Borjas2003; Martínez Reference Martínez2005; Greene Reference Greene2007; Mossige Reference Mossige2013). Many scholars treat the PRD as a failure or underachiever, citing repeated presidential defeats (Greene Reference Greene2007; Bruhn Reference Bruhn1997) and chronic infighting (Martínez Reference Martínez2005; Mossige Reference Mossige2013). It has not won the presidency, having paid a price for catering to its base (Greene Reference Greene2007). Recently, the party has lost electoral support, and its two main leaders, Andrés Manuel López Obrador and Cárdenas, have defected (in 2012 and 2014, respectively). López Obrador’s personalistic vehicle, the National Regeneration Movement (MORENA), currently rivals the PRD in national elections and has overtaken the PRD in Mexico City, the PRD’s bastion.

Yet the PRD’s problems should not be overstated. Unlike MORENA, the PRD is institutionalized. It remains a national contender, having passed the 10 percent threshold in Congress continuously since 1994 with a higher average vote share than Brazil’s PT. Since 1998, it has won more than 15 gubernatorial elections (three in coalition with the PAN). Since 1997, it has held the Mexico City mayoralty, one of Mexico’s top offices.

For the PRD, collapse remains unlikely. Partisan voters give the PRD a reliable share of congressional seats, which, due to landmark 1996 electoral reforms, ensures steady access to financial resources. PRD foot soldiers remain a major electoral asset. In sum, the PRD is institutionalized, and perennially contends for national power. Few recently born Latin American parties have achieved as much. Comparatively, the PRD is an unmistakable case of successful party building, and Cárdenas’s external appeal and internal dominance during the formative period proved critical for this success.

Lula and Brazil’s PT

Brazil’s PT was established in 1980 under the military dictatorship (1964–85). It was founded by groups that led Brazil’s grassroots democratizing movement: the militant labor movement and the “new unionism,” together with the Catholic and Marxist left.

During its formative years, the PT electorally depended on its founding leader, Lula. In the lead-up to the PT’s formation, Lula, as leader of the new unionism, became the national face of the prodemocracy struggle. In the early PT, his coattails substituted for a strong brand: “the party had to include Lula to get off the ground” (Keck Reference Keck1992, 81). Lula “enjoyed more societal support than the party” (Hunter Reference Hunter2010, 3), and “all PT candidates depended on [his] electoral performance” (Secco Reference Secco2010). Lula’s coattails generated cohesion. Members coordinated around the goals of electing Lula governor of São Paulo (1982) and president (1989, 1994). Radical PT leader Hamilton Pereira (Reference Pereira2008) stated retrospectively that the PT’s “high standard of discipline and unity of action” was “largely generated by the expectation to elect Lula president” 270).

Lula was also internally dominant (Hunter Reference Hunter2010, 3, 36, 122). Asked to explain the PT’s survival, PT historian Lincoln Secco highlighted the role of Lula, who had both “external and internal charisma” (Secco Reference Secco2010). Lula’s humble origins, working-class bona fides, and background as movement leader gave him moral authority. The PT cast itself as Brazil’s first authentically popular party. The São Paulo labor leaders, led by Lula, incarnated this founding myth. They made the PT’s animating narrative credible and provided the higher cause that motivated early members.

Notably, PT radicals, who criticized Lula’s moderation, respected him and accepted his leadership. “However much the left might criticize what it called his vacillation, it recognized that Lula was still the authentic working-class leader par excellence” (Keck Reference Keck1992, 81). Radicals rarely publicized their critiques of Lula (Betto Reference Betto2010) and virtually never questioned his leadership (Ribeiro Reference Ribeiro2012).

Lula also had strong crossfactional ties. As leader of the grassroots democratizing movement, he, before the PT’s founding, built relationships across PT feeder organizations, including rural, industrial, and middle-sector unions and left Catholic, Marxist, and student groups. As PT leader, he “was always one of the few ‘glues’ between the factions, above them all, unifying the party” (Ribeiro Reference Ribeiro2012), and he acted as the central negotiator and guarantor of agreements internally (Rodríguez Reference Rodríguez2010, 208).

