Case Selection

Bolivia offers an ideal setting for evaluating the proposed argument. The country has a permissive institutional environment that facilitates party-building and electoral engagement (Albó Reference Albó and Sieder2002; Komadina and Geffroy Reference Komadina and Geffroy2007; Van Cott Reference Van Cott2005; Zuazo Reference Zuazo2009). It transitioned to democracy during the 1980s and, in the 1990s, implemented policies decreasing barriers to new party entry and increasing indigenous movements’ opportunities for electoral participation. During this period, Bolivia’s party system fragmented and collapsed, resulting in the erosion of traditional partisan attachments (Madrid Reference Madrid2012; Reference Madrid2005). As traditional parties disappeared, the open institutional environment interacted with the representational void to enable the emergence of multiple proto-parties (Van Cott Reference Van Cott2005). Established literature on new party entry suggests that parties seeking to fill ideological openings should be more successful than those attempting to enter already crowded ideological positions (Lago and Martínez Reference Lago and Martínez2011; Laroze Reference Laroze2017; Tavits Reference Tavits2006). In Bolivia, the representational arena was wide open during this period and opportunities for growth were relatively even across parties.

Focusing on Bolivia also allows me to control for the role of conflict in party-building processes. Levitsky, Loxton, and Van Dyck (Reference Levitsky, Loxton, Van Dyck, Levitsky, Loxton, Van Dyck and Domínguez2016) found the presence of explosive social conflict to be a critical factor explaining successful party-building in Latin America since democratization. Bolivia fits this story: it is characterized by a salient ethnic cleavage that has systematically structured political behavior (Madrid Reference Madrid2012; Van Cott Reference Van Cott2007; Yashar Reference Yashar2005). Coupled with the institutional environment, the well-established salience of this cleavage allows me to set aside the question of “were conditions favorable?” to evaluate how network structures shape party-building outcomes.

In this context of institutional openings, party system collapse, and increased conflict, multiple proto-parties emerged to compete for votes from Bolivia’s indigenous populations. The three most significant projects were the Assembly for Peoples’ Sovereignty (ASP), MAS-IPSP, and Pachakuti Indigenous Movement (MIP) (do Alto and Stefanoni Reference do Alto and Stefanoni2010).Footnote ⁷ These parties were similar in important ways. They emerged during the same period and responded to the same representational void. The three proto-parties originated from indigenous peasant organizations with strong regional bases and extensive mobilization capacity. Their origin organizations were either regional or subregional branches of CSUTCB, Bolivia’s most important rural organization, and their organizational resources were ideal for establishing a movement-based party with extensive appeal (Anria Reference Anria2018; García Yapur, García Orellana, and Soliz Romero Reference García Yapur, Orellana and Romero2014; Van Dyck Reference Van Dyck2014).

The three proto-parties were also led by influential leaders. Felipe “Mallku” Quispe (MIP), Evo Morales (MAS-IPSP), and Alejandro “Alejo” Véliz (ASP) were at the forefront of a historical cycle of indigenous protests. Notably, both Quispe (MIP) and Véliz (ASP) were elected to CSUTCB’s top positions during this period, serving as executive secretary and first secretary, respectively. Morales (MAS-IPSP) led the powerful coca growers’ movement (Coordinadora), a much smaller but highly mobilized and politically influential organization. These leaders therefore enjoyed enormous legitimacy and mobilization capacity among Bolivia’s indigenous peasantry (García Yapur, García Orellana, and Soliz Romero Reference García Yapur, Orellana and Romero2014; Gutiérrez Aguilar Reference Gutiérrez Aguilar2008; Patzi Reference Patzi1999).

The three proto-parties also had comparable programmatic positions that found their inspiration in a wave of national indigenous mobilizations. They advocated and mobilized for issues including gas nationalization, rejection of free trade agreements (and neoliberal policies broadly), new communal land rights, and elimination of coca eradication policies, all central issues during the protest cycles (Albó Reference Albó2009). They also demanded a Constitutional Assembly, increased indigenous political participation, and proposed other reforms considered radical by traditional parties. Their agendas and discourse incorporated demands articulated during protests and shared by indigenous and left-leaning sectors. The similarities across the parties were such that ASP and MAS-IPSP are typically treated as the same in literature, with ASP seen as the precursor to MAS-IPSP (García Yapur, García Orellana, and Soliz Romero Reference García Yapur, Orellana and Romero2014; Gutiérrez Aguilar Reference Gutiérrez Aguilar2008; Van Cott Reference Van Cott2005). Morales himself argued that “the emergence of MAS-IPSP is simply the continuity of ASP, which has already met its function” (Burgoa Moya Reference Burgoa Moya2016, 78). While MAS-IPSP is better conceived as an off-shoot of ASP—the division between them generated two distinct proto-parties that participated in elections separately and competed for control of indigenous peasant support bases—the prevalent view highlights the similarities between these proto-parties which, along with Quispe’s MIP, are part of the same wave of indigenous party-building.

