INDIGENOUS VOICES AND THE DEFEAT OF THE NEW CONSTITUTION

This section lists and groups different explanations related to indigenous affairs offered by indigenous personalities after the September 4, 2022, plebiscite. I include views expressed in the press by the indigenous representatives elected to the reserved seats for the 10 legally recognized indigenous peoples in the Constitutional Convention,Footnote ³ as well as statements by other indigenous public figures.

A prevalent explanation underscores the importance of the Mapuche conflict. Violence during the process symbolized resistance to institutional solutions, influencing some to vote against the draft (Pairican Reference Pairican2022). Hugo Alcamán, who campaigned for the “Reject” option, stated that “the Mapuche people is against violence, as it demonstrated in the plebiscite” (Soto Reference Soto2022). Convention candidate Pedro Cayuqueo (Reference Cayuqueo2022) noted how the conflict caused disillusionment with political pathways among the Mapuche. Héctor Llaitul, leader of one of the main Mapuche militant organizations, criticized plurinationality for being “too statist and gradual” (Bauer Reference Bauer2022, 437). Representative Elisa Loncon pointed out that the Mapuche militant sector’s refusal to dialogue favored rejection (Cooperativa 2022). Additionally, others believe that indigenous representatives did not adequately condemn violence, distancing voters (Millaleo Reference Millaleo2022).

Another perspective highlights a disconnect between indigenous representatives and voters. The constitutional debate became elitist, using academic and technical language (Millaleo Reference Millaleo2022). Some claimed that “Plurinationality and other demands would only be the ideas of a handful of disconnected ÑuñoaFootnote ⁴ intellectuals who are out of touch with the real world” (Cayuqueo Reference Cayuqueo2022). Alvarado Lincopi and Llancaman (Reference Alvarado Lincopi and Llancaman2022) stressed the role played by “the lack of political instruments that link organizational and territorial processes with institutional and democratic political disputes.” Luis Penchuleo, director of Corporación Nacional de Desarrollo Indígena (CONADI),Footnote ⁵ emphasized the need for better political education and campaigning (Oviedo Reference Oviedo2022). Ariel León Bacián (Reference León Bacián2022), former Quechua legislative advisor to the Convention, also highlighted lackluster campaigning in indigenous territories.

Others believe that the content of the constitutional draft did not align with indigenous voters’ preferences. Yaghan representative Lidia González suggested that the negotiation process resulted in the inclusion of provisions that did not represent the people (La Prensa Austral 2022). Loncon discussed the influence of Western thought within indigenous communities, which distanced them from the draft’s principles (Biobio Chile Reference Chile2023). For some indigenous citizens, voting Reject might have been a way to assert their Chileanness (Cayuqueo Reference Cayuqueo2022), which Loncon mentioned as potential “indigenous whitening” (Huenchumil Reference Huenchumil2022). These testimonies align with survey data from Southern Chile (Centro de Estudios Públicos 2022), which indicates that the majority of Mapuche respondents identify either exclusively as Chilean or as both Chilean and Mapuche. Moreover, most express a preference for a multicultural state or one without distinctions among peoples rather than a plurinational state.

Other explanations revolve around the characteristics of indigenous communities. Javier García, the Aymara mayor of Colchane, noted that issues affecting his community (e.g., migration) were not addressed in the process (Mayorga Reference Mayorga2022). Aymara representative Luis Jiménez highlighted the Aymara’s conservative nature, preferring gradual change (Miranda Reference Miranda2022). Former governor Richard Caifal (Reference Caifal2022) also stated that the Mapuche favor less disruptive changes. Likewise, historian Sergio Caniuqueo argued that Mapuche society is socially conservative, which collided with the progressive constitutional text—particularly among Mapuche evangelicals (Miranda Reference Miranda2022). The presence of long-standing patron–client relationships with political parties in the Reject camp also could have had a decisive impact on the indigenous electorate (Cayuqueo Reference Cayuqueo2022). By contrast, in the case of Rapa Nui, the Convention’s text recognized it as a special territoryFootnote ⁶ and acknowledged the Agreement of Wills with the Chilean State (1888), which was considered “dignifying” by Pedro Edmunds Paoa, the island’s mayor (Gálvez, Canales, and Mayorga Reference Gálvez, Canales and Mayorga2022).

