Introduction
Inequality has become an indelible feature of Latin American democracies. Its political and socioeconomic implications are latent across the region and the recent social uprisings in several countries bear witness to the growing discontent resulting from the highly unequal structures rooted in the colonial period. Extensive literature has broadly covered the topic of the economic and political costs of inequality (Przeworski Reference Przeworski2011; Rothstein and Uslaner Reference Rothstein and Uslaner2005; Tilly Reference Tilly1998; Uslaner and Brown Reference Uslaner and Brown2005), and burgeoning scholarly works on its potential effects on undermining democracy were released in the last decades (Nolan et al. Reference Nolan, Salverda, Checci, Marx, McKnight, Tóth and van de Werfhorst2014; Piketty Reference Piketty2014; Sánchez-Ancochea Reference Sánchez-Ancochea2021).
Because inequality is a multidimensional phenomenon, and it presents different manifestations that are deeply intertwined, in this article we approach the concept by disaggregating it into different structures that have historically shaped power relations in Latin America to evaluate their potential impacts on political trust. In particular, building on horizontal inequalities studies (Stewart Reference Stewart2008), we consider the interplay between social, economic, and ethno-racial inequalities as structural hierarchies that shape citizens’ relations with the political system. Institutions should reflect a cross-section of all citizens and ensure the political inclusion of all groups by promoting redistribution. Yet, given the legacy of colonial exploitation, Latin American political systems often perpetuate racialization, and this might eventually impact how marginalized collectives interact with political processes.
Against this backdrop, we ask whether members of racialized groups establish different connections with the political system. Do they exhibit varying attitudes towards institutions? Do they present lower levels of political trust? Our interest is in explaining how structural inequalities shape citizens’ perceptions across Latin America, examining socioeconomic as well as ethno-racial differences in trust. We theorize that both socioeconomic and ethno-racial structures of inequality are reinforcing dynamics that affect power relations and impact differently citizens’ relations with the political system. The persisting income inequality in the region has its roots in long-standing class and ethnic structures (Hoffman and Centeno Reference Hoffman and Centeno2013), and these relations keep perpetuating socioeconomic and political exclusion that undermine the legitimacy of democratic institutions. A vast literature in comparative politics has linked individuals’ attitudes to democratic quality, stating that citizens are the origin of democratic performance (Inglehart Reference Inglehart2015; Welzel Reference Welzel2013; Norris Reference Norris1999). Thus, one might expect that if members of racialized groups feel excluded by the system in systemic ways, their perceptions towards it will impact negatively on democratic legitimacy.
Documenting variations in political attitudes across ethnic groups is a vibrant area of analysis and considerable scholarly research concerning this issue has been released (Abramson Reference Abramson1983; Avery Reference Avery2006; Citrin et al. Reference Citrin, McClosky, Shanks and Sniderman1975; Hetherington Reference Hetherington1998; Lijphart Reference Lijphart1977; Thorp and Paredes Reference Thorp and Paredes2010; Silver and Dowley Reference Silver and Dowley2000), however, due in part to the challenges involved in measuring ethnic inequalities, many of its potential effects on political trust still await systematic cross-national exploration. In this sense, this article aims at contributing to these debates by offering new evidence on the relevance of the interplay between socioeconomic and ethno-racial inequalities as key factors in explaining attitudes towards institutions across Latin America.
Contemporary Latin America is ethnithically and racially diverse but this diversity also encompasses large and profound dynamics of discrimination that can affect how individuals establish their relations with the politcal system. If institutions are “avenues of inclusion” (Munck and Luna Reference Munck and Luna2022), they should be able to represent the interests of all groups and and ensure legitimacy. However, Latin American political systems do not often meet these ideal standards since their institutional designs are based on structural relations of power. This may influence the connections of racialized individuals with the political system and alter their views on its democratic institutions, reflecitng underlying power dynamics.
Racial and ethno-social distinctions, along with their impact on power relations, are key factors in explaining the political, social, and economic trajectories of Latin America. Some scholars have defended that discrimination against ethno-racial groups has contributed to income inequality and resulted from it (Sánchez-Ancochea Reference Sánchez-Ancochea2021; Solis et al. Reference Solis, Güemez Graniel and Holm2019). Therefore, we believe that the ethno-racial component should be considered in the social analyses of the region’s political reality, especially when measuring its connections with inequality. Our guiding expectation is that ethno-racial structures, along with other economic and sociopolitical dimensions of inequality, shape individuals’ attitudes towards the democratic institutions of the political system.
Building on survey data from four waves of the Latinobarometer project and other contextual indicators from different sources, we conduct regression analyses of political trust across 18 Latin American countries. Notwithstanding political distrust seems to be a trend in the region, we find significant differences in means among groups within countries. Results suggest that structural inequalities affect political trust and impact on the formation of different relations with the political system depending on the racialized group and its interactions with other socioeconomic structures at the individual level. We believe that this finding is largely relevant since it would imply that even when increasing the socioeconomic status, ethno-racial structures remain affecting citizens’ relations with the political system. While our results contribute directly to the empirical literature on mass political culture in Latin America, we believe they also might have implications beyond the region.