Finally, the parties also pursued comparable party-building strategies. They identified the need to make inroads with urban and labor movements and sought to compensate for their limited appeal in various ways. Their strategies toward this end included: the enhancement of alliances with the national workers’ confederation (COB), efforts at attracting support from working-class leaders and intellectuals by inviting them to run on party lists (Anria Reference Anria2013; Reference Anria2018; Madrid Reference Madrid2012), a broadening of parties’ agendas to include issues of interest to urban and working-class sectors (Albó Reference Albó2009), and the adoption of party rhetoric that heavily emphasized unity among indigenous and working classes. Altogether, the various similarities—in timing, resources, leadership legitimacy and influence, programmatic views, and party-building strategies—serve as crucial controls for examining how the proto-parties emerged from indigenous organizations and used their network environments to expand and compete.

The final condition that makes Bolivia ideal for evaluating the proposed theory turns to differences between these party-building projects. As I show, despite having strong roots in Bolivia’s indigenous peasant movement, the structural resilience of proto-parties’ network environments nonetheless varied significantly. Controlling for various factors enables me, to the extent that a nonexperimental setting allows, to isolate the variable of interest—network environments’ structural resilience—and trace its relationship to party-building outcomes.

Before proceeding, there are potential confounding factors to consider. Particularly, despite the significant similarities across these proto-parties, there are nonetheless notable differences in leadership styles and party-building strategies that could account for MAS-IPSP success. Madrid (Reference Madrid2012) highlights Quispe’s exclusionary ethnopopulism and argues that Morales’ inclusionary approach contributed to MAS-IPSP’s success. Anria (Reference Anria2018), for his part, attributes MAS-IPSP’s organizational expansion into noncore constituencies to a party strategy involving the “negotiation of spaces of power and influence within the government” with organizational leaders (88). Undoubtedly, Morales’ charisma, organizational skills, and inclusive populist discourse, along with MAS-IPSP’s political linkage strategies, are important factors contributing to MAS-IPSP’s success. These explanations are not incompatible with the proposed argument. However, as I will show, variation among them cannot effectively account for the major differences in proto-parties’ network environments. Indeed, this study’s findings suggest that these alternative explanations may be doing much of their work through proto-parties’ organizational networks. The opportunities and constraints that network structures offer proto-parties may condition the design and effectiveness of organizational efforts, linkage tactics, and appeal strategies.

Data Collection and Methodology

The analyses that follow adopt a mixed-methods strategy, using process tracing and network analysis to assess the association between network structures and party-building outcomes. I employ data gathered through extensive fieldwork in Bolivia between 2013 and 2015, including newspaper and electoral data, organizational documents, and secondary sources. I use process tracing to identify the origins and central features of variation in network environments, trace how parties grow through networks, and assess how competition alters pathways to expansion.

For the network analysis, I constructed an original organizational network dataset using newspaper articles gathered through Bolivia’s Center for Documentation and Information’s 10 Years of History, 30 days of news (2002–2011) newspaper database (CEDIB 2012).Footnote ⁸ The network dataset traces positive intra- and inter-organizational interactions in 2003 (Giusti-Rodríguez Reference Giusti-Rodríguez2023). Positive interactions are those that reveal productive collaboration among organizations. It includes coparticipation in meetings, public statements, negotiations, or protests where the outcome was a joint effort, agreement, or alliance. Positive interactions contrast with negative ones, which are those interactions where the outcomes signal an organizational break within or between organizations. Examples include internal disagreements or splits, failed negotiations, and statements negating mobilization support to organizations requesting it. In this study, I focus on positive interactions. The resulting organizational network has 51 organizational nodes and 255 ties (positive interactions) between these.Footnote ⁹ The network has a mean degree centrality of 10 and median degree of 5. I use network analysis to evaluate each organization’s tie density, centrality, and connectivity, and assess networks’ structural vulnerabilities.

The network dataset focuses on the 2003 protest cycles. In addition to this being the first year for which full newspaper articles are available, it also offers an effective context for evaluating structural resilience. The 2003 Gas Wars involved persistent protests and inter- and intra-organizational interactions that spanned the Bolivian territory, paralyzed the country on multiple occasions, and resulted in a presidential resignation. Most organizations associated with the ideological left and indigenous communities participated. Because of its intensity and reach, the 2003 protest cycle revealed extensive organizational activity and made visible numerous network ties and organizational actors that would otherwise be difficult to capture in newspaper data during periods of limited organizational activity.

The 2003 focus nonetheless raises several challenges. One concern is that, by this point, MAS-IPSP may have already been too far ahead of its counterparts. Both MAS-IPSP and MIP performed well in the 2002 elections, but MAS-IPSP obtained a significantly larger vote share. In that same election, ASP participated in an alliance with Nueva Fuerza Republicana (NFR), running its own candidates in regional elections (Véliz was elected to the legislative assembly for Cochabamba) and supporting the NFR candidate for president. Thus, the potential risk here is that 2003 is inadequate for comparing these network environments because, by then, MIP and ASP had fallen too far behind.Footnote ¹⁰ From a leadership and organizational standpoint, however, the mobilization dynamics of the 2003 protest cycle suggest that these remained viable political projects. The three party leaders and their organizations had prominent roles throughout the protests and displayed persistent mobilizational power. Quispe, for instance, carried out the hunger strike that sparked the Gas War protest cycle and remained a core leader throughout. Writing about Quispe and Morales in the aftermath of these protests, (Van Cott Reference Van Cott2005, 98) states that “the leaders of the two new ethnic parties, while maintaining their identity as outsiders, were firmly established as central political figures. In a conjuncture where political parties had been rendered almost meaningless […] all eyes were on Morales and Quispe.”Footnote ¹¹ The 2003 protest cycle, therefore, may well have been an equalizing moment, particularly for MIP and MAS-IPSP, giving them greater influence and political momentum. Beyond this, however, there is no evidence to suggest that network environments were altered significantly with the decline of these proto-parties.