His inclusive, nonconfrontational leadership style reinforced his crossfactional ties. He “put himself above the factions” and refrained from explicitly “taking sides” (Secco Reference Secco2010). For more than a decade, he did not join or directly participate in any faction (De Azevedo Reference De Azevedo1995, 154), instead authorizing aides—mainly José Dirceu—to advance his agenda and “use the iron fist when necessary” (Ribeiro Reference Ribeiro2012). He abstained from contentious votes. He never supported expelling or silencing opposing factions (Betto Reference Betto2010; Donato Reference Donato2010) and rarely conflicted directly with opponents. Instead, he embraced the PT’s ideological diversity. Radical PT leader Hamilton Pereira retrospectively extolled Lula for his inclusivity and for clearly defining the PT at the outset as a political rather than ideological pact (Pereira Reference Pereira2008, 264).

Lula ideologically represented the core PT base. He was a non-Marxist and widely known to be a relative moderate (Keck Reference Keck1992; Hunter Reference Hunter2010; Ribeiro Reference Ribeiro2010; Secco Reference Secco2011). The rank and file, too, were predominantly moderate, drawn over-whelmingly from the new unionism and, secondarily, the Catholic left (Secco Reference Secco2011, 49). In 1991, only 10 percent of PT members had roots in Marxist organizations of the “extreme left” (Secco Reference Secco2011, 48). Moderate factions dominated internal PT elections and consistently controlled the national party apparatus, save for one brief period (1993–95).

Lula’s assets and characteristics made him dominant, and he anchored the PT’s controlling bloc from the early 1980s on. In 1982, new unionists founded the Articulação faction, positioned in the PT’s center. Articulação won every internal election until 1993 and, after a brief period of radical control (1993–95), regained power under the Campo Majoritário label. Without the backing of Lula, their “symbolic” leader (Secco Reference Secco2011, 93), Articulação and Campo Majoritário could not have dominated the PT. The legitimacy of these groups followed from the working-class origins, record of struggle, and ideological positions of their leaders, above all Lula (Keck Reference Keck1992, 116).

The main purpose of forming these dominant coalitions was to limit the radical left’s influence. Articulação was designed to “restrain the activity” of Marxist tendencies, which, it was argued, had unduly influenced the PT’s development from 1980 to 1982 (Ribeiro Reference Ribeiro2010, 186–87; Hunter Reference Hunter2010, 114, 118–19, 121; Secco Reference Secco2011, 123). Similarly, Campo Majoritário was conceived as an instrument for institutionalizing moderate control of the PT after the radical left’s two-year stint in power and Lula’s defeat in the 1994 presidential election. It succeeded, transforming the PT into a center-left electoral-professional party that held the presidency from 2002 until 2016 (Hunter Reference Hunter2010).

Tellingly, even when radicals controlled the PT (1993–95), Lula used his “electability” (external appeal) and “popularity among petistas” (internal dominance) to force the PT to moderate (Hunter Reference Hunter2010, 3, 6, 36, 120–26). He had a “moderating influence” on the “program, tone and tactics” of the PT’s 1994 campaign (Hunter Reference Hunter2010, 120). He went outside formal party channels to make alliances to the PT’s right, and the radical PT leadership granted him this leeway (Hunter Reference Hunter2010, 121–22).

Lula was uncontested in internal elections and as the PT’s presidential nominee. He won the presidential nomination by overwhelming consensus five times (1989, 1994, 1998, 2002, 2006), even though the radical left controlled the party in 1994, and even though his defeats in 1994 and 1998 damaged his image of electoral prowess. Most revealingly, in 1993, at the peak of internal moderate-radical tensions and just months after the PT’s “extreme left” factions had won control of the party, Lula’s candidacy for the PT presidency “met with overwhelming internal consensus” (Hunter Reference Hunter2010, 121–22).

