Introduction

Participation and representation are fundamental pillars of democratic governance. Reflecting on these democratic ideals, countries across the globe have undertaken concerted efforts to enhance the political representation of historically marginalized communities (Kapiszewski et al., 2021; Echeverri-Pineda, 2018; Van Cott, 2007). One of these efforts materialized in creating special districts for selecting reserved seats for historically oppressed groups (Htun, 2016; Echeverri-Pineda & Villamizar-Chaparro, 2018). However, participation levels vary widely within these districts, suggesting that not all eligible voters choose to cast their ballots in them.Footnote 1 In this paper, we examine which factors explain why some people choose to vote in ethnic districts for reserved seats.

This paper posits that district choice is contingent upon individuals’ level of group consciousness. This dynamic underscores the significance of the politicization of identity as an incentive driving political engagement. Specifically, individuals exhibiting heightened group consciousness will prefer to vote in ethnic districts to advocate for enhanced representation and the preservation of such representation. Furthermore, our results illustrate that membership in ethnic organizations does not generate preferences for voting in ethnic districts. We argue that this results from a combination of organizational, institutional, and individual factors that generate incentives for members of these organizations to participate in other districts.

We concentrate on the case of Colombia’s reserved seats for Blacks in the House of Representatives. These seats were created in 1991 to enhance the political participation of the Black community. The two reserved seats are selected in a national district, which entails that any Colombian can choose to vote for them at the expense of their vote in their territorial district for the same political body. Furthermore, rules established that traditional political parties cannot have candidates running for these seats, and only organizations that have proven continuous work with Afro-Colombian communities may do so (Villamizar-Chaparro & Echeverri-Pineda, 2022).

To test our argument, we use a mixed-methods approach where we combine a subgroup analysis of a forced-choice conjoint experiment with a series of semi-structured interviews. The conjoint was embedded in a survey with a convenience sample of 1,200 Afro-Colombians. In this experiment, individuals were asked to choose between two hypothetical districts, and we analyze how group consciousness and organizational membership affects the choice of an ethnic district. We complement this quantitative dimension of our research with semi-structured interviews with Afro-Colombian political elites. The insights gained from these interviews serve to capitalize on the wealth of knowledge these key informants possess regarding the intricacies of electoral politics within the ethnic district.

This article provides several contributions. Previous work outside the United States has tended to focus on the role of the elite’s race in political processes (Bueno & Dunning, 2017; Janusz, 2022), and how racial fluidity and racial identification alters political preferences (Johnson, 2020; Mitchell, 2010). In this paper, we provide micro-level evidence of how an individual’s group consciousness affects district choice. Second, our findings further our understanding of the debate about the contextual role of group consciousness in ethnic politics (Smith et al., 2023; Mitchell, 2018) in a context where race is more fluid. Lastly, our findings dialogue with the scholarship on membership to ethnic organizations and political preferences (DellaPorta & Diani, 2006; Poertner, 2021; Madrid, 2005), by indicating that the link between membership in these entities and political choice is more complex and needs to account for the agency of the individuals, the organizational dynamics and the institutional incentives were these organizations operate.

Demands for Representation, Ethnic Districts, and Preferences for Inclusionary Institutions

According to Urbinati and Warren (2008), territorial residence has been the fundamental condition for political representation. Studies regarding identity and representation in geographic districts have shown that voters’ preference for local candidates is an expression of their place-based social identity (Schulte-Cloos & Bauer, 2021), and voters expect that a local candidate will be a legitimate representative of the interests of their local community (Campbell et al., 2019). However, this mode of representation has been questioned by historically marginalized groups who have struggled to become part of decision-making bodies (Mansbridge, 1999; Phillips, 1998). On this search for both descriptive and substantive representation, countries worldwide have created inclusionary institutions to promote political representation and participation of traditionally excluded groups (Htun, 2016; Paredes & Došek, 2020; Protsyk, 2010). One of these types of institutions is the electoral ethnic district.

We define an ethnic district as an electoral district whose constituency is built around descriptive traits related to ethnic identity. These institutions are intended to further the descriptive representation of a determined ethnic group by selecting the representatives of some reserved seats. Additionally, these districts allow ethnic groups to have a voice and be able to vote in legislative arenas with some autonomy from the existing party system (Htun, 2004). At the same time, they include the interests of minority groups within the legislative branch (Broockman, 2013; Griffin & Keane, 2006).