Another concern centers on whether historical differences between organizations provided MAS-IPSP with a stronger and more cohesive organization and, in turn, shaped party-building outcomes. Existing evidence, however, indicates that the three organizations were equipped with impressive resources. During the period of analysis, CSUTCB was Bolivia’s most organizationally powerful and densely interconnected organization. Within CSUTCB, too, the La Paz and Cochabamba federations—the origin organizations of MIP and ASP, respectively—were the most powerful, cohesive, and mobilized branches. As top executives of CSUTCB, and leaders of their departmental federations, both Quispe and Véliz were arguably better positioned than Morales to access and mobilize a vast network structure. Morales himself led a powerful but drastically smaller, geographically concentrated organization that was located on a lower rung of CSUTCB’s organizational chart within Veliz’s FSUTCC. These dynamics highlight that the challenge for MIP and ASP did not stem from their organizational weakness or increasing irrelevance relative to MAS-IPSP, nor from their leaders’ lack of mobilization capacity. Rather, the challenge was in their inability to translate significant organizational strengths and resources into political support.

A final concern centers on endogeneity issues, specifically, on whether the network ties observed in 2003 were bolstered in the aftermath of MAS-IPSP’s 2002 electoral performance. While available data do not allow network analysis of earlier years, the mixed-methods approach enables me to examine various empirical expectations associated with this argument and related concerns. I structure the analyses and discussion to address these potential issues.

First, I use process tracing to identify the origins of network environments. This allows me to demonstrate that the most critical ties—and the main differences across network environments—were present well before proto-parties’ emergence. I show that this variation was unintentional and originated from formal organizational structures and established patterns of alliance development, rather than from transformations following the 2002 election. I also employ process tracing to examine how Coordinadora’s ties shaped MAS-IPSP’s expansion opportunities and leadership structures throughout the party’s early years. I find a strong association between historically institutionalized ties and patterns of party growth, consistent with the idea that proto-parties grew through pre-existing network ties. Third, I combine network analysis and process tracing to assess how competition shapes networks. I find significant and varied evidence in support of expectations both in pre-2003 patterns of organizational fractures and 2003 measures of networks’ structural vulnerabilities.

Finally, I also find important evidence consistent with the idea that inter-organizational cooperation was common throughout protest cycles and that political rivalries did not significantly decrease the likelihood of inter-organizational cooperation. Notably, Morales’ Coordinadora and CSUTCB, with Quispe and Véliz at its helm, cooperated—as did their leaders—consistently throughout the 2003 protests. Along with COB, Coordinadora and CSUTCB were the most important organizational nodes in the network and were strongly interconnected, both through direct and indirect ties. These varied insights demonstrate that Bolivia’s proto-parties did not generate their social networks, nor drastically altered patterns of inter-organizational cooperation. Instead, proto-parties sought to adapt to, and grow through, their pre-established network ties.

Resilience of Network Environments

The three political projects I examine—ASP, MAS-IPSP, and MIP—emerged from regional organizations embedded in CSUTCB. ASP was founded in 1995 and had its roots in FSUTCC, Cochabamba’s departmental indigenous peasants’ federation. Its leader was Alejo Véliz, then FSUTCC executive secretary. MAS-IPSP was based in Cochabamba’s Chapare and used the coca growers’ Coordinadora—of which Evo Morales was executive secretary—as its organizational base (Komadina and Geffroy Reference Komadina and Geffroy2007). Finally, MIP was founded by Felipe Quispe in 2000 and traced its origins to FDUTC-LP “TK,” La Paz’s indigenous peasant federation, which Quispe led. Table 1 summarizes the three parties’ organizational roots, founding years, and leaders.

Table 1. Indigenous Political Projects in Bolivia

Each of these organizations was embedded in a network environment—with distinct nodes and ties—that provided the landscape of opportunities for proto-parties’ expansion and consolidation. These network environments shared important similarities. The three parties emerged from regional or subregional branches of Bolivia’s powerful indigenous peasant organization, CSUTCB. Their ties to CSUTCB provided a critical pathway for expansion; they connected each organization into a vast organizational network that included institutionalized ties to local, subregional, and departmental CSUTCB nodes spread across Bolivian territory, along with national and subnational ties, resulting from conjunctural or stable alliances to organizations outside CSUTCB’s structure. Through CSUTCB, they were also connected into Bolivia’s workers’ confederation, COB, which provided proto-parties with a pathway into urban and working-class sectors.

Their network environments, however, also differed in important ways. One difference was that each organization was connected into the larger network from different departmental and subdepartmental levels. CSUTCB’s La Paz and Cochabamba federations have historically been the most important departmental CSUTCB branches. They have displayed consistent and extensive organizational capacity and made important intellectual contributions to the indigenous peasant movement (Chávez León and Costas Monje Reference Chávez León, Monje and Linera2005; Rivera Cusicanqui Reference Rivera Cusicanqui1984). MAS-IPSP’s Coordinadora, for its part, was a subdepartmental organization embedded within CSUTCB’s Cochabamba federation. The Coordinadora was much smaller and lower down in the organizational hierarchy than its counterparts. It was also newer (it consolidated in 1992). Nonetheless, it represented one of the most militantly organized sectors in Bolivia at the time. Figure 1 visualizes each organization’s position within CSUTCB’s structure.