Our contribution to political culture studies is twofold. First, our analyses strengthen the evidence of how ethno-racial and socioeconomic inequalities interact with power dynamics and systemic processes of racialization that erode democracy. Second, they provide an explicit focus on the relations between ethno-racial inequality and political trust and on how race and ethnicity as historical social structures might shape attitudes towards the democratic institutions of the political system. By emphasizing this issue, we draw attention to the potential risks of persisting structural discrimination for democracy.
The article proceeds as follows. The next sections briefly introduce the theoretical framework of political trust and discuss its relations with structural inequalities and the ethno-racial component. Thereafter, drawing from the above-mentioned literature and our expectations, we posit our hypotheses and introduce our data and methodology. We then present the results and conclude by discussing our findings and their implications.
Why Do People Trust Institutions?
In the last decades, the concept of political trust has been profusely analyzed in political science. Understanding factors affecting political support is important since legitimacy is a key element for democracy to survive (Almond and Verba Reference Almond and Verba1963; Citrin and Stoker Reference Citrin and Stoker2018; Linz and Stepan Reference Linz and Stepan1996). Scholars have traditionally identified large levels of distrust in institutions as a sign of instability, illegitimacy, and poor functioning of the democratic system (Di Palma Reference Di Palma1970; Norris Reference Norris1999; Torcal and Montero Reference Torcal and Montero2006; Van der Meer and Zmerli Reference Van der Meer, Zmerli, Van der Meer and Zmerli2017).
Following scholarly contributions on political disaffection (Torcal and Montero Reference Torcal and Montero2006; Harteveld et al. Reference Harteveld, Van der Meer and De Vries2013; Zmerli et al. Reference Zmerli, Newton, Montero, Van Deth, Montero and Westhol2007), we approach the concept of political trust departing from Easton’s (Reference Easton1975) notion of political support, defined as citizens’ evaluations of the performance and outputs of political authorities and institutions (specific support), and the perceptions of more basic and fundamental aspects of the political system for what it is and represents (diffuse support). While the former is considered more volatile and subject to change depending on the context and on specific actions, the latter is more stable, enduring, and based on fundamental beliefs that make it less susceptible to short-term fluctuations.
The theoretical debate on the explaining elements of political trust is divided into two traditions: the “traditional-culturalist” and the “rational-culturalist.” The supporters of the first model argue that political trust is the result of a long-term process of socialization and attitudes would then be less likely to change over time (Almond and Verba Reference Almond and Verba1963; Lerner Reference Lerner1958). On the other hand, those who adhere to the latter model defend that political trust may exhibit longitudinal variations as a consequence of citizens’ evaluations of the institutional performance (Lane Reference Lane1992; Lipset and Schneider Reference Lipset and Schneider1983; Norris Reference Norris1999; Mishler and Rose Reference Mishler and Rose2001). Deriving from social capital theory (Putnam Reference Putnam2000; Newton Reference Newton2006), scholars have also found that high levels of interpersonal confidence are positively associated with individuals’ trust in democratic institutions (Bjørnskov Reference Bjørnskov2008; Dalton Reference Dalton2004; Glanville and Paxton Reference Glanville and Paxton2007; Newton and Zmerli Reference Newton and Zmerli2011; Uslaner and Brown Reference Uslaner and Brown2005).
In the last decades, the “trust-as-evaluation” approach (Van der Meer and Hakhverdian Reference Van der Meer and Hakhverdian2017), which supports the view that political trust can be explained by democratic systems’ social and economic performance, has increasingly gained ground (Clarke et al. Reference Clarke, Dutt and Kornberg1993; Cusack Reference Cusack1999; Anderson Reference Anderson2009). According to these perspectives, trust in institutions would differ depending on individual perceptions and evaluations about the system’s performance and outcomes. This explanation would be grounded in the presence of short-term variations in the levels of political trust (Zmerli and Newton Reference Zmerli, Newton, Zmerli and Hooghe2011). In this sense, dissatisfaction with the socioeconomic situation among citizens would lead to a decline in political trust.
Among the set of factors used in scholarly works to account for variations in political trust, inequality has drawn special attention in the last decades (Anderson and Singer Reference Anderson and Singer2008; Newton and Zmerli Reference Newton and Zmerli2011; Tyler et al. Reference Tyler, Rasinski and McGraw1985; Uslaner Reference Uslaner, Zmerli and Hooghe2011). In general terms, interest in inequality has increased in recent years with a widespread consensus about its rise both in developed and developing countries (Piketty Reference Piketty2014; Milanovic Reference Milanovic2016). In Latin America, as Prados de la Escosura (Reference Prados de la Escosura, Hatton, O’Rourke and Taylor2007) states, economic inequality grew during the twentieth century, however, during the first decades of the new millennium it showed a decrease in the context of what was known as “the commodities boom” during the so-called “turn to the left” (Sánchez and García Montero Reference Sánchez, García Montero, Sánchez and Montero2019). Nevertheless, the end of the commodities’ cycle along with new events, such as the COVID-19, and together with the historical socioeconomic patterns of the region, reversed this trend. Latin America ranks at the top of any ranking of inequality and is distinguished from other areas of the world for its income concentration within a small elite (Sánchez-Ancochea Reference Sánchez-Ancochea2021). The region would thus lie in Uslaner’s (Reference Uslaner, Zmerli and Hooghe2011) “inequality trap,” where great levels of inequality lead to lower trust and more corruption.
