In October 2019, massive demonstrations erupted across Chile to protest inequality and out-of-touch political elites who pushed for a transportation fare hike that would have placed a substantial burden on a working class already in a precarious economic situation. These protests unfolded in one of the most economically unequal countries in the world (Pérez-Ahumada Reference Pérez-Ahumada2014; Posner, Patroni, and Mayer Reference Posner, Patroni and Francois Mayer2018), where labor representatives are often excluded from policymaking and political parties with historical ties to labor unions have a poor track record of successfully passing reforms to benefit the working class (Barría Traverso, Araya Moreno, and Drouillas Reference Barría Traverso, Araya Moreno and Drouillas2012; Frank Reference Frank2002). This economic landscape has produced a crisis of representation and increasing disenchantment with formal institutions of representation in Chile, especially among the working class (Olavarría Reference Olavarría2003; Siavelis Reference Siavelis2016).
As in Chile, citizens across Latin America are disillusioned with democracy. This discontent with the way democracy works is fueled by persistently high levels of economic inequality and a perception that political elites are disconnected from the daily concerns of the poor and working-class majorities. It is no wonder that higher levels of poverty and inequality are associated with lower support for democratic norms across the region (Carlin Reference Carlin2006). In fact, Latin America is one of the most unequal regions in the world (ECLAC 2018). Economic inequality has produced political inequality (Boulding and Holzner Reference Boulding and Holzner2021; Cole Reference Cole2018; Solt Reference Solt2008), where the majority of working-class people are poorly represented in legislatures across the region. In Chile, more than 75 percent of the population is working class, but only about 1 percent of legislators are workers.
More generally, Latin American legislators are drawn from a narrow set of elites who are largely out of touch with the everyday lives of average citizens (Taylor-Robinson Reference Taylor-Robinson2010). Upper-class politicians often fail to address the interests of the majority. When political institutions “become discredited in the eyes of citizens, populist presidents may displace these institutions as representatives of the people” (Taylor-Robinson Reference Taylor-Robinson2010, 14). Thus, it is no surprise that the poor and working class may turn to populist leaders who can claim to represent workers more authentically, even if these leaders have questionable democratic credentials.
Some of the most enduring political figures of the last two decades, like Luiz Inácio Lula da Silva of Brazil, Hugo Chávez of Venezuela, and Evo Morales of Bolivia, rose from poor, working-class backgrounds (Anria and Cyr Reference Anria and Cyr2017; Ellner and Hellinger Reference Ellner and Hellinger2003; Hunter Reference Hunter2010). Although these presidents garnered substantial attention worldwide, political inclusion and representation of poor and working-class citizens is not the norm. Indeed, an individual’s ability to rise from the working class to the presidency in several countries is exceptional in a region where politicians typically hail from the economic elite (Barnes and Saxton Reference Barnes and Saxton2019; Carnes and Lupu Reference Carnes and Lupu2015; Taylor-Robinson Reference Taylor-Robinson2010). In Latin America, where working-class citizens make up anywhere from 60 to 80 percent of the labor force, fewer than 5 percent of legislators have a working-class background (see Figure 1.1).
Figure 1.1 Share of working-class legislators in Latin America
Note: Bars represent the percentage of deputies with working-class backgrounds in the lower chamber of the legislature. Data come from the authors’ coding of the University of Salamanca’s Survey of Parliamentary Elites, which asks, “What was your occupation prior to being elected deputy?” The figure reflects waves 4 and 5, the most recent Parliamentary Elites in Latin America (PELA) data used in our analysis for each country.
This political marginalization of the working class undermines one of the fundamental norms of democracy: inclusion. The principle of inclusion suggests that those directly affected by policy outcomes should be included in the decision-making process (Young Reference Young2000). In the aforementioned example from Chile, political elites made a policy decision without including the perspectives of the working-class people most affected by it. This lack of inclusion raises the question: How does the exclusion of the working class from political power influence citizens’ perceptions of representation?
We argue that the exclusion of the working class contributes to a crisis of representation. At its root, the crisis of representation in Latin America, as well as in other democracies around the world, is a story about disenchantment with political parties and legislatures (Carlin Reference Carlin2006; Reference Carnes2018; Luna Reference Luna2016; Mainwaring Reference Mainwaring2006; Mair Reference Mair2013; Przeworski Reference Przeworski2019; Tanaka Reference Tanaka, Mainwaring, María Bejarano and Pizarro Leongómez2006). Despite the importance of these institutions in Latin America, they are generally viewed with distrust (Mainwaring, Bejarano, and Leongomez Reference Mainwaring, María Bejarano and Pizarro Leongómez2006; Seligson Reference Seligson2007). This distrust threatens democracy (Claassen Reference Claassen2020).