Figure 1. CSUTCB Organizational Structure

A second key difference in network environments centers on origin organizations’ connectivity to other organizations. As regional CSUTCB branches, FSUTCC and FDUTC-LP “TK” relied heavily on CSUTCB to organize and mobilize. While both organizations had significant autonomy and strength to participate in events and join causes, as formal regional branches, their autonomous mobilizational efforts were generally limited to their particular regions. Ties to their network environment, and particularly to those organizations outside their own regions (even other CSUTCB regional branches), ran through CSUTCB. They were therefore both deeply connected into and strongly dependent on CSUTCB’s organizational structure. When new political parties emerged within these regional organizations, CSUTCB represented their main pathway to expansion.

Coordinadora’s network environment was different. Much like FSUTCC and FDUTC-LP “TK,” Coordinadora had an expansion pathway through CSUTCB by way of Cochabamba’s FSUTCC. Yet, in contrast to its counterparts, Coordinadora’s network environment was characterized by numerous ties that ran both through and outside CSUTCB and connected it to national and regional organizations in and outside Cochabamba.

The differences between Coordinadora’s and other organizations’ network environments were produced by two dynamics: Coordinadora’s historical development and its organizational culture of alliances. Coordinadora was founded in 1992 as an effort to bring together six different and independent coca growers’ federations that emerged throughout Chapare since the 1960s to organize the exponential growth in coca producers. These federations remained fragmented through the 1980s (Spedding Reference Spedding2005) and during this period, each one established organizational affiliations and institutional ties that embedded them into the indigenous peasant organizational landscape (see Appendix A). In the process, each federation became affiliated with a different regional and national indigenous peasant organization (Gutiérrez Aguilar Reference Gutiérrez Aguilar2008, 188). Two of the six local federations became affiliated with Cochabamba’s FSUTCC and, through this, with CSUTCB. The other four coca federations affiliated themselves to FSCC, Cochabamba’s interculturales federation, and departmental branch of the national interculturales’ confederation (CSCB) (Chávez León and Costas Monje Reference Chávez León, Monje and Linera2005; Spedding Reference Spedding2005).Footnote ¹²

Coordinadora thus emerged in 1992 as an umbrella organization for six previously independent federations to facilitate a more coordinated response to harmful state policies. Critically, however, when it formed, rather than canceling its member federations’ prior affiliations and establishing its own institutional ties, it maintained the institutional status quo and remained integrated into both FSUTCC and FSCC in Cochabamba and, through them, to CSUTCB and CSCB structures (Appendix A). These two national indigenous peasant organizations were also independently connected to COB (see Figure 2).

Figure 2. Coordinadora’s Network Environment

Thus, in marked contrast to FSUTCC and FDUTC-LP “TK,” Coordinadora’s network environment had multiple institutional ties connecting it to regional and national indigenous peasant organizations and the workers’ confederation. This historical development reduced Coordinadora’s dependence on any given organizational node, provided it with multiple pathways to the national arena, and afforded it with a more resilient network environment.

Critically, the key elements of this network environment became defined with the consolidation of Coordinadora in 1992 and predated the formation of MAS-IPSP in 1999. They were also an unintentional byproduct of fractured processes of organizational development. Indeed, this unusual structure was perceived as “very problematic” by Morales, then Coordinadora’s Executive Secretary (Chávez León and Costas Monje Reference Chávez León, Monje and Linera2005, 391). Yet, it was this precise structure that provided Coordinadora with a more resilient network environment than its counterparts. The fact that CSUTCB and CSCB had different geographical strengths and weaknesses also meant MAS-IPSP could rely on different regional ties to grow into different regions.

The second factor that contributed to Coordinadora’s different network environment was the organization’s pattern of engagement with other organizational nodes outside its institutional structure. Both Coordinadora and CSCB have been particularly apt historically at establishing stable ties with other organizations. They achieved this by collaborating closely with, and often mobilizing for, other organizations. CSCB, for instance, had a strong and stable collaboration with CONAMAQ and worked consistently with powerful region-level organizations such as Santa Cruz Ethnic Communities’ CPESC (Chávez León and Costas Monje Reference Chávez León, Monje and Linera2005). Coordinadora, for its part, often engaged in protests supporting other organizations and policy demands outside their realm and region (Gutiérrez Aguilar Reference Gutiérrez Aguilar2008); through this, they grew and solidified connections in their network environment. Examples of this include their participation in the 2000 Water Wars, where Coordinadora developed a strong and durable alliance with Cochabamba’s Water Coordinator and the irrigators’ federation—the “main mobilizational force of the Cochabamba valley region”—among others (Chávez León and Costas Monje Reference Chávez León, Monje and Linera2005). Ties with working-class sectors were established through alliances with workers’ departmental CODES, six departmental federations within COB (La Voz 2003; Opinión 2003). In 2001, Coordinadora also integrated COMUNAL (which became Estado Mayor), an alliance originally established with other Cochabamba sectors.