After 2002, the PT became Brazil’s most successful party. It “adapted,” accepting macroeconomic orthodoxy and modernizing its campaigns (Samuels Reference Samuels2004; Hunter Reference Hunter2010). These adaptations led to Lula’s 2002 presidential victory and repeat PT victories in 2006, 2010, and 2014. Since 1994, the PT has won 10 to 20 percent of congressional vote share. It holds numerous governorships. Despite President Dilma Rousseff’s recent impeachment and a drop in partisan identification since 2013, the PT remains Brazil’s most institutionalized party. It has a territorial organization, committed activists, strong institutions of internal democracy, and a still-powerful brand likely to ensure electoral relevance for decades.

Comparative Inferences, Objections, and a Shadow Case (Venezuela’s LCR)

The three case studies compared show that Lula and Cárdenas were internally dominant while Barrantes was not.Footnote ¹⁰ Lula and Cárdenas controlled their parties’ tactics and platforms and won nominations and internal elections with ease, even when their opponents held power (Lula in 1993 and 1994) and after their images of electability had been tainted. By contrast, Barrantes did not control IU’s tactics and platform; he was not, like Cárdenas, his party’s arbiter; and he faced major resistance to IU’s presidential nomination. This variation matters. During the PT’s and PRD’s formative periods, Lula and Cárdenas never had strong incentives to defect. By contrast, Barrantes’s failures to moderate IU and impose himself as nominee triggered his fatal defection.

It is noteworthy that Lula had clout with PT radicals despite being a moderate, and that Cárdenas was his party’s most electable figure despite being a radical. These facts suggest that in a possible world, an IU radical could have been externally popular, or an IU moderate—perhaps one with greater moral authority and stronger crossfactional ties than Barrantes—could have held sway over IU radicals.

One might object, though, that the case of IU is simply too different from the cases of the early PT and PRD to justify such comparative arguments. IU was not a party but a coalition of parties.Footnote ¹¹ Circumstances in Peru in the late 1980s were unique: the country was suffering hyperinflation, recession, and the escalating violence of the Shining Path insurgency. These crises polarized IU radicals and moderates and worsened voter perceptions of the left. They brought Peru to the brink of state collapse and contributed to the entire party system’s eventual breakdown. Thus, one might argue, first, that given heightened moderate-radical polarization in the late 1980s, IU was bound to split; second, that IU collapsed not due to the absence of a particular type of leader but due to the same external factors that brought down Peru’s entire party system; and third, that no leader in the late 1980s could have simultaneously dominated the radical IU and been electable.

These counterarguments are reasonable but unpersuasive. First, the argument that an IU schism was inevitable due to heightened moderate-radical polarization in the late 1980s is ahistorical. IU moderates did not split from IU radicals; a minority subset of IU moderates did (i.e., Barrantes and a club of allies). The neutral bloc, a much larger group of moderates, remained in IU (i.e., in alliance with the radical bloc). Second, IU’s collapse did not result from the same external factors that brought down Peru’s party system.Footnote ¹² Unlike APRA, AP, and the Popular Christian Party (PPC), IU split, and it split in mid-1989, well before the entire system collapsed.Footnote ¹³ Furthermore, IU was in the opposition during the late 1980s and therefore did not bear responsibility for Peru’s crises nearly to the degree that the governing party, APRA, did. Indeed, IU stood to gain from the discrediting of its main rival, APRA, as evidenced by polls in 1988 and early 1989 that forecast a first- or second-place finish for IU’s Barrantes in the first round of the 1990 presidential election (Taylor Reference Taylor1990, 113; Cameron Reference Cameron1994, 93; Tanaka Reference Tanaka1998, 135). For all these reasons, the question, why did IU split? merits individual analysis.

Third, the assertion that by the late 1980s, no leader could have simultaneously dominated IU internally and been popular externally is untenable. This claim implies, first, that no member of Peru’s radical left could have been externally popular, and second, that IU radicals would not have accepted a moderate as IU leader. The first premise is speculative and far from obvious. The second is false: IU radicals opposed Barrantes’s nomination but did not demand his expulsion or proscription; they insisted on fielding their own candidate but agreed to accept the outcome of the primary election; and after Barrantes’s exit, they assented to the nomination of a moderate (Pease).