Ethnic districts are sometimes considered measures of affirmative action that could improve the regime’s legitimacy because there are more members in the decision-making process; because they promote group empowerment, and because they foster increased political engagement (Mansbridge, 1999; Pitkin, 1967). However, despite the political advancements brought by these institutions, the performance of some of them is modest compared to the higher goals of equality and better representation from which they were created (Godfrank, 2021). Some recent scholarship has highlighted the fact that these institutions fail to promote accurate representation for different ethnic groups (Htun, 2016; Paredes & Došek, 2020), help reinforce racial hierarchies (Telles, 2014; Krook, 2008), have been co-opted by traditional sectors (Roberts, 2021), and cannot by themselves solve all the problems around demands for more representation (Echeverri-Pineda & Villamizar-Chaparro, 2018).

Group Consciousness, Organizational Membership, and District Choice

We argue that choosing to participate in an ethnic district is correlated with two factors. Firstly, individuals with a stronger sense of group consciousness are more likely to vote within these institutions aimed at promoting the descriptive and substantive representation of their own ethnic group. Secondly, we argue that individuals affiliated with ethnic grassroots organizations are more inclined to participate in ethnic districts, as these organizations offer greater information and actively encourage member participation in these political arenas. We elaborate on these concepts below.

Group Consciousness and District Choice

We define racial group consciousness as the feeling and awareness that a person’s identity is intertwined with a disadvantaged racial minority and wants to achieve some explicit racial-group-based objective (McClain et al., 2009; Smith et al., 2023). According to Miller et al. (1981), this concept can be decomposed into four components. First, there is a sense of belonging and group awareness (group identification). Second, there is a preference for members of one’s group and a dislike for those outside their group (polar affect). Third, there is discontent with the group’s status (polar power). Lastly, there is a sense that the group’s lower status is a consequence of systemic inequalities (individual vs society). Under this definition, we understand the concept of linked fate as a sub-category of group consciousness related to the group identification component.Footnote 2

What are the effects of group consciousness on political preferences and political behavior? While most of the studies have focused on the effects of group consciousness on Black communities, more recent scholarship has found these group heuristics to be prominent in other ethnic groups (Sanchez & Vargas, 2016; Berry et al., 2022; McClain et al., 2006). Furthermore, previous studies have found that higher levels of group consciousness have been related to increased support for policies and co-ethnic politicians (Dawson, 1944; Barreto & Pedraza, 2009; Luján, 2020), increased political participation, and the definition of political attitudes (Chan & Jasso, 2021). Additionally, other work has contended that group consciousness increases solidarity across ethnic groups and increased awareness of the group’s issues (Davenport et al., 2021), changes perceptions of a candidate’s success (Burge et al., 2020), and determines party identification and membership (Wamble et al., 2022; Anria, 2018; Van Cott, 2007; Yashar, 2005).

Besides the effects on political attitudes and behavior, group consciousness has been tied to increased demands for political representation of ethnic groups. For instance, Casellas and Wallace (2014) and Wallace (2014) argue that Latinos and Black US citizens think it is essential to have more significant descriptive representation in their districts and the political system as a whole. Other researchers have found that high levels of group consciousness or self-identification as Black can explain support for inclusionary policies such as affirmative action (Mitchell-Walthour, 2011).

Although there haven’t been specific studies on how group consciousness affects district choice, we propose that it functions as a mediating factor in determining ethnic district preferences. Notably, Black individuals exhibiting higher levels of group consciousness are more inclined to prefer voting in ethnic districts. Their motivation stems from the objective of bolstering their ethnic group’s representation within the political system. Heightened group consciousness is associated with greater demands for representation, increased preferences for co-ethnic voting, and more interest in the issues affecting their communities.

Individuals with stronger group consciousness will gravitate towards voting in ethnic districts in order to search for a larger representation of their own group. We also argue that this remains the case where there is more racial fluidity, like in Latin AmericaFootnote 3. Despite the broader range of options for racial self-identification facilitated by mestizaje in these places, we suggest that group consciousness still shapes political preferences as individuals choose to align themselves with specific ethnic groups. Consequently, we expect that:

  • Hypothesis 1: Individuals who have higher levels of group consciousness will prefer to vote in ethnic districts.