Between 2003 and 2005, Coordinadora also mobilized to support protests organized by FDUTC-LP “TK,” CSUTCB, FEJUVE-El Alto, COR-El Alto, Huanuni miners, and the Gas Coordinator. Because of its tremendous mobilizational capacity, Coordinadora’s participation in protests enabled movements—often based in La Paz—to gain national interconnectivity and visibility. Significantly, however, these were instances of Coordinadora actively throwing its mobilization capacity behind other actors, often outside Cochabamba, in a manner consistent with its organizational culture.Footnote ¹³

Coordinadora and CSCB’s tie-building culture contrasts somewhat with FDUTC-LP “TK” and FSUTCC’s. Partly due to their structural position within CSUTCB, both organizations tended to invest less in the establishment of stable inter-organizational alliances outside their core structure and, especially, outside their region. Their independent ties as departmental federations were generally with other intra-departmental organizations and focused on departmental issues. When they supported causes outside their regions, they tended to do so through the national CSUTCB. CSUTCB, for its part, while powerful and deeply interconnected has nonetheless had challenges establishing stable ties outside its own core network environment (Chávez León and Costas Monje Reference Chávez León, Monje and Linera2005, 211). Its significant organizational strength may have limited its need to invest in establishing stable alliances with organizational counterparts.

In sum, despite important similarities in their societal and organizational foundations, the three proto-parties nonetheless became embedded in distinct organizational network environments that shaped their dependency on CSUTCB’s structure and connectivity to other organizations. These differences were rooted in historical processes and organizational cultures that ultimately limited the pathways connecting FSUTCC and FDUTC-LP “TK” to their organizational networks while also providing the Coordinadora with greater interconnectivity and structural resilience.

The analysis of these organizations’ network environments in 2003 is consistent with these historical and structural dynamics. Figure 3 uses network data to explore variation in network environments across these organizations. It visualizes the network environments of (a) Coordinadora, (b) FDUTC-LP “TK,” and (c) FSUTCC, highlighting their direct ties to organizational nodes in the network. The figure shows that the three organizations are connected into the same organizational community, but their positions and connectivity within this network vary substantially. Whereas Coordinadora is strongly interconnected, FDUTCP-LP “TK” and FSUTCC have connections with important nodes but have fewer ties to the organizational network.

Figure 3. Network Environment of Indigenous Parties’ Origin Organization

Note: Figure plots (A) Coordinadora, (B) FDUTCP-LP “TK,” and (C) FSUTCC’s network environments. Small circles represent organizational nodes. The gray lines indicate organizational ties, capturing positive interactions. Dark lines highlight direct ties connecting each proto-parties’ origin organization to larger organizational community.

Furthermore, while during the period of analysis, all three organizations were deeply engaged in mobilizations, meetings, and negotiations, Coordinadora did so directly—without having to go through CSUTCB—and with a more diverse set of actors. During this year, its densest ties (with multiple lines) were established with national organizations such as CSUTCB, COB, MST (Landless Peasants), and Estado Mayor, and with regional ones like COFECAY and COR (El Alto regional workers), both in La Paz. Other ties with major national organizations such as CSCB, rural and urban teachers’ unions, and regional organizations including FSUTCC and the Cochabamba, Oruro, and La Paz departmental workers’ federations were also evident. This robust network environment connected Coordinadora to organizations with national and regional presence that extended well beyond Cochabamba and into other departments, reaching diverse social sectors.

In marked contrast, FSUTCC and FDUTC-LP “TK”’s network environments were limited and more regionally grounded. During the period of analysis, FSUTCC established direct ties with six organizations, including Coordinadora, CSUTCB, and Estado Mayor, and La Paz and Cochabamba’s departmental workers’ federations. FDUTC-LP “TK” engaged with seven organizations including CSUTCB, COB, the rural teachers’ union, El Alto university teachers’ union, COR-El Alto, transportation carriers, and Warisata community. While these ties embedded both FSUTCC and FDUTC-LP “TK” in a remarkable organizational landscape, the analyses reveal significant differences across network environments, with Coordinadora’s being significantly more interconnected than its counterparts, being less dependent on particular connecting nodes, and having more diverse pathways to expansion. Notably, the variation in connectivity level did not imply significant differences in organizational activity. As the most important regional branches of CSUTCB, both FSUTCC and FDUTC-LP “TK” were highly mobilized throughout this period. But, because of their institutional ties and structures, their mobilization activity was generated within and through CSUTCB. These dynamics are consistent with the three organizations’ structural position and historical patterns of development.

Various measures of node centrality confirm Coordinadora’s greater interconnectedness. Node centrality measures reveal the importance of organizational nodes within the network structure. There are three main node centrality measures: degree, eigenvector, and betweenness.Footnote ¹⁴ Degree centrality captures the number of edges adjacent to a node; by counting how many direct ties nodes have to the network, the measure reveals information about node connectivity and importance. Eigenvector centrality incorporates data about other nodes’ importance and calculates centrality by weighing more important nodes more heavily than more marginal ones (Bonacich Reference Bonacich1987). The more connected a node is to other important nodes, the higher its eigenvector score. The final measure, betweenness, captures the number of times a node lies in the shortest path between other nodes. It reflects other nodes’ dependence on a node for communication (Ward, Stovel, and Sacks Reference Ward, Stovel and Sacks2011).