This book examines how the near exclusion of working-class citizens from legislatures affects how citizens perceive political representation. In doing so, we tackle three important questions. Our first question is: Do citizens want to be represented by members of the working class? We argue that yes! – voters want to be represented by working-class political representatives. The reason is twofold: (1) The presence of working-class legislators may signal a more inclusive policymaking process. That is, simply having more workers in office conveys to citizens that policymakers care about their experiences, preferences, and policy needs. (2) Given that working-class legislators often have different policy preferences and advance different policy agendas than middle- and upper-class representatives from the same political party, working-class representatives may also enhance policy responsiveness. Using survey data from across Latin America, and original survey data from Argentina and Mexico, we demonstrate that citizens want to see more workers in office, that the presence of workers in office is associated with better evaluations of legislatures and political parties, and that many people believe workers are better suited to represent the policy needs of working-class citizens.
On the face of it, this may seem intuitive; upon closer inspection, however, the aforementioned argument raises two additional questions. The presence of workers in office can only foster better-perceived representation via policy responsiveness if working-class legislators advocate on behalf of the working class. That said, legislators face a number of competing incentives and demands. Not all working-class legislators have the same motivations or opportunities to represent the working class. Consequently, we ask: Will any worker do? In other words, how do citizens evaluate workers who do not represent working-class policy interests? Likewise, simply having more working-class legislators in office can only alter how people feel represented if they are actually aware of working-class representation. For workers’ presence in office to signal a more inclusive policymaking process, citizens must be aware of workers’ presence in office. Yet working-class status is arguably more difficult for citizens to observe than characteristics such as race, ethnicity, or gender. Thus, we ask: How do citizens know workers are in office?
Will any worker do? We explain that even though all workers come to office having unique lived experiences that better position them to represent the needs of working-class citizens, once in office, workers are faced with a range of competing incentives and opportunities that sometimes preclude their desire and ability to represent working-class interests. We argue that for working-class legislators to effectively improve evaluations of representative institutions, they should have strong relationships with the disadvantaged group they represent and be committed to advocating for their policy interests (Dovi Reference Dovi2002). Where working-class politicians enter office and defect – prioritizing their party or other economic interests instead of the working class – we anticipate working-class representatives will invoke backlash and diminish evaluations of institutions. In contrast, where workers strive to cultivate working-class support and represent workers’ policy interests, we posit that working-class politicians will improve perceptions of institutions. Drawing on an original dataset of the class backgrounds of Argentine and Mexican national deputies over a twenty-year-period, survey data from the Latin American Public Opinion Project (LAPOP), and original survey experiments in Argentina and Mexico, we demonstrate that where representation of the working class is accompanied by policy responsiveness to workers’ interests, working-class representation improves evaluations of democratic institutions. Absent policy representation, however, workers’ presence in office only moderately improves evaluations of institutions, and in some cases, generates backlash.
This leads to our final question: How do citizens know workers are in office? We argue that even if people do not have perfect information, there are a number of ways that citizens learn about working-class representation (Bernhard Reference Bernhardn.d.; Bernhard and Freeder Reference Bernhard and Freeder2020; Hinojosa and Kittilson Reference Hinojosa and Caul Kittilson2020). Since parties and candidates have political incentives to make politicians’ working-class status known, news sources and governments make information about legislators available and digestible, and people are surprisingly adept at inferring class status from readily available heuristics such as facial images and speech patterns, we explain that citizens have a number of opportunities to learn about the presence – or absence – of workers in office. Leveraging examples from news sources, surveys of government websites, and a unique experiment in Argentina and Mexico that tests respondents’ abilities to classify the class status of deputies based only on facial images, we provide strong evidence that citizens are capable of detecting workers’ presence in office. Latin American citizens know workers are in office because they learn about the class status of politicians through numerous information sources, and they can infer class from images of their representatives. Furthermore, using survey data from across Latin America, we demonstrate that the relationship between working-class representation and positive evaluations of institutions is strongest among individuals who are most likely to be aware of legislators’ class status – that is, citizens with high levels of political interest and avid news followers.