Ethnic Organizations: Information, Identity, and Membership

Engaging in collective action is a costly and difficult endeavor. However, one of the ways to overcome these issues is through membership in ethnic grassroots organizations (EGO). In particular, previous work has shown that by mobilizing resources, asserting political demands, having an incidence on local politics, and connecting their members to electoral politics, EGOs have been able to shape members’ political preferences and incite political participation (Edwards & McCarthy, 2004; Poertner, 2021; DellaPorta & Diani, 2006).

EGOs provide information and political cues to members by imparting knowledge about electoral processes and endorsing specific political organizations and candidates. Some scholars argue that in contexts of scarce information, social identities become salient and help determine political preferences (Chandra, 2004). Other set of works focuses on how endorsements from co-ethnics or EGOs help determine political choices (Boudreau et al., 2019). We extend this argument and contend that even in environments with low information, membership in EGOs shape individual’s district preferences. Studies conducted in Colombia, Bolivia, and the US, show how endorsements or presence of EGOs shape vote choice (Villamizar-Chaparro & Echeverri-Pineda, 2022; Poertner, 2021).

EGOs also have an educational purpose: they socialize citizens into their group’s issues, and in doing so, they enhance the salience of group identity (Palmer-Rubin et al., 2020; Andrews et al., 2010; DellaPorta & Diani, 2006). This has been particularly important in Latin America, where EGOs have been vehicles to demand political rights, promoters of political participation, formulators or political platforms, and bridges between ethnic groups and the government (Palmer-Rubin et al., 2020; Yashar, 2005; Van Cott, 2007; Poertner, 2021). Overall, EGOs aggregate demands and link their individual members to electoral politics, influencing their political choices. Based on the above, we argue that because EGOs provide information, influence identity formation, and increase the political salience of ethnicity, its members will be more prone to choose to vote in ethnic districts. Especially under the circumstances where EGOs play the role of political parties* or when they endorse certain candidates. More formally:

  • Hypothesis 2: Individuals from Afro-Colombian EGOs will prefer to vote in ethnic districts.

Colombia’s Ethnic District for the Afro-Descendants Reserved Seats

Colombia’s 1991 Constitution established three districts for ethnic minorities in Congress. The first one reserves two seats in the Colombian Senate for Indigenous peoples, while the second allocates one seat in the Congress’ lower house for the same ethnic group. The third district, also created in the House of Representatives, allocates two seats for the Afro-Colombian population. Importantly, it should be noted that these institutions were designed to be national, thereby rendering all Colombian citizens eligible to cast their votes in them if they decide to do so. This means that every four years, Colombian adults decide whether to vote in their territorial district or in one of these ethnic districts for their vote in the House of Representatives.

These ethnic districts were formulated to redress the historical disenfranchisement of ethnic communities within the realms of electoral participation and decision-making processes. In this respect, traditional parties are excluded from participating in these districts. The law dictates that only entities with a verifiable history of advocating for the Afro-Colombian community, officially registered with the Ministry of the Interior, are eligible to participate and endorse candidates within these districts. These organizations (EGOs) comprise both ethnic organizations and Community Councils. The former represents grassroots organizations dedicated to advancing the well-being of the Afro-Colombian population, while the latter functions as governance entities for collective territories inhabited by Afro-Colombian communities.

When looking at the number of people participating in the district, we find that engagement in it remains relatively stable over time, accounting for approximately 3% of the total valid votes in the House of Representatives. However, despite these districts promoting the descriptive representation of Afro-Colombians, the attainment of substantive representation proves to be a formidable challenge. Previous research argues that these districts, due to the permissiveness of rules governing candidate selection (Basset, 2011; Escandón Vega, 2011), and the absence of mechanisms ensuring accountability for the elected representatives (Echeverri-Pineda & Villamizar-Chaparro, 2018), do not effectively promote the inclusion of ethnic minority issues within the broader national political agenda.