Table 2 presents the results of the centrality analyses. Across measures, Coordinadora stands out as a more interconnected and central organizational node. It has significantly more network ties (degree centrality); Coordinadora’s 43 ties contrast with FDUTC-LP “TK”’s nine and FSUTCC’s six ties. When weighted for connectivity to central nodes (Eigenvector), it is approximately five times more central to the network than FDUTC-LP “TK” and eight times more so than FSUTCC. And when we consider the extent to which each organization serves to connect other organizations (Betweenness), Coordinadora again stands out for having a substantially higher betweenness value than its counterparts, meaning that other organizations regularly relied on Coordinadora to reach other nodes. This is consistent with Coordinadora’s alliance culture, which often drove it to mobilize behind other groups and bring actors together through sustained coordination (Chávez León and Costas Monje Reference Chávez León, Monje and Linera2005; Komadina and Geffroy Reference Komadina and Geffroy2007).

Table 2. Proto-Parties’ Origin Organizations’ Centrality Measures

By contrast, FDUTC-LP “TK” and FSUTCC had fewer direct ties to other organizations in the network and most connections were with regional actors. These ties maintained the two organizations in a marginal position within the network, despite being the two most powerful regional CSUTCB branches. This highlights the difference between organizational capacity and structural resilience. The two departmental federations were critical mobilizational forces—arguably CSUTCB’s muscles—during the 2003 protest cycle. And yet, they remained weakly connected and marginal within the network, relying heavily on CSUTCB to access other organizations. Coordinadora would have shared a similar structural vulnerability had it relied on its FSUTCC ties to connect to the national arena. But, because of its history and culture of alliances, it had greater network connectivity, both through and outside CSUTCB’s structure.

Collectively, these outcomes reveal critical and, thus far, overlooked differences in the three organizations’ network environments. Specifically, they reveal Coordinadora’s significant structural resilience and FDUTC-LP “TK” and FSUTCC’s contrasting weaknesses and structural vulnerabilities. Because CSUTCB represented their only feasible pathway to their network environment, any threat to this tie meant these organizations could become isolated, undermining their value as channels for party-building. In the next section, I evaluate how the emergence of the three projects within CSUTCB undermined this pathway’s utility and impacted proto-parties’ expansion prospects.

Fracturing CSUTCB: Political Parties Competing for Control

As an organization gathering Bolivia’s indigenous peasantry with enormous mobilization capacity and significant community-level presence, CSUTCB offered an ideal network for the growth and establishment of an indigenous party (Chávez León and Costas Monje Reference Chávez León, Monje and Linera2005). The powerful organization offered proto-parties born within its structure potential access to extensive resources, mobilization capacity, and strong partisan foundations (Anria Reference Anria2018; Poertner Reference Poertner2021). It also offered an institutionalized pathway to COB, which would enable proto-parties to expand beyond indigenous rural areas and into urban and working-class sectors.

Figure 4 illustrates CSUTCB’s critical position in the network environment. Figure 4a visualizes the network environment using degree centrality—which counts the number of edges that are node-adjacent—to determine node size (larger nodes have higher values). Relevant nodes—CSUTCB, COB, Coordinadora, FSUTCC, and FDUTC-LP “TK”—are in white (others are in gray). The figure shows that CSTUCB has the greatest number of direct ties to other organizational nodes, surpassing even COB and Coordinadora, the two other nodes with distinctly high degree centrality. Figure 4b plots CSUTCB’s direct ties (black) and underscores CSUTCB’s value by displaying its interconnectivity to other important nodes. Collectively, the figures highlight that CSUTCB was the most important node in the organizational network.

Figure 4. Organizational Network and CSUTCB Connections

Note: Figure (A) represents the complete organizational network, with important nodes in white circles and node size determined by degree centrality. Figure (B) plots this network, highlighting (in black) CSUTCB’s direct ties.

I further evaluate CSUTCB’s position within the network using bridge centrality metrics, which are community-centric measures that assess how important nodes are to the communication between different organizational communities. The greater a node’s bridge centrality value, the greater its role connecting node communities to each other and enabling communication across them. Bridge centrality arguably offers proto-parties distinctly valuable interconnectivity because it provides access into different organizational communities.

Figure 5 plots three bridge centrality measures. These capture the number of ties connecting each node to nodes outside its community (Strength), the number of times each node connects nodes in different communities (Betweenness), and the distance between a node and every node outside its organizational community (Closeness) (Payton Reference Payton2023). The results further evidence CSUTCB’s value, demonstrating that it serves as a critical organizational bridge. CSUTCB has 31 ties connecting it to nodes in other organizational communities (strength), 6 more than COB, and substantially more than all other organizations. CSUTCB also stands out as the node with the highest bridge betweenness; it is on the shortest path between nodes in different communities 630 times, again outperforming COB (577) and drastically surpassing others (all below 200). Bridge closeness measures are consistent with this assessment: CSUTCB has the lowest average distance connecting it to all nodes outside its community (closeness measure reflects inverse). Having access to CSUTCB and, through it, to COB should have therefore been a strong predictor of party-building success, precisely because, through CSUTCB, proto-parties could reach most corners of the network environment.