Indeed, several studies have empirically proved that socioeconomic inequality leads to lower levels of citizen’s trust in institutions (Anderson and Singer Reference Anderson and Singer2008; Uslaner Reference Uslaner2008; Zmerli and Castillo Reference Zmerli and Castillo2015). These conceptions are related to the concept of distributive justice further developed by Rawls (Reference Rawls1971) who links discontent towards institutions to perceptions of unfairness in resources redistribution. Following this line, for the Latin American case, Zmerli and Castillo (Reference Zmerli and Castillo2015) found a robust negative association between political trust and perceptions of unfair income distribution.
Structural Inequalities and the Ethno-racial Component
Apart from socioeconomic disparities, when assessing the relation between inequality and political trust in Latin America, it is fundamental to consider the role of ethnicity/race as a social structure. The racial component has historically been at the forefront of explaining social and political processes in the region. Since colonization, income and racial inequalities have come together in Latin America (Sokoloff and Engerman Reference Sokoloff and Engerman2000; Acemoglu et al. Reference Acemoglu, Johnson and Robinson2001), when racialized groups (fundamentally Indigenous and Afro-descendants) lied at the very bottom of the social structure. These racial hierarchies were intensified on the eve of independence when nonwhites were relegated to an inferior legal and social status (Munck and Luna Reference Munck and Luna2022). The miscegenation (mestizaje in Spanish) project promoted by the elites across the region in the twentieth century spread the idea of the existence of a “mixed race.” This “cosmic race” would have eliminated all ethno-racial hierarchies by achieving inclusion and leading to the so-called “racial democracies” (Hooker Reference Hooker2017).
However, as can be noted in Table 1, more than 42% of those self-reported as Indigenous in Latin America describe themselves as belonging to the poorest steps on the wealth ladder,Footnote 1 whereas among those self-reported as White the percentage is 23%. As Solis et al. (Reference Solis, Güemez Graniel and Holm2019) point out, the interplay of social class inequalities and ethno-racial relations should be considered when analyzing inequality since in the region, skin color and social status are generally connected (Hall and Patrinos Reference Hall and Patrinos2012; Telles and Paschel Reference Telles and Paschel2014). Following these conceptions, we include ethnicity/race as a variable measuring structural inequality to assess differences in political trust emphasizing the relevance of racialized structures in shaping power relations.
Table 1. Perceived Wealth by Self-reported Race in Latin America (%)
Source: Own elaboration from Latinobarometer (2020).
The meanings and constructions of ethnic and racial identitiesFootnote 2 in Latin America face many complexities and are deeply intertwined with historical legacies, defying categorizations. Scholars have extensively analyzed how these identities are not only a matter of biological features but are also and especially shaped by social, economic, and political forces (de la Cadena Reference De la Cadena2000; Knight Reference Knight and Graham1990; Telles Reference Telles2014; Wade Reference Wade2010). These works underscore the diversity of ethno-racial categorization and the difficulties of racial dynamics across the region. It is important to understand these identities not as fixed categories but as evolving constructs that reflect power dynamics. Consequently, recent contributions show that operationalizing inequalities between ethnic groups presents several challenges and existent cross-national measures of ethnic socioeconomic and political inequality differ significantly in terms of empirical scope, conceptualization, and methodology (Leipziger Reference Leipziger2023).
Some of the most prominent explanations on income distribution in Latin America revolve around social class (González Casanova Reference González Casanova2003), however, scholars have questioned these arguments and highlighted the role of ethno-racial notions in the analysis of Latin American social reality showing that race encompasses both physical and cultural aspects and its meaning may vary across countries due to different historical and political trajectories (Aguilar Reference Aguilar2011; Guerreiro Ramos Reference Guerreiro Ramos1957; Lancaster Reference Lancaster1991; Telles and Paschel Reference Telles and Paschel2014). As shown by Zizumbo-Colunga and Flores Martínez (Reference Zizumbo-Colunga and Flores Martínez2017) for the case of Mexico, skin color is more important than geography, economic sector, or gender in explaining wealth and education gaps. These notions are related to the concept of “horizontal inequalities” developed by Stewart (Reference Stewart2008) that refers to the intergroup differences based on a socio-cultural identity, such as ethnicity. Horizontal inequalities are multidimensional, encompassing not only socioeconomic disparities but also other cultural and political cleavages. This explanation is theoretically close to Tilly’s (Reference Tilly1998) conception of categorical inequalities that focuses on how different social categories (ethnicity, gender, class) become the basis for systematic divergences in access to material and symbolic resources. Building on this theoretical framework, we approach the concept of structural inequality by including wealth, social class, and race/ethnicity as structures overlapping in power relations in Latin America.