It was a product of misfortune more than necessity that IU, instead of having one leader who combined external appeal and internal dominance, had one popular leader (i.e., Barrantes) and other leaders with separate sources of internal power. Several figures in IU had strong ties across the radical and moderate blocs, even in the late 1980s (e.g., Henry Pease, Rolando Ames, Jorge del Prado). Other figures had moral authority among IU’s core base (e.g., Javier Díez Canseco). To imagine an externally appealing, internally dominant IU leader, we need only conceive a hypothetical scenario in which one of these leaders (e.g., Pease, Díez Canseco) also happened to be popular, or potentially popular, with voters. To call such a scenario structurally implausible seems an overstatement.

Fourth, although it is tempting, with the benefit of hindsight, to argue that the failure of IU and success of the PT and PRD were predictable, I would argue otherwise, given the significant challenges faced by the PT and PRD and IU’s considerable assets and advantages. The PT and PRD, like IU, were founded by conflicting factions with contrasting views on the relative importance of elections and social mobilization. Both also suffered life-threatening electoral setbacks. Thus it was not obvious during the first half of the 1980s that the PT would survive intact (Keck Reference Keck1992, 18, 152–53, 156; Bruhn 1998, 250; Pereira Reference Pereira2008, 264; Ribeiro Reference Ribeiro2010, 186; Secco Reference Secco2011, 48–49), or in the early 1990s that the PRD would survive intact (Borjas Reference Borjas2003, 301, 450, 457–58, 518; Prud’homme Reference Prud’homme2003, 118, 118 n. 4; Martínez Reference Martínez2005, 59). Equally, it was far from obvious in the late 1980s that Barrantes would defect (Tanaka Reference Tanaka1998, 137; Herrera Reference Herrera2002, 379).

It is also critical to highlight IU’s assets and advantages. First, territorial organization is crucial for party building (Levitsky et al. 2016), and IU had one of the largest grassroots infrastructures on the Latin American left during the third wave, with local branches across Peru and, by the late 1980s, 130,000 to 150,000 active members (Tanaka Reference Tanaka1998, 135; Pease Reference Pease2010, Reference Pease2011). Second, IU developed a strong left brand during the 1980s. IU was in the opposition and consistently antineoliberal during its formative decade, unlike other new left parties that quickly attained national power, adopted unpopular austerity policies, and diluted their left-wing brands (Lupu Reference Lupu2016; Roberts Reference Roberts2016). Third, as we have seen, Peru’s crises in the late 1980s discredited IU’s main rival, APRA, which, as evidenced by polls throughout 1988 and early 1989, gave IU an opportunity to establish itself as Peru’s strongest party.

Fifth, even if we concede that IU, because it was a coalition of parties and developed during a major crisis, was more prone to collapse than the PT and PRD, it does not follow that IU’s collapse was inevitable—only that it was more likely than in the cases of the PT and PRD. Outcomes have multiple causes, and in analyzing the origins of IU’s fatal schism, this article has sought to identify one cause, not the only cause.

Sixth, although this article has focused on the unsuccessful case of Peru’s IU, a brief review of another unsuccessful new left case, Venezuela’s Radical Cause (LCR), suggests the argument’s wider applicability. Like IU, LCR was a mass-based party; it rose to electoral prominence on the coattails of a popular but internally nondominant leader (Andrés Velásquez); it consisted of moderate and radical factions with contrasting views on the permissibility of armed struggle; and it fatally split.