Figure 1 illustrates the number of organizations (Fig. 1a) and candidates (Fig. 1b) participating in elections within the Afro-Colombian ethnic district spanning the years 2006 to 2022. We can observe that despite having a magnitude of two seats, on average, 42 organizations are competing in this district. Those same organizations have presented, on average, 112 candidates each election cycle. This aligns with media reports highlighting the perceived leniency in nomination practices among these organizations (Medrano & Velandia, 2022). These statistics illuminate the notably fragmented nature of the district, as evidenced by an average fragmentation index of approximately 0.89.Footnote 4 This competitive landscape is further accentuated by the presence of unsuccessful parties constituting nearly 76% of the total number of organizations.

Fig. 1
figure 1

Number of Candidates and Organizations running for office in Colombia’s Black District (2006–2022). Panel (a) shows the number of Black organizations competing for the two Afro-reserved seats since 2006. The horizontal line shows the average number of organizations for the whole period. Panel (b) presents for the same period the number of candidates running in the ethnic district. Data for both graphs comes from the Registraduría Nacional del Estado Civil

Full size image

Research Design

This study employs a mixed-methods design to gain a comprehensive insight into the factors influencing Afro-Colombians’ district choice, thereby strengthening the validation of the study’s findings. Particularly, we implemented a parallel research design where we independently ran a conjoint experiment and a series of elite interviews. The combination of voter information gathered through surveys, along with semi-structured interviews with organizational leaders, facilitated a nuanced exploration of electoral dynamics in the Afro-Colombian district. This approach provides information on individual electoral decisions while also offering broader and contextualized perspectives from leaders based on their organizational experience in the electoral process. The analysis derived from the interviews was aligned with the quantitative data analysis, serving to contextualize and elucidate the patterns observed in the quantitative dataset.

Survey and Sample

The quantitative component of this research comes from a survey of 1,200 Afro–Colombian adults we conducted between October and November 2021. Participants were recruited with the help of a survey firm and had to self-identify as Afro–Colombian before answering the survey. Individuals who responded to our survey got credits that could be exchanged for prizes with the survey firm.

Our convenience sample roughly aligns with Colombian census data. Our respondents were primarily single women with an average age of 32. According to Colombia’s 2018 census, men aged 15–64 comprise 65.2% of the total Afro-Colombian men. Meanwhile, the same statistic for Afro-Colombian women was 64%. Age pyramids also show more Afro–Colombian women in the 25–29, 30–34, and 35–39 age bands. Respondents in our sample are also more educated; while the national average for Afro-Colombians with no education is 6.1%, only 4.4% of respondents have no education. Working Afro-Colombians are around 38%, while the proportion of respondents working in our sample is 43%.

Table A2 presents some descriptive statistics from our sample. While respondents in our study demonstrate a limited identification with specific political parties, a significant proportion is affiliated with community councils (56%) and ethnic organizations (47%). Additionally, 63% of our sample believe there has been historical discrimination against Afro-Colombians in the country, and 60% believe Blacks are treated worse than Whites. Furthermore, 49% of our sample has gone through at least one episode of racial discrimination, and 51% of respondents feel self-identification as Afro-Colombians is essential to their identity.

Choosing Between Districts: Conjoint Experiment

We embedded a forced-choice discrete conjoint experiment in our survey. For this experiment, respondents were presented with a choice between two hypothetical districts that varied in seven attributes, as explained in Table 1. In particular, we varied the type of district, the number of candidates running, the candidate’s policy experience, the extent of clientelism, the candidates’ ethnicity, a measure of constituency work for the Afro-Colombian community, and a measure of socialization into the district. Respondents were exposed to two profiles with randomly assigned attribute levels, maintaining consistent attribute order. Respondents were then prompted to select their preferred district for voting, and this process was iterated three times. A separate working paper elaborates further on the key findings derived from these results of the conjoint experiment.Footnote 5

Table 1 Conjoint Choice Table
Full size table

In this article, our primary focus is the implementation of subgroup analyses on the conjoint results to test our hypotheses pertaining to group consciousness, organizational membership, and district choice. To do so, we estimate the Conditional Average Marginal Component Effects (CAMCE) to understand the causal impact of different factors on district choice (Leeper et al., 2020). CAMCEs give us a causal interpretation of the effect certain attribute levels have on district choice. However, CAMCEs might not be the best for describing what people generally prefer for each factor because the interpretation of the results can be sensitive to the baseline category, especially in subgroup analyses. So, alongside CAMCEs, we also calculate Marginal Means (MM) to talk about how much people generally like or dislike districts with certain features (Leeper et al., 2020).