Ethno-racial Inequality and Political Trust
Recent contributions have unveiled how increasing inequalities between ethnic groups are associated to democratic erosion (Houle Reference Houle2015; Panzano Reference Panzano2023; Wrigley-Field Reference Wrigley-Field2020). In parallel, literature on the psychology of intergroup relations defend that ethno-racial inequality may shape individuals’ political attitudes (Allport Reference Allport1954; Pettigrew and Tropp Reference Pettigrew and Tropp2006). Scholars in Canada and the United States show that different measures of ethnicity (i.e., race, origin, and language) have impact on trust (Roseborough and Breton Reference Roseborough, Breton, Blishen, Jones, Naegele and Porter1968; Simeon and Elkins Reference Simeon and Elkins1974). For the Latin American case, a growing body of research exploring how ethno-racial inequality affects mass political culture has emerged in the last decades (Aguilar Reference Aguilar2011; Bailey Reference Bailey2009; Bueno Reference Bueno2012; Clealand Reference Clealand2017; Janusz Reference Janusz2022; Madrid Reference Madrid2005 and Reference Madrid2012; Mitchell-Walthour Reference Mitchell–Walthour2017; Morgan and Kelly Reference Morgan and Kelly2017). These works explore the formation of ethno-racial identities and examine the relations between ethnicity and a set of political attitudes identifying racial gaps and gathering evidence on the existence of sociopolitical exclusion for members of marginalized racial groups.
At the aggregate level, some studies have evaluated the relation between exclusion and some democratic values like political tolerance across Latin America and found that intolerance is higher when ethno-racial hierarchies are pronounced (Morgan and Kelly Reference Morgan and Kelly2021). Regarding political trust, through an analysis of racial dynamics in Brazil, Telles (Reference Telles2014) demonstrates how racial and ethnic identities impact trust in political institutions, revealing a complex interplay between social stratification and political engagement. These works collectively highlight the challenges and opportunities for building cohesive societies in ethnically diverse contexts.
Hypotheses: The Effects of Structural Inequalities on Political Trust
Couched on the theoretical arguments outlined above and considering the interplay between income, social class, and ethnicity, our first set of hypotheses looks at how structural inequalities shape political trust. Since we regard inequality as a multidimensional phenomenon, we consider the effects of economic as well as sociopolitical inequalities over political trust including also the racial issue. Previous studies show that economic inequalities between ethnic groups are likely to increase the probabilities of democratic decline and onset of autocratization (Houle Reference Houle2015; Panzano Reference Panzano2023). Similarly, recent scholarly works have linked sociopolitical ethnic exclusion to the erosion of the quality of democracy (Bornmann Reference Bornmann, McCulloch and McGarry2017; Cederman et al. Reference Cederman, Hug and Wucherpfenning2022). In this sense, we expect that:
H1a High levels of economic inequality between ethnic groups are negatively associated with political trust.
H1b High levels of sociopolitical inequality between ethnic groups are negatively associated with political trust.
Given the substantial importance of ethno-racial inequality in Latin America in explaining economic and sociopolitical processes, we consider that a study which investigates the existing connections between inequality and political culture should include the ethno-racial dimension as a key variable in the analyses. However, we argue that the effect of racial identity on trust will yield divergent results depending on the depth of those inequalities between ethnic groups. Our suspicion is that:
H2a As economic ethnic inequality increases, those self-reported as belonging to historically racialized groups, namely Indigenous and Afro-descendants, will exhibit lower trust in contrast to those self-identified as Whites.
H2b As sociopolitical ethnic inequality increases, those self-reported as belonging to historically racialized groups, namely Indigenous and Afro-descendants, will exhibit lower trust in contrast to those self-identified as Whites.
Due to the complexities behind cross-national racial categorization, in contrast to H1, which relies on a series of theoretical and empirical works, H2 is rather an explorative hypothesis. We theorize that there are distinct trust groups that reflect different types of relations within the political system. On the other hand, at the individual level, we also establish a set of expectations. Even though increases in socioeconomic status have been found to correlate with higher levels of political trust in a vast number of studies, when including the ethno-racial component, this trend is not as evident. Scholars have explored how racialized experiences influence political attitudes and found that systemic barriers persist regardless of individual qualifications or socioeconomic status (LaVeist and Isaac Reference LaVeist and Isaac2013; Mendelberg Reference Mendelberg2001; Swain Reference Swain1993; White and Laird Reference White and Chryl2020). Within this framework, we derive the following hypothesis:
H3 The likelihood of trusting institutions for racialized groups does not increase significantly even when perceived wealth levels and subjective social status are rising.
We suspect that when Whites increase their socioeconomic status, they will express comparatively higher trust in institutions. However, we might expect that when these indicators are modified in the same direction for the members of racialized groups, levels of political trust would not suffer significant increases. Insofar as individuals see they belong to a group who has been historically discriminated by the system, they may remain suspicious of its institutions. In short, we posit three groups of hypotheses: the first pertains to the macro-level, the second addresses cross-level interactions, and the third examines interactions at the individual level.