Figure 5. Bridge Centrality of Organizational Nodes

Note: Bridge strength is calculated as the absolute value of edges between a node and all nodes outside of that node’s community. Bridge betweenness considers the number of times a particular node is in the shortest path between two nodes in different communities. Bridge closeness is calculated as the inverse of the average distance from a particular node to all nodes outside its own community. Organizational communities are defined through random walks.

All three proto-parties recognized CSUTCB’s value for party-building and had access to it through their origin organizations and key leaders (Burgoa Moya Reference Burgoa Moya2016; Chávez León and Costas Monje Reference Chávez León, Monje and Linera2005). But, precisely because they recognized CSUTCB’s value, proto-parties also fought fiercely for control of this organizational space (Burgoa Moya Reference Burgoa Moya2016; Quispe Reference Quispe2003). Leaders’ deep knowledge of CSUTCB meant they were all effective, at least, at attacking and canceling this critical node within the organizational network. Ultimately, rather than serving as a direct path to expansion and consolidation, CSUTCB instead became internally fractured.

The first fracture appeared within FSUTCC in 1997 and resulted from a split between Véliz—then FSUTCC executive secretary and ASP president—and Morales (García Yapur, García Orellana, and Soliz Romero Reference García Yapur, Orellana and Romero2014; Komadina and Geffroy Reference Komadina and Geffroy2007; Quispe Reference Quispe2003). The fracture boiled to the national CSTUCB when Véliz and Morales’ supporters engaged in physical confrontations at the 1998 CSUTCB Congress (Quispe Reference Quispe2003). Two competing parties, both from Cochabamba, emerged from this fracture. Véliz kept the ASP name while Morales founded MAS-IPSP. This split created an intense division within FSUTCC and CSUTCB—which had a former FSUTCC executive secretary at its helm (Rasguido Reference Rasguido2006)—as the two proto-parties sought to assert their influence at the regional and national levels (Albó Reference Albó2009).

After the 1998 CSUTCB Congress confrontation, COB intervened as mediator, calling for a CSUTCB Unity Congress. Quispe was elected to lead CSUTCB as the compromise candidate between the two political factions (García Yapur, García Orellana, and Soliz Romero Reference García Yapur, Orellana and Romero2014; Gutiérrez Aguilar Reference Gutiérrez Aguilar2008). Quispe, however, while not siding with ASP (and initially showing some support for Morales’ project), quickly began to challenge MAS-IPSP’s expansion within CSUTCB.

Network attacks by the different projects took various forms. In 1999, for instance, MAS-IPSP supporters convinced nongovernmental organizations that assisted CSUTCB financially to suspend their aid; as a result, CSUTCB had to stop paying its employees and had its electricity and phone lines cut (Gutiérrez Aguilar Reference Gutiérrez Aguilar2008).Footnote ¹⁵ That same year, MAS-IPSP supporters attempted to force Quispe’s resignation from CSUTCB and the installation of a Morales ally as executive secretary. The following year, during an explosive social conflict—which the Coordinadora and FDUTC-LP “TK” spearheaded—Quispe negotiated with the government and then “left Morales alone with his blockades, even though [Morales] had acted in solidarity with CSUTCB’s leader. This forced [Morales] to negotiate with the government alone and from a weak position. Soon after, Morales disavowed Quispe’s leadership and organized a parallel organizational direction” (Komadina and Geffroy Reference Komadina and Geffroy2007, 47).

This generated the second fracture, which led to the formation of parallel CSUTCB organizational structures, both claiming legitimacy. One bloc brought Quispe and Véliz together in a hostile alliance motivated by their interest in limiting MAS-IPSP influence within CSUTCB and advancing alternative political projects (ANF 2001). The other bloc was controlled by MAS-IPSP supporters, who were also looking to assert control of the organizational space and deny access to their counterparts. CSUTCB had splintered along political lines and, while the three proto-parties competed fiercely, even physically, for its control, none of them could successfully achieve full control of the organization. By the time Quispe founded MIP in November 2000, he was leading an internally fractured CSUTCB which, in 2001, split into two factions that would continue operating separately for some time (Correo Del Sur 2003).

The CSUTCB factions that emerged placed a cap on proto-parties’ capacity for expansion. The resulting parallel organizations reflected proto-parties’ spheres of influence, but also delineated the limits of their expansion opportunities through CSUTCB. The Morales-allied CSUTCB bloc included Santa Cruz, Chuquisaca, Tarija, Pando, Potosi, and parts of Cochabamba and Oruro’s departmental federations. Quispe and Véliz’s projects had control over the La Paz and Beni federations, parts of the Cochabamba and Oruro ones, and several provincial federations (Chávez León and Costas Monje Reference Chávez León, Monje and Linera2005, 124). These alliances are represented in Figure 6, which maps the geographical distribution of proto-parties’ CSUTCB control.

Figure 6. CSUTCB Departmental Federations’ Political Alliances

Note: Figure uses 2001 Bolivian census data (INE 2001) to map proto-party political control in rural municipalities by identifying municipalities that are over 75% rural—a proxy for CSUTCB reach—and assigning political control to each party based on known power coalitions among department federations.

Consistent with expectations, then, the competition that ensued for control of CSUTCB led it to fracture internally along political lines and canceled it as a feasible pathway for party expansion and consolidation.