1.1 Evidence of Working-Class Representation and Citizens’ Evaluations
We leverage a combination of qualitative and quantitative data to answer these three pressing questions. Our empirical analysis is situated in Latin America, where we analyze citizens’ evaluations of representative institutions – namely, legislatures and political parties – across the entire region to capture variation in working-class representation. We draw on quantitative evidence from elite surveys with legislators and mass public opinion surveys across eighteen countries from 2008 to 2014 for a total of forty-eight country-years. We bolster this cross-national evidence with original survey experiments and novel survey questions that we fielded in Mexico and Argentina in 2019, and a newly constructed dataset of the class backgrounds of over 4,400 Mexican deputies serving from 1997 to 2018 and about 1,800 Argentine deputies from 2002 to 2016.
The qualitative evidence comes from multiple sources as well. First, in 2019, more than 5,000 individuals in Argentina and Mexico responded to open-ended survey questions explaining their perceptions of class representation. We utilize this rich collection of citizen-level responses throughout the book to contextualize many of the empirical findings from our quantitative analyses. Second, we provide in-depth case studies of the history of working-class representation in Argentina and Mexico to explain key variation in workers’ ties to political parties and in representatives’ track record of policy representation. Finally, throughout the book – and particularly in Chapters 5 and 6 – we draw heavily on historical examples and vignettes to illustrate our findings and bolster our conclusions. Although we do not rely on the open-ended survey responses or examples to make inferences or empirical claims – all our conclusions are drawn from the quantitative analyses – they help to illustrate our empirical results.
1.1.1 Leveraging Cross-National Data from Latin America
To empirically test our expectations about working-class inclusion, we utilize a variety of cross-national data sources. From a research design perspective, cross-national data are important for several reasons. First, by drawing on elite and public opinion survey data across forty-eight country-years, we leverage far more variation than would be available if we were only looking at one or two cases in depth. The elite data come from the University of Salamanca’s Survey of Parliamentary Elites (PELA), which conducts anonymous surveys of a representative sample of legislators in each legislative session from all eighteen countries included in our cross-national analysis. Mass public opinion data come from LAPOP’s AmericasBarometer, which has been conducting nationally representative surveys across Latin America since 2004. Using these two sources of cross-national survey data, we capture important variation across space and time in working-class representation from the PELA surveys (our key independent variable), as well as citizens’ evaluations of political institutions from the LAPOP surveys (our key dependent variable). This variation is critical for empirically testing our expectations and drawing valid inferences (King, Keohane, and Verba Reference King, Keohane and Verba1994).
Second, cross-national data are important for demonstrating the generalizability of our findings. In particular, we demonstrate that working-class legislators improve citizens’ perceptions of representation across a variety of economic contexts and institutional arrangements, including different electoral rules, party systems, levels of development, and labor union density. In addition to elite and mass public opinion survey data, we also draw on vignettes and journalistic accounts from across the region to contextualize our statistical analysis and to demonstrate the generalizability of our findings.
1.1.2 Original Surveys from Argentina and Mexico
Scholars interested in understanding the relationship between representation and people’s evaluations of institutions are plagued with the challenge of identifying causality. Even when taking the best of care to measure variables sequentially, maximizing variation in both the dependent and independent variables, and considering potential sources of spuriousness, scholars studying political behavior and public opinion are thwarted by threats of endogeneity – particularly reverse causality and omitted variable bias. With respect to reverse causality, it is entirely plausible that where citizens feel better represented by parties and legislatures, workers are more likely to pursue political office. Of course, in our cross-national analysis, we attempt to address this with a careful research design that measures working-class representation in the time period before citizens are asked to evaluate institutions, but both trends in citizens’ attitudes and workers’ access to office tend to change slowly over time. Thus, it is possible that this process is reversed.
As for omitted variable bias, most social scientists (and curious people in general) can easily think of a number of factors that may theoretically improve both workers’ access to office and people’s evaluations of representative institutions – thus potentially explaining away the link we observe between these two factors. Although we do our best to address potential omitted variables by controlling for observables in our cross-national analyses, there may be unobservable factors we cannot account for – a known limitation of correlational studies. Beyond these limitations, the questions we can ask are constrained by data availability in publicly available surveys and respondents are constrained in their answers by predefined multiple-choice options.
To address these challenges and improve our ability to draw inferences, we designed and fielded original surveys in Argentina and Mexico that contain a series of novel survey experiments, original survey questions, and open-ended survey questions. In particular, we designed two survey experiments: one to assess the causal mechanism linking higher levels of working-class representation to individuals’ perceptions of representation, and a second to evaluate citizens’ ability to identify working-class representatives. We developed new questions to assess individual evaluations of working-class deputies, and we incorporated open-ended survey questions to provide insights into the underlying factors that individuals claim inform their perceptions of working-class deputies.