Data and Method
To test our posited hypotheses, we rely on individual-level public opinion data from the Latinobarometer project along with contextual-level data from different sources. Latinobarometer has conducted cross-national surveys across Latin America and Spain since 1995 and it includes a range of measures of political trust as well as different proxies of inequality. We employ the four most recent available installments of the survey that include all our measures of interest: 2011, 2013, 2018 and 2020. Our sample consists of 72,455 individuals across 18 Latin American countries.Footnote 3
Consistent with the aforementioned theoretical conceptions, to operationalize our dependent variable (political trust) we selected four items from a battery of questions in the survey measuring trust in different institutions and actors of the political system (government, congress, judicial power and political parties). Given the high correlation of the four variables (0.5 or above in all cases), we first considered adding them into a mean index. However, as different scholars have shown, these four concepts are theoretically and empirically different and trust in different objects may have different implications (Lipset and Schneider Reference Lipset and Schneider1983; Denters et al. Reference Denters, Gabriel, Torcal, van Deth, Montero and Westholm2007). For this reason, we decided to run independent models for each institution/actor.Footnote 4 In the four questions, respondents are asked to indicate how much trust they have in each group/institution on a four-point Likert-type scale where 1 represents “a lot,” 2 “some,” 3 “little,” and 4 “none.” We reverse coded the four variables so that greater scores always indicate higher levels of trust. Cronbach’s alpha scores for each individual item exceed 0.86 in all cases, reflecting strong internal consistency.
Our main explanatory variables of interest are those related to structural inequalities. At the country level, our analyses include two measures: economic ethnic inequality and sociopolitical ethnic inequality. To identify potential economic disparity among ethno-racial groups across Latin America, we introduce a between-group inequality measure (BGI) (Baldwin and Huber Reference Baldwin and Huber2010). This indicator ranges from 0 to 1 and is similar to a group Gini coefficient but differs in that it calculates inequality by assigning the mean income of each group to all its members, rather than using individual incomes. It is computed by determining the average income for each ethno-racial group in each country and then considering how much each group’s average deviates from the national mean.Footnote 5 The differences in these average incomes are adjusted by twice the mean income of the society, enabling comparisons of BGI scores across countries with varying income levels.
Additionally, in order to assess the effect of sociopolitical ethnic inequality cross-nationally, we utilize the n-star indicator (Cederman and Girardin Reference Cederman and Girardin2007) from the Ethnic Power Relations (EPR) dataset that measures power relations between ethnic groupsFootnote 6 and ranges between 0 and 1 depending on the number of excluded groups in the country.
At the individual level, to measure wealth across the selected four waves of the Latinobarometer, we include self-perceived wealth using a 1 (poorest) to 10 (richest scale). Additionally, we use a measure of distributive justice as proxy for evaluation of income inequality following Uslaner’s (Reference Uslaner, Zmerli and Hooghe2011) conception on the consequences of individuals’ perceptions of unfairness in income redistribution. To do so, we use a 4-point scale where 1 means “very fair” and 4 “very unfair.” We proxy social status on a 5-point scale where 1 means “upper class” and 5 “lower class.” We reverse coded both scales for the analyses. For ethnicity, we use self-reported ethnicity (categorial measure) including the following categories: “White,” “Mestizo,” “Indigenous,” “Afro-descendant” (grouping “Black” and “Mulatto”) and “Other race” (grouping “Asian” and “Other races”).Footnote 7 We treat “White” as the reference category.
It is important to mention the challenges entailed in the quantitative study of this variable. To begin, since race and ethnicity are social constructions, their dimensions also reproduce real social boundaries that define groups and maintain systemic inequalities (Paredes Reference Paredes2018). Empirical studies regularly proxy this variable by using self-reported racial categories or chromatic ranges from skin-color palettes that are often directly associated with certain socioeconomic status. This could lead respondents to be susceptible to social desirability bias. On the other hand, in a region like Latin America, where the myth of mestizaje became the basis of national projects promoting the idea of racial mixing (Telles Reference Telles2014; Wade Reference Wade2010; Wade and Moreno Figueroa Reference Wade, Figueroa, Mónica, Figueroa and Wade2022), the mestizo classification may operate as a category where individuals fell seeking upward social mobility and eluding discrimination. Indeed, 46.74% of the respondents from our sample are self-reported as Mestizos, this being the larger category. Sadly, being mestiza has not made Latin America less racist. This is what is called the “mestizaje trap” (Amselle Reference Amselle and Kandé1999).
Finally, our analyses include a range of controls including interpersonal confidence (dichotomous) and sociodemographic factors like age (continuous), self-reported sex (1 woman), and education (1 elementary 2 secondary and 3 higher). At the aggregate level, to measure inequality at the country level, we also control for the Gini Index, based on the Standardized World Income Inequality Database (SWIID) (Solt Reference Solt2020), along with the logarithm of GDP per capita as an indicator of economic development.
Figure 1 displays each country’s mean and standard deviation for the dependent variable (descriptive statistics for the four questions can be consulted in the Online Appendix, Table A2) and offers support for our conceptual contribution as it presents distrust as a constant across the region, but it does vary within countries. The standard deviation bands cross the mean in every case except for El Salvador and Uruguay for trust in the government and the judiciary, suggesting that more variation exists within rather than between countries. This would imply that even though there is not considerable cross-national variation in trust, inside the countries individuals would hold diverging behaviors towards institutions. We are thus concerned with understanding the logics behind these groups’ differences and identifying whether socioeconomic and ethno-racial hierarchies might impact them.
Figure 1. Country’s Average Political Trust with the Horizontal Axis Presenting the Total Mean and Bands Showing the Standard Deviation.