How did the collapse of the CSUTCB pathway impact proto-parties’ network environments? To answer this, I first evaluate the network environment’s vulnerability to the removal of particular nodes. Figure 7 examines the implications of intentional network attacks and random breakdowns and reveals what happens to the network when an increasing number of organizational nodes are canceled out (Lhomme Reference Lhomme2015). I am particularly interested in the betweenness, degree, and cascading measures included in the plot because these capture what happens to networks when competing projects emerge within critical nodes. Note that all three proto-parties recognize the value of CSUTCB and COB as essential pathways to consolidation. As such, they are not only looking to grow through these organizational spaces, but also to impede other parties’ access to, and control of, them. Efforts at undermining other proto-parties’ access to CSUTCB’s structure represent instances of intentional network attacks. The betweenness, degree, and cascading measures reveal how “attacks” on critical nodes impact networks. They collectively illustrate how quickly the network loses connectivity when we remove the most important nodes within it. Results show that, despite the network’s many organizational actors, it is highly reliant on the connectivity provided by a small number of nodes. Targeted attacks aimed at these nodes can quickly undermine network connectivity.

Figure 7. Attack Tolerance of Organizational Network

Note: Plot assesses network vulnerability to intentional or random attacks. Each line captures connectivity loss resulting from the removal of individual nodes randomly (Random) or according to levels of degree centrality (Degree), degree betweenness (Betweenness), and recalculated degree betweenness after node removal (Cascading).

The removal of critical nodes, however, has drastically different consequences for the various proto-parties. For MIP and ASP, the implications of attacks that canceled access to CSUTCB and COB were severe: not only did they lose access to the most important organization, but the network that remained was now dominated by Coordinadora and its Estado Mayor inter-alliance network, MAS-IPSP’s foundational platforms. The organizations’ few remaining ties provided poor conduits for party-building. In the case of FDUTCP-LP “TK,” three of its five remaining ties were to La Paz-based organizations. FSUTCC was left even more weakly positioned, with two of its five remaining ties being to Coordinadora and Estado Mayor.

With CSUTCB fractured, FSUTCC and FDUTCP-LP “TK” lost connectivity to their organizational network environment and this undermined ASP and MIP’s expansion prospects. The lack of pathways constrained these projects to their bases. For ASP, this was particularly challenging; unable to establish full control of Cochabamba’s peasant federation, ASP failed to grow beyond that partial base. It obtained significant electoral support in rural Cochabamba in the 1997 presidential and 1999 municipal elections but failed to grow beyond this area and eventually collapsed.Footnote ¹⁶ MIP, for its part, successfully established political control of FDUTCP-LP “TK.” Geographically, this gave the party a strong base in rural La Paz, where it obtained over 60% in some municipalities in 2002 (Molina Reference Molina, Orellana and Yapur2010; Romero Ballivián Reference Romero Ballivián2003, 309). After a promising electoral showing, however, the proto-party proved unable to grow beyond La Paz. In the 2005 presidential election, it obtained 2.15% and, soon thereafter, collapsed.

MIP’s collapse was not for lack of strategy. MIP actively sought to build alliances that would provide it with pathways into other organizational and societal sectors. For instance, in the 2002 election, Quispe sought to run on the same ticket, first with Morales and then with Domitila Chungara, a powerful symbol of Bolivia’s workers’ movement. When neither of those alliances consolidated, he chose an intellectual Quechua woman: “[Quispe] was very interested in finding an electoral partner in Cochabamba for two reasons: first, that is the territory of Evo Morales and Alejo Véliz […]; second, because of the Quechua population” (La Prensa 2002). Each of these vice-presidential candidate choices reveal what Quispe recognized as his representational gaps and organizational paths (coca growers, urban working classes, and Quechua populations [Quispe being Aymara], respectively). In the 2005 election, we would see a similar logic drive Quispe’s attempts, first, at establishing an electoral ticket with Morales (La Voz 2005); when this failed, at having the COB executive secretary as his running mate (La Razón 2005a); and when that strategy, too, failed, running with an indigenous woman from El Alto that had founded the indigenous peasant women’s confederation. MIP’s structural environment helped define both the strategic alliances it pursued—Coordinadora and COB—and the outcome of these efforts. Ultimately, despite Quispe having played a leading role during the 2003 protest cycles as CSUTCB’s executive secretary, he lacked the organizational ties necessary to convert popular support in the streets into electoral votes, much less committed partisans. After the 2005 election, unable to escape the boundaries of rural La Paz and nationalize its reach, MIP collapsed.

By contrast, absent access to CSUTCB as a bridge to expansion, Coordinadora remained remarkably well-positioned to facilitate MAS-IPSP’s growth and consolidation. Coordinadora sustained ties to the powerful CSCB—which had its strongest presence in La Paz, Cochabamba, and Santa Cruz—and, through CSCB, to COB. Coordinadora also sustained ties to regional organizations within Cochabamba (e.g., the influential factory workers’ union and other major working-class organizations), La Paz (e.g., COFECAY), and other departments (e.g., Santa Cruz’s CPESC). Crucially, many of these ties were institutionalized—whether formally (CSCB, for instance) or informally through a long and stable relationship of mutual support (e.g., COFECAY). The structural resilience of Coordinadora’s network environment enabled MAS-IPSP to overcome the attacks from competing parties and continue its expansion through alternative pathways.