We use a conjoint experiment over other observational and experimental designs for three reasons. First, these experiments help disentangle multidimensional treatments into average treatment component effects for each attribute (Hainmueller et al. 2014). Choosing a district might entail more considerations than the type of district, such as candidates’ experience and the number of candidates running. The conjoint allows us to disentangle the effects of every attribute on the probability of choosing a district and focus on the subgroup analysis of only one of the attributes under a causal framework.

Second, conjoints are well suited to deal with social desirability biases (Horiuchi et al., 2022) and avoid satisficing behavior (Bansak et al., 2021), concerns that could be present in our convenient sample. Thirdly, the conjoint simulates an actual decision every Colombian has to make every time they vote for Congress. In this respect, we are not asking our respondents to think about a hypothetical situation they never experienced or recall their vote of a few years back but, rather, a familiar choice. This allows our results to be more robust when compared to other types of designs.

Variable Operationalization

The main dependent variable for this paper is the change in the probability of choosing a particular district. Given our interest in examining the effects between district choice and indicators of RGC and group membership, we conduct a subgroup analysis of the conjoint results. This analysis explores the interaction between the conjoint attributes and individual characteristics of our respondents. We compute the Average Marginal Component Effect (AMCE) and the Marginal Means (MM) for each interaction model to facilitate interpretation. We present estimates detailing the differences between AMCEs and MMs across groups for a specific conjoint attribute (Type of district). The comprehensive set of results for all conjoint attributes is available in the appendix in Figures A3 to A23. However, many are not discussed due to the absence of statistically significant differences among groups.

We measure group consciousness in three ways that resemble the characteristics identified by Miller et al. (1981). First, as a measure of group identification, we used the answers to a question asking respondents about believing in linked fate. To gauge individual vs. system blame, we employ a question asking respondents if they perceive Afro-Colombians as being treated worse than Whites. Lastly, to assess polar power, respondents are questioned about their belief in historical discrimination against the Black community in Colombia. These questions, adapted from the ANES, are measured on a 5-item agreement scale. To make interpretation easier, we create indicator variables and assign a value of one if the respondent agrees or strongly agrees with each question and zero otherwise. It is noteworthy that results remain consistent when using the original agreement scales.

As explained beforehand, two types of EGOs can endorse candidates for Colombia’s Black ethnic district: community councils and grassroots ethnic organizations. To measure membership in these organizations, we asked our respondents how often they attended meetings held by these organizations. We then created indicator variables that took the value of one if the respondent assisted at least once a week and zero otherwise.

Elite Interviews with Afro–Colombian Leaders

Alongside the conjoint, we conducted eleven interviews with Afro-Colombian leaders. These interviews were crucial for reflecting on the complexity of factors explaining the electoral preferences of Afro–Colombians because they provide valuable insights into the needs, concerns, and aspirations of the communities. Additionally, as leaders serve as intermediaries connecting political institutions and individuals, their thoughts provide insights grounded in their experience and knowledge of the internal dynamics within Black organizations. Furthermore, being Afro–Colombians themselves, they hold unique perspectives on the role of RGC and its relationship with district choice.

We created a convenience sample of potential interviewees, focusing on two distinct groups. First, we identified organizations that participated in the ethnic district between 2006 and 2018, prioritizing those with a history of winning a seat or participating in multiple elections. The second group consisted of non-electoral entities formally registered as Community Councils or grassroots organizations by the Ministry of the Interior. We prioritized organizations registered in places with the largest number or percentage of Afro-Colombians. Challenges in reaching some organizations arose due to factors such as organizational disappearance and outdated contact information.

The interviews took place from August to December 2021 and were primarily conducted via telephone and video calls due to ongoing COVID-19 restrictions, which also posed challenges in contacting some organizations. On average, each interview lasted 30 min. The final sample of participants consisted of 9 men and 2 women, all serving as representatives or former representatives of Afro-Colombian organizations. During the interviews, participants were queried about their opinions regarding participation in the ethnic district and explored the role of EGOs in electoral decisions.