Source: Own elaboration based on Latinobaromenter.
Taking into consideration all of the above, we test our assumptions using mixed-effects multilevel modeling. Since individuals are nested in countries, which are in turn nested within years, this strategy allows us to respect the three-level structure of our data. First, we estimated a full model with fixed effects at the country/year level to control for time-invariant characteristics and specific country and year features, isolating the effect of our variables of interest (Model 1). Additionally, we estimated two Random Effects Between country (REB) models for each institution/actor including two cross-level interactions between the self-reported race of the respondent and the BGI, on the one hand, and the sociopolitical ethnic inequality, on the other (Model 2). Finally, we estimated a model with interaction terms at the individual level between self-reported race and respondent’s perceived wealth/subjective social status (Model 3).
In Models 2 and 3, we include a random component at the country level along with the fixed country/year effects. This approach allows us to control for unobserved cross-country heterogeneity and control for invariant factors within countries and years. Since we theorize that both socioeconomic and ethno-racial structures are dynamics interviewed, the purpose of the interactions is to explore whether between-group inequalities exert varying effects based on racial identity and altering socioeconomic conditions at the individual level impacts racialized groups in distinct manners.
Before testing our hypotheses in multivariate models, we examined whether differences in means on political trust exist among ethno-racial groups. Data reveal intergroup divisions among respondents on our dependent variables. One-way ANOVAs were performed to compare the effect of ethno-racial categories on the four variables and results confirm statistically significant differences in means (p < 0.000 in all cases).
Results
Table 2 estimates the effects of structural inequalities on political trust (Model 1). The model includes country/year fixed effects (not shown)Footnote 8 with both individual and country-level measures for each of our four dependent variables, while also controlling for the indicators specified above.
Table 2. Effects of Structural Inequalities on Political Trust
+p < 0.10; *p < 0.05; **p < 0.01; ***p < 0.001. Standard errors in parentheses. Source: Own elaboration from Latinobarometer.
Consistent with most previous studies, in general, at the individual level, results confirm that socioeconomic features have an impact on political trust. Higher perceived wealth, social status, and perception of fairness in the distribution of outcomes are positively and statistically significantly associated with greater confidence in the democratic institutions. These results suggest that individuals with lower levels of wealth and social status who perceive the income distribution as unfair are more likely to present lower political trust than those who report greater wealth and social class and believe distributive justice exists.
Looking at the contextual level, our two main measures of inequality between groups yield different findings. First, regarding economic ethnic inequality, results support H1a for the cases of the government (where we find the stronger effect), the political parties, and the judicial power, with the coefficients showing a statistically negative relation between BGI and trust in these three institutions. As seen in each model, increasing economic ethnic inequalities would lead to lower levels of confidence in all our considered institutions but the congress. This finding may also suggest that, as was explained earlier based on Easton’s approach, different institutions represent different objects of political support, and this would impact their relations with economic ethnic inequality.
On the other hand, our expectations for sociopolitical ethnic inequality (H1b) cannot be fully supported. Coefficients remain negative for all institutions but the congress, although they only exert a statistically significant effect for the cases of the government and the judiciary. Despite previous contributions having found statistically negative relations between political exclusion and a set of political attitudes (Morgan and Kelly Reference Morgan and Kelly2017 and Reference Morgan and Kelly2021), recent scholarship also found negative but statistically insignificant relations between political and social inequalities and democratic erosion (Panzano Reference Panzano2023). Therefore, further analyses ought to tackle these concerns by comparing results from different measures of social and political ethnic inequalities.
Additionally, in the five cases with statistical significance, the fact ethno-racial inequalities decrease political trust despite the inclusion of controls may suggest that the effect of these structural disparities is unlikely to be solely driven by the respondent’s socioeconomic background or the general economic situation/level of inequality in the country. This may indicate that the ethno-racial component plays a significant role in shaping political trust beyond regular economic conditions, and it is its interplay with other forms of economic and sociopolitical inequalities that impacts power dynamics in Latin America. Thus, these results manifest the importance of our argument of considering inequality as a multidimensional phenomenon where ethno-racial structures along with other economic and sociopolitical features shape individuals’ political attitudes.
Moving now to the controls, sociodemographic proxies reveal again different findings for diffuse and specific support. Whereas trust in the government increases among older respondents, the direction for the rest of institutions/actors is the opposite (although only significant for the judicial power). In parallel, being a woman predicts lower trust in the government and the political parties, while higher levels of education are negatively and statistically associated with trust in the parties and the congress.
This last finding goes in line with previous research whose results also show a negative effect of education on political trust (Seligson Reference Seligson2002), although there is a large body of empirical work puzzling over the variability of the effect of education on political culture (Verba et al. Reference Verba, Nie and Kim1978; Powell Reference Powell1986; Gallego Reference Gallego2010). For our case, we believe that highly educated people are more likely to be suspicious of institutions since they are considered to reproduce unequal structures. In addition, social trust is a relevant predictor of institutional trust, exhibiting a statistically positive association. Finally, both country-level controls have the expected effect. In countries with higher levels of inequality (higher Gini), individuals are likely to be less trustful of institutions, whereas higher GDP per capita predicts higher trust.