A DISTRICT-LEVEL ANALYSIS OF THE REJECT VOTE

To understand the link between indigenous peoples and Chile’s 2022 plebiscite results, I analyzed data from Chile’s 345 municipalities (comunas) using official sources (Instituto Nacional Electoral 2018; Servicio Electoral 2022) and protest-event data (Centro de Estudios de Conflicto y Cohesión Social 2020). This analysis sheds light on the indigenous dynamics in the outcome of the constitutional proposal (Disi Pavlic Reference Disi Pavlic2023b).

The bivariate analysis indicates that indigenous regions primarily rejected the constitutional draft (figure 1). The highest rejections occurred in Colchane, Camiña, and General Lagos, the three northern municipalities with significant Aymara populations (Disi Pavlic Reference Disi Pavlic2023a). A simple linear regression confirms this correlation’s statistical significance.

Figure 1 Reject Vote by Percentage of Indigenous Population

The analysis supports indigenous leaders’ claims that indigenous-related conflicts influenced the plebiscite. Districts with the most conflicts—Ercilla, Contulmo, and Lumaco—showed Reject rates of 72.6%, 81.2%, and 80.5%, respectively (figure 2). However, other variables should be considered. Given the intersection of ethnicity with variables such as poverty (Agostini, Brown, and Roman Reference Agostini, Brown and Roman2010) and electoral preferences (Fuentes, De Cea, and Teitelboim Reference Fuentes, De Cea and Teitelboim2019), controls are crucial.

Figure 2 Reject Vote by Number of Violent Protests with Indigenous Claims per 1,000 Inhabitants (2008–2021)

I used weighted linear (WLS) and generalized linear models (WGLM) with robust standard errors to evaluate the relationship between indigenous variables and plebiscite voting.Footnote ⁷ Indigenous populations were disaggregated by the 2017 census categories (i.e., Mapuche, Rapa Nui, Aymara, Atacameño, Quechua, Colla, Diaguita, Kawésqar, Yahgan, and OthersFootnote ⁸), acknowledging differing political inclinations across communities. The controls used were local mean years of schooling, population size, poverty and rural population rates, and share of votes for right-wing candidate Jose Antonio Kast in 2021, which reflected partisan stability in preexisting preferences in referendums (Leduc Reference Leduc2002). A measure of protests during the 2019 “social outburst” (Sáez-Vergara, Godoy, and Disi Pavlic Reference Sáez-Vergara, Godoy and Disi Pavlic2022) also is considered due to its association with preferences in the plebiscite (Disi Pavlic Reference Disi Pavlic2023a, 112). Figure 3 illustrates the regression coefficients of the indigenous variables in the WLS and WGLM models.

Figure 3 Weighted Least Squares and WGLM Regressions of Reject Vote by Violent Protests with Indigenous Claims (2008–2021) and Percentage of Indigenous Peoples, with 95% Confidence Intervals

Regression results counter the bivariate analysis and the testimonies of indigenous figures, which suggests a negative link between indigenous violence and the Reject vote. Each additional violent event corresponds to an approximate 1.1-percentage-point decrease in Reject votes. Individually assessing indigenous groups also adds complexity: a negative association between indigenous populations and Reject support was evident for both the Mapuche and the Rapa Nui.Footnote ⁹ This suggests, as some indigenous personalities noted, that factors beyond indigenous identity—such as preexisting political preferences, community size, and education—may have influenced indigenous districts against the draft. However, the positive correlation among the Diaguita and Kawésqar indicates that some indigenous groups opposed the draft, giving credence—for at least some groups—to the argument that the constitutional draft was not valued or understood by indigenous voters.

Most indigenous people reside in urban zones, but their presence in rural areas also is significant. The implications of this divide are politically consequential (Aravena Reference Aravena2003; González Reference González2021). The split also is important in economic terms because the socioeconomic exclusion of indigenous peoples is explained partly by rural residence (Valenzuela, Toro, and Rojo Reference Valenzuela, Toro and Rojo2017). A WLS model was used with an interaction term between the shares of the indigenous and rural populations. The predicted Reject-vote values were estimated across the observed ranges of indigenous population and for selected percentages of rural population (figure 4).Footnote ¹⁰ The results show how the Reject vote’s association with the indigenous population depends on rurality. The association was negative in rural areas but slightly positive in highly urban municipalities, which suggests that the constitutional draft resonated more with some indigenous voters in traditional settings than with the majority residing in cities. This result indicates that the claim that indigenous voters did not approve of the draft has more support in urban areas.

Figure 4 Adjusted Predictions of Reject Vote by Indigenous and Rural Populations