Following the interviews, we systematically identified key themes, categorized responses into relevant topics, and used matrices to organize the information efficiently. Subsequently, a content analysis was employed to scrutinize responses within each category, unveiling trends and variations in participants’ answers for comparison with quantitative data. For confidentiality, the evidence presented excludes identifying information. Further details on fieldwork procedures and an ethics statement, along with a discussion on interviewee safety and the questionnaire used, can be found in the Appendix.

Results

Group Consciousness and Preferences for the Ethnic District

To check for statistically significant differences between subgroups, we conducted simple difference in means tests using the cregg package in R. Figure 2 presents the results of this procedure by each of the RGC variables. The right panels present the differences between CAMCE for each of the subgroups, indicating the difference in effect sizes for the conjoint results. In contrast, the left panels present the differences between the marginal means, indicating the significance of the respondent’s preferences. The solid grey lines mark a zero difference, meaning there are no statistically significant differences across groups. Lines around the estimates are 95% confidence intervals. For this figure, we focus only on the differences in the attribute named type of district; the complete set of results for all of the conjoint attributes can be found in the appendix in Figures A3 – A11. In this paper, we focus on explaining the patterns for the conjoint’s district type attribute.Footnote 6

Our results show that individuals with higher group consciousness tend to prefer ethnic districts over territorial districts, as AMCEs for the ethnic district (Afro) are positive and statistically significant. More specifically, our results illustrate that shifting a district’s type from territorial to ethnic increases district choice by ten percentage points for individuals with high RGC (See Table A4). Substantively, these results mean that individuals with stronger RGC will choose districts with an ethnic component more often. To contextualize this effect, it is important to note that the estimated difference is comparable in magnitude to the individual effect of the Clientelism attribute in the overall conjoint results. This particular effect of clientelism stands out as the largest observed effect in the entire experiment, as illustrated in Figure A1.

The right panels of Fig. 2 present the MM differences between individuals with high and low RGC levels who chose territorial or ethnic (Afro) districts. Among those individuals who chose an ethnic district, having high levels of RGC increases the probability of choosing a district by around four percentage points. This entails that individuals with higher RGC have stronger preferences when choosing ethnic districts. The magnitude of this result is higher in those individuals with stronger feelings of linked fate (group identification) −5.2 pp- and lower for the polar power measure −3.7 pp. On the other hand, the results from those who selected territorial districts show that individuals with higher RGC have an increased aversion to choosing these kinds of districts. In particular, individuals with high RGC prefer territorial districts four percentage points less than their non-RGC counterparts. In general, these results align with what has been found in the literature about RGC, representation, and preferences for a specific district (Wallace, 2014; Casellas & Wallace, 2014; Schulte-Cloos & Bauer, 2021) and show that our results are robust to changes in the baseline categories.

Fig. 2
figure 2

Differences between AMCEs and MMs for Group Consciousness Variables. The Right panel presents the differences between the AMCEs for each subgroup, and the left panel the difference between marginal means. The solid grey lines mark a difference of zero, which means there are no differences across groups. Lines around the coefficients are 95% confidence intervals. Historical discrimination was coded as one if a respondent answered that they agreed or agreed a lot with the fact that centuries of slavery and discrimination have created conditions that make it harder for Afrocolombians to get a job that allows them to exit poverty. Linked fate was coded as 1 if respondents agreed that their lives are affected by what generally happens to other Afro-Colombians. Finally, we asked respondents to answer how much they would agree that in Colombia, black individuals are treated worse than white people. We coded all those who "agreed" or "agreed on a lot" as 1 and the rest as 0 for the present sub-group analysis

Full size image

Our statistical outcomes align with the results from our qualitative work as explained by a member of an ethnic organization: "It is completely clear and proportional that the greater the ethnic identity, the greater the possibility of voting for the ethnic district" (Interview 11). Moreover, our interviewees agreed that people who vote in the ethnic district recognize the importance and symbolism of this institution after years of struggles for Black representation. Interviewee 5, leader of a Concejo Comunitario, also emphasized how the ethnic district belongs to Black Colombians and that it incentivizes participation. Similarly, interviewee 10, an Afro-Colombian activist, said, "The people who vote for the district [...] understand that the district implies an exercise of inclusion and participatory democratic openness"

In terms of group identification, our qualitative work shows that feelings of linked fate and suffering historical discrimination are key in the creation of group consciousness. Highlighting the importance of group identification and the group’s polar power, Interviewee 6, a member of an Afro-Colombian women’s organization, explained that the existence of this group awareness: "has to do with factors of poverty, factors that are marginalizing the ethnic population." Similarly, Interviewee 10 explains that the process of political transformation and larger group awareness starts when "a person who feels that has been attacked because of their skin color, who suffers from discrimination and structural racism" makes choices about who to select for decision-making bodies in a way that can "properly represent us; so that the spaces for participation can raise these issues on the legislative agenda." Other interviews highlighted that individuals who vote in the ethnic district are people who are more "aware of their blackness" (Interview 2) or "more rooted in their culture" (Interview 8).

Finally, our results agree with McClain et al. (2009) in the fact that group consciousness is not the same as group identity since it has been politicized with beliefs for better social standing. We fail to find any statistically significant differences between the preferences of those individuals who believe being Afro-Colombian is important for their identity and those who do not (see Figures A21 – A23) as well as those individuals living in departments with a larger proportion of Afro-Colombians (see Figure A32).Footnote 7 In contrast, we do find strong evidence that the RGC variables are correlated with different preferences for the ethnic district.

Broken Linkages Between Organizations and Electoral Politics

Regarding membership in EGOs, we hypothesize that their members will vote more in ethnic districts. Vote recall data from our survey shows that the probability of voting in the ethnic district increases when people belong to either grassroots organizations or community councils (Table A3). However, our experimental findings in Fig. 3 illustrate that those respondents who belong to EGOs do not prefer ethnic districts over territorial ones (see Figures A12 – A17 for the complete set of results). Our findings indicate no differences in CAMCEs or MMs between individuals who belong to these organizations and those who do not. When looking at community councils, we see similar patterns.Footnote 8.

How do we square the differences between the vote-recall data and the experimental results? We argue that the disconnection between participation in ethnic organizations and district preferences can result from three interrelated processes at the institutional, organizational, and individual levels. First, at the institutional level, we contend that many small ethnic organizations competing in the ethnic district lead to disorganization and fragmentation during elections (Echeverri-Pineda & Villamizar-Chaparro, 2018; Escandón Vega, 2011; Basset, 2011). Elections in ethnic districts produce many electoral losers, generating weaker mandates and lower accountability as voters lose interest in the representatives themselves. In the words of Interviewee 11: "When there are 120 candidates for two seats and only two win, the other 118 feel like losers. When you lose, you don’t feel that the one who wins represents you. For those of us who reach seats, this means that we do not have the approval and support from all those who voted."

Fig. 3
figure 3

Differences between AMCEs and MMs for Group Consciousness Variables. The Right panel presents the differences between the AMCEs for each subgroup and the left panel the difference between marginal means. The solid grey lines marks a difference of zero which means the no differences across groups. Lines around the coefficients are 95% confidence intervals. Belonging to an ethnic organization was coded as 1 if the respondent attended a meeting from said organization at least once or twice a year. Belonging to a community council was coded as 1 if the respondent attended a meeting from said council at least once or twice a year

Full size image

Second, the organizational dynamics of these ethnic entities can also generate incentives to participate in either district. On the one hand, these organizations can endorse candidates, but they sometimes lack the capacity to be helpful during election time, given their geographical or financial reach. A former holder of one of the reserved seats told us they could get more than 20,000 votes, but only 70 or 80 of them came from the community council that endorsed them despite being supportive every step of the way. They also mentioned that asking a community council today to pay for a campaign is "impossible, impractical, and not real." On the other hand, the lack of companionship from ethnic organizations during elections creates incentives for personalistic campaigns. Some of our interviews underscored that organizations within the black community had become endorsing machines and allowed the rise of electoral casicazgos ("chiefdoms") that work as "political markets in favor of certain sectors of society" (Interview 10). Interviewee 6 also explained that in the ethnic district, "everyone wants to be a candidate, everyone wants to be a protagonist. However, there is no common thread linking all of those agendas."

Finally, a third reason that could explain why membership in an ethnic organization does not affect district choice is the existence of strategic voters within these groups. Given that there will be two reserved seats for the Afro–Colombian community in every election, some voters might vote in their territorial districts to maximize the probability of more extensive ethnic representation in Congress. This is aided by the fact that ethnic organizations and community councils can support candidates in the territorial districts or even be members of the lists of traditional parties in the territorial districts. In this respect, Interviewee 6 told us that traditional political parties usually explore avenues of cooperation with black leaders like themselves to gain their electoral support.