1.1.3 Observing Working-Class Representation
How do we identify a member of the working class? The concept of class can be somewhat ambiguous, but, fortunately, researchers have developed numerous ways to conceptualize and operationalize which legislators come from the working class. Similar to various other studies on the class backgrounds of political elites, we rely solely on legislators’ occupational status, rather than relying on income or some other measure of socioeconomic well-being. Although most legislators would be coded as belonging to the same (elite) class if we relied on income-based measures of class, the reality is that people with different occupational backgrounds have different life experiences, opportunities, social circles, and economic and political interests (Evans and Tilley Reference Evans and Tilley2017; Hout, Manza, and Brooks Reference Hout, Manza and Brooks1995; Mood Reference Mood2017; Stephens, Markus, and Phillips Reference Stephens, Rose Markus and Taylor Phillips2014; Weeden and Grusky Reference Weeden and Grusky2005).Footnote 1
Occupation-based conceptualizations of class are fundamentally different from socioeconomic-based approaches (Kitschelt and Rehm Reference Kitschelt and Rehm2014). This distinction between socioeconomic-based and occupational approaches is paramount for evaluating both legislators’ ties to citizens and the policies they represent when in office. Above and beyond a person’s income, the work people do to earn a living establishes their position in society (Manza and Brooks Reference Manza, Brooks, Lareau and Conley2008). Indeed, the theory of occupational socialization argues that time spent in an occupation should shape political preferences and behavior (Barnes, Beall, and Holman Reference Barnes, Beall and Holman2021; Barnes and Saxton Reference Barnes and Saxton2019; Best and Cotta Reference Best and Cotta2000; Carnes Reference Carlin2012, Reference Carnes2013; Oesch and Rennwald Reference Oesch and Rennwald2018; O’Grady Reference O’Grady2019; Vivyan et al. Reference Vivyan, Wagner, Glinitzer and Eberl2020). People who work in the same industry have shared experiences that both define what they see as the range of potential political concerns and also influence their preferences and priorities.
In terms of shared experiences, the socialization that members of different social classes receive in their occupations plays a fundamental role in shaping distinct preferences and priorities. Individuals employed in working-class occupations often have lower incomes, face heightened employment insecurity, and are more likely to depend on social safety net programs, such as unemployment insurance, that are funded via redistributive policies and require state intervention in the economy (Evans and Tilley Reference Evans and Tilley2017). It is perhaps unsurprising, then, that individuals’ occupations often correlate with their social policy preferences (Kitschelt and Rehm Reference Kitschelt and Rehm2014). This same relationship between citizens’ occupations and policy preferences is evident among legislators as well. When legislators hail from working-class backgrounds, they too experience a similar socialization process from their occupations, and thus have fundamentally different policy preferences, especially regarding economic issues, than their colleagues with white-collar occupational backgrounds (Barnes, Beall, and Holman Reference Barnes, Beall and Holman2021; Carnes Reference Carlin2012; O’Grady Reference O’Grady2019).
While we identify members of the working class based on their occupation, we do not further subdivide workers between those who work in the informal and formal sectors of the economy in this study. In Latin America, anywhere from a quarter to around half the population in most countries is employed in the informal sector. It is possible that members of the working class from the informal and formal sectors have distinct preferences based on their occupational socialization. Access to social safety net policies, such as unemployment insurance or health insurance, is sometimes limited to formal sector workers, although reforms have been made to increase access to these benefits for informal workers in several countries (Posner, Patroni, and Mayer Reference Posner, Patroni and Francois Mayer2018).
Nonetheless, research that examines differences between the political preferences of formal and informal sector workers in Latin America finds very few differences, as these two sectors of the economy in Latin America are highly integrated (Baker and Velasco-Guachalla Reference Baker and Ximena Velasco-Guachalla2018; Palmer-Rubin and Collier Reference Palmer-Rubin and Berins Collier2022). Workers may shift from one sector to the other over time, with many workers “participate[ing] in a mix of formal and informal activities” in the labor market (Hummel Reference Hummel2017, 1525). Similar to formal workers, informal workers organize in most countries. They are often unionized, and they bargain directly with the state over self-regulation (Holland Reference Holland2015; Hummel Reference Hummel2017, Reference Hummel2021). The adoption of neoliberal reforms across Latin America has made formal sector employment much more precarious and unstable, lessening differences across sectors. Thus, policy preferences are unlikely to vary between workers in formal and informal