Born in 1971, LCR operated for two decades on the margins of Venezuela’s two-party system, which consisted of the historically center-left Democratic Action (AD) and center-right Independent Electoral Political Organization Committee (COPEI). LCR was a “movement of movements” (Crisp and Levine Reference Crisp and Levine1998), incorporating student activists, the urban poor, intellectuals, and most centrally, iron- and steelworkers in the Greater Guyana region of Bolívar state (Nogueira-Budny Reference Nogueira-Budny2014, 114). Velásquez, who became a nationally influential progressive figure as leader of Venezuela’s largest steelworkers’ union in Guyana, was LCR’s “most successful and well-known member” (Nogueira-Budny Reference Nogueira-Budny2014, 114). After the 1983 death of LCR founder Alfredo Maneiro, he became the “de facto party leader” due to his “charisma, electability and popularity” (Nogueira Budny Reference Nogueira-Budny2014, 114), and would contest the presidency three times, in 1983, 1988, and 1993.

In the late 1980s and early 1990s, LCR rose “meteorically” (Crisp and Levine Reference Crisp and Levine1988) on Velásquez’s coattails. In 1989, Velásquez won the governorship of Bolívar state; in 1992, LCR’s Aristóbulo Istúriz won the Caracas mayoralty; and in 1993, Velásquez won 22 percent of the presidential vote, turning LCR into a national contender.

Although widely acknowledged as LCR’s most electable member, Velásquez was not internally dominant, and his internal power diminished in the late 1980s and 1990s as LCR’s divisions widened. Like most left parties in Latin America during the third wave, LCR consisted of moderate and radical elements, with moderates prioritizing the pursuit of electoral success and radicals seeking to effect change through social mobilization, even armed revolt. Moderate-radical differences grew and hardened in the late 1980s. After the congressional elections of 1988, amid profound, protracted economic crisis, eroding support for Venezuela’s traditional parties, and declining political participation, LCR split into radical and moderate factions (Buxton Reference Buxton2001, 160–61). Velásquez led the moderate faction, while Pablo Medina—an ex-Communist ex-guerrilla with little visibility or popularity outside radical left activist circles—was LCR’s leading radical member (McCaughan 2004, 80–81; Nogueira-Budny Reference Nogueira-Budny2014).

As Venezuela descended into extreme political instability—capped by two failed coup attempts and the impeachment of President Carlos Andrés Pérez (AD)—LCR moderates and radicals came to differ on “the wisdom of maintaining links to army coup plotters while simultaneously courting respectability through the democratic process.” Velásquez favored a categorical rejection of armed struggle and “was anxious to press home the party’s electoral advantage.” Medina, by contrast, “continued to meet with [Hugo] Chávez” during the late 1980s and early 1990s. Velásquez blocked Medina’s attempts to organize a coup within LCR, and after “Medina’s group played a small role in the two coup attempts in 1992,” “it was almost expelled from the party” (all quotations from McCaughan 2004, 80–81).

Despite Velásquez’s unrivaled popularity with voters, LCR’s radical bloc was larger and more mobilized than the moderate bloc (Buxton Reference Buxton2001, 178–79; López Maya Reference López Maya2005, 189). Consequently, Velásquez was not ideologically representative; his moderate positions did not resonate with most of LCR’s base. Moreover, Velásquez did not function as a crossfactional broker or mediator, especially during the early to mid-1990s. Rather than standing above faction and leading the whole party, he gradually became the leader of one (numerically disadvantaged) faction.

The fissure between Velásquez and the radical core of LCR, led by Medina, became especially nasty and public in the lead-up to and aftermath of the 1995 municipal elections. LCR’s weak performance “brought the longstanding rivalry between Medina and Velásquez into the open with mutual recriminations over policy and tactics” (McCaughan 2004, 80). LCR’s fatal schism followed shortly afterward. In the lead-up to the 1997 general election, Velásquez and allied cadres reached an impasse in regard to the radical faction. Unable to impose a moderate party line, Velásquez’s group “[expelled] those who didn’t follow their ideas” (López Maya Reference López Maya2005, 189). Medina defected to form Fatherland for All (PPT), taking Aristóbulo Istúriz, ex-mayor of Caracas (1993–96); Alí Rodríguez Araque, another leading figure in LCR; and the “bulk of the movement” with him (Buxton Reference Buxton2001, 178–79). Eviscerated, LCR receded into electoral marginality.Footnote ¹⁴