Individuals can also use information strategically. Notably, people involved in ethnic organizations have additional information about the quality of the candidates and representatives that could make them prefer to vote in the territorial district. Similarly, previous studies argue that under a low-information context, voters will tend to vote more often for candidates of their same ethnic identity (Chandra, 2004). Table 2 shows the results of a series of correlations between voting in the territorial district and measures of trust in the ethnic district for EGO members. Overall, we can see that individuals with higher trust in the reserved seats representatives (Trust Ethnic Reps) tend to vote less in the territorial district. We also see that younger and poorer individuals tend to vote in the territorial district more often; this could result from clientelistic appeals in the territorial district. In the case of community councils, we see that members who believe people in reserved seats are doing a lousy job vote more in the territorial district.

Table 2 Correlations between voting in the territorial district and trust in ethnic representatives within ethnic organizations
Full size table

What are the preferences of members of EGOs with more RGC Figures A24 – A31 in the Appendix explore the interactions between our measures of group consciousness and being part of EGOs. We find suggestive evidence about a stronger preference for the ethnic district amongst individuals who are both members of an ethnic organization and have a strong RGC. The effect is more notable and statistically significant when we use an index of group consciousness that adds the individual indicator variables. This result implies the importance of group consciousness even within ethnic organizations.

Concluding Remarks

With the rise of a liberal world order, many countries have established institutions intended to facilitate the inclusion of traditionally marginalized groups. In Latin America, scholars have characterized this as a second inclusionary turn, leading to significant transformations of the political participation structures (Kapiszewski et al., 2021). These new institutions advanced inclusion in three dimensions: recognition in the political arena, access to policy-decision levels, and inclusion in the country’s distributional schemes. Despite the advancements made by these institutions, the degree of real and meaningful inclusion has been low (Godfrank, 2021), has led to a deficit in representation of said groups in the political arena (Htun, 2016) and has been hindered by organizational diffuseness (Roberts, 2021).

This article studies the importance of RGC and EGO membership in choosing to participate in one of these institutions: the reserved seats for Afro-Colombians. This paper finds that group consciousness is a key determinant of district choice. Using a subgroup analysis of a forced-choice conjoint experiment and a series of elite interviews, it suggests that individuals who feel part of a broader marginalized group will have stronger incentives to participate in institutions intended to foster the representation of said group in politics.

Our study sheds light on the intricate nature of the relationship between membership in EGOs and political choice. Our results illustrate that under contexts of heightened party competition where there are chances of supporting ethnic candidates in multiple districts, there is not a strong correlation between organizational membership and district choice. These findings emphasize the necessity of considering the political agency of individuals within these organizations and how intra-organizational dynamics matter. Furthermore, our findings underscore the importance of the institutional incentives that shape the operational landscape of these ethnic organizations during elections. By doing so, we advocate for a comprehensive understanding that recognizes the interplay between individual agency, organizational dynamics, and institutional frameworks in analyzing the relationships between organizational membership and district choice.

In exploring these issues, our results have implications beyond Colombia’s reserved seats by deepening our understanding of the impacts of race relations in Latin American politics. Previous work in the region has focused on the salience of race in the political elite (Bueno & Dunning, 2017; Janusz, 2022). However, less is being done about how race affects mass political participation and preferences for district choice. The existence of a group consciousness creates a baseline that politicians could use to rally voters and push for reforms and policies that could be shaped in their favor. Furthermore, our results imply that representative institutions work best when they appeal to RGC.

Future research should place greater emphasis on investigating the underlying factors and strategic considerations that influence voters’ choices across different districts. An intriguing avenue for exploration involves examining the intersectionality between racial and economic identities and its potential impact on the determinants of district choice and participation in ethnic districts. While our study provides a snapshot of voters in 2021, a longitudinal approach could offer valuable insights into how voters’ strategic considerations and shifts in the strength of their group consciousness over time contribute to long-term determinants of district choice. Analyzing these temporal dynamics would enhance our understanding of the evolving nature of voters’ decision-making processes and support for the ethnic district.