Concerning H2, we found mixed support for our expectations. We specified and estimated one model for each dependent variable in which we added cross-level interactions between racial identity and both economic and sociopolitical ethnic inequalities (Model 2 see Table B2 in the Online Appendix for the complete regression outputs). Figure 2 plots the marginal effects of these interactions.Footnote 9 First, for the economic dimension, results show a statistically significant association (p <.000) between increasing levels of between-group economic inequality and membership of the Indigenous group in the reference category (White) for all cases but the political parties. Nevertheless, the direction of this association differs based on the institution.
Figure 2. Marginal Effects: Economic and Sociopolitical Ethnic Inequality Effects on Political Trust across Self-reported Race. Adjusted Linear Predictions with 95% Cis.
Source: Own elaboration based on Latinobarometer.
For the government and the judicial power, political trust decreases as BGI rises for both groups; however, the direction of the association is contrary to what was expected,Footnote 10 suggesting that, in contexts of great economic ethnic inequality, the Indigenous would exhibit greater levels of political trust than Whites. In the case of the congress, results are not the expected since they suggest that not only do Indigenous individuals report higher levels of trust than Whites in contexts of heightened economic ethnic inequality, but their trust levels also progressively increase as BGI rises.
Although this finding may entail further theoretical discussions and case analyses, we hypothesize as a possible explanation that the impact of economic ethnic inequality might be mediated by the sizes of the samples and the interaction of other social and economic indicators with self-reported race. Additionally, these results might also be connected to the literature on redistributive preferences and political engagement in multi-ethnic societies (Hooker Reference Hooker2005; Kinder and Kiewiet Reference Kinder and Kiewiet1979).
Moving to the interaction between sociopolitical ethnic inequality and self-reported race, results in Figure 2 indicate that a statistically significant association exists exclusively for trust in the government, although between distinct racial categories. Results of this estimation indicate that as disparities increase, political trust achieves lower levels for Afro-descendants than for Whites (interaction term statistically significant at p < .000).
The marginal effects plot suggests that whereas rising social and political inequalities would not alter Whites’ levels of political trust, those self-reported as Afro-descendants will become less trustful in the government. Thus, H2b can only be partly supported since differences in trust between racial groups widen in contexts of greater economic and sociopolitical inequality but in different ways. Members of historically marginalized groups, namely Indigenous and Afro-descendants, do not exhibit lower levels of trust as ethno-racial inequalities rise, as we theorized. However, we do find nuanced associations for the Indigenous membership and the economic dimension, and the Afro-descendants category and the sociopolitical sphere. We believe that this finding is worth further investigation, and we aim in future works to keep exploring the variations in the political attitudes across different marginalized groups.
In relation to prior research, two points merit attention. First, findings from Models 1 and 2 confirm, in general, the connections between horizontal inequalities and political trust, identifying a direct effect on diminishing confidence levels (H1). Second, related to the mixed support we find for H2 on specific differences between ethnic groups in contexts of structural inequalities, across the three models, members of historically racialized groups show on average higher levels of political trust than White respondents.
This resonates with the theoretical debate that supports that in multi-ethnic democracies, minority groups may exhibit higher levels of political trust, especially when their engagement is fostered through targeted policies, representation, or community mobilization efforts (Hooker Reference Hooker2005; Miller et al. Reference Miller, Gurin, Gurin and Malanchuk1981; Van Cott Reference Van Cott2005). When racialized groups perceive that the political system is responsive to their demands, their respect for political actors and institutions may increase (Collins Reference Collins1990; Crenshaw Reference Crenshaw1991; Dawson Reference Dawson1994; Jones-Correa Reference Jones-Correa2005).
In relation to the individual-level interactions, Figure 3 plots the marginal effects of perceived wealth by racial identity for political trust.Footnote 11 Results show a statistically significant negative association for the Indigenous category in all institutions/actors but the judiciary (although coefficient is also negative). Figure 3 shows that as perceived wealth increases, political trust achieves higher levels for White respondents than for Indigenous respondents (interaction term statistically significant, p < .010 in all cases). Indigenous evidence significantly lowers increases in trust as perceived wealth rises than Whites. This trend is particularly steady for the case of the parties.
Figure 3. Effects of Racial Identity on Political Trust Mediated by Perceived wealth. Adjusted Linear Predictions with 95% Cis.
Source: Own elaboration based on Latinobarometer.
These associations might imply that, when income intervenes, political trust does increase for those self-reported as Whites, but for members of the Indigenous group this increment is less pronounced. In other words, climbing the wealth ladder increases trust in institutions for Whites, but not as much for Indigenous individuals. Variations in trust when perceived wealth grows are different depending on the racial category.
Likewise, the model reveals a similar effect on the direction for Afro-descendants, however, the coefficient is only statistically significant for trust in the government. This association would support our argumentation described earlier that the historically deep colonial roots of ethno-racial hierarchies are still relevant to explain political dynamics and relations in Latin America. Recall that in H3 we hypothesized that the likelihood of trusting institutions for racialized groups does not increase significantly even when perceived wealth levels and subjective social status are rising. This expectation is partly supported since results of the interaction terms between self-reported ethnicity and social status are not statistically significant for members of the marginalized groups.Footnote 12 However, we believe the strongest and most powerful empirical finding here is the consistent association between ethno-racial and socioeconomic individual inequalities and the varying effects of their interactions between racial groups on political trust.
Additionally, interactions terms further validate Easton’s differentiation between “diffuse” and “specific support.” Notably, trust in the government emerges where we find stronger significant cross-level interaction terms and where all racial categories show statistical significance at the individual level. This finding suggests distinct dynamics in how citizens engage with institutional frameworks, shaped by how they perceive the objects of trust for what they represent or for their specific actions and performance. Trust in the government might appear more closely linked to confidence in particular political figures, performances, or outcomes, likely influenced by evaluations of governmental efficacy. In contrast, trust in the other three institutions—presumably congress, political parties, and especially the judiciary where significant associations were not found—may reflect a broader, more enduring form of support, less swayed by immediate political outcomes and more anchored in the fundamental principle of the political system (diffuse support).
Discussion and Conclusion
Political trust is a crucial dimension testing the democratic quality of political systems. Our data point to an overall poor trust in institutions among Latin Americans. Yet, although we do not observe great cross-country variability in levels of trust, our preliminary analysis suggests that larger variation exists within groups, and that is why we pay special attention to the individual level along other contextual factors. Our analysis endeavors to bring the role of the ethno-racial component and its associations with socioeconomic hierarchies in shaping political trust.
Our findings show that individual-level predictors along with contextual factors of ethno-racial and socioeconomic inequalities may help explain variations in political trust, but nuanced effects are found based on the institution and the connections between the dimensions of inequality. In our analyses, interaction terms significantly support Easton’s distinction between “diffuse” and “specific support,” with governmental trust uniquely demonstrating significance at both cross-level interactions and uniform statistical significance across racial categories at the individual level. The other three institutions present different relations, especially in the case of the judiciary where no statistically significant associations were found at the individual level, and the congress, where the cross-level interaction with the economic dimension reveals a positive relation with the Indigenous category. This underscores unique engagement patterns with political institutions, driven by perceptions of their broader principles or specific performances.
In line with our theoretical expectations, results suggest a consistent negative association between economic ethnic inequality and trust in the four institutions. The relation between sociopolitical ethnic inequalities and lower levels of political trust is not obvious since we did not find a direct effect in our analyses, coinciding with the mixed findings of recent works (Panzano Reference Panzano2023). Nevertheless, interacting this dimension with self-reported race does indicate that when inequality rises, trust decreases for all racial categories except for the confidence in the congress of the Indigenous membership. These mixed results require further systematic analyses, especially the case of the congress, and the interactions between BGI and the membership to marginalized groups.
Interestingly, contrary to our expectations, we found that in contexts of greater economic ethnic inequality, those self-reported as Indigenous exhibit higher trust than Whites. Even though this may be explained by the literature on political culture in multi-ethnic societies, future work on this subject should thus take care to compare our findings using additional measures and more rigorous methods. On the other hand, interaction terms at the individual level report relevant findings. The analyses suggest that, when perceived wealth levels increase, variations in trust are different depending on the ethno-racial category, with higher levels of confidence reported for Whites compared to Indigenous and Afro-descendants.
We believe this article has important implications for our theoretical understanding of the relations between structures of inequalities and political trust and may contribute to mass political culture studies in two main ways. First, our work shows that both socioeconomic and ethno-racial inequalities operate together and have implications for citizens’ relations with the political system. Hence, when studying inequality, especially in a region like Latin America, the ethno-racial component needs to be considered as a key component of the social structure. Our evidence shows that ethnicity may shape socioeconomic and political relations and entails important implications for power dynamics.
Second, we believe that our main finding is that, at the individual level, wealth affects political trust differently depending on the ethno-racial identification. We theorized that traditional racialized groups experience structural discrimination, which impacts their political attitudes and exerts varying effects when interacting with the socioeconomic condition. Overall, results support our theoretical argumentation that socioeconomic and ethno-racial inequalities are intertwined. We consider this finding relevant for citizenship formation since it suggests ethno-racial inequalities yield different outputs across groups and this fosters distinct relations with the political system.
To conclude, our study provides a useful framework for understanding political citizenship formation in contexts of inequality. Our findings suggest that it is important to investigate potential heterogeneities in citizens’ relations with the political system and the ethno-racial factor emerging as a relevant structural inequality in shaping political culture. The impacts of discrimination are not the same for every ethno-racial group and this may have important implications for democracy since, as our analyses reveal, it creates different types of relations with the system and, therefore, distinct varieties of political citizenship. Perpetuating racialization poses a threat to the democratic quality of the political systems. Scholars could further investigate why different forms and levels of political trust emerge in different contexts among groups and compare the results with other regions where ethno-racial inequalities are also relevant.
Supplementary Material
To view supplementary material for this article, please visit https://doi.org/10.1017/lap.2024.29
Acknowledgments
We are grateful to the editors and anonymous reviewers for insightful feedback and a rigorous review process that have significantly enhanced the quality of our manuscript. This research was conducted within the framework of a predoctoral contract co-funded by the University of Salamanca and the Santander Bank.
Competing interests
The authors declare they have no competing interests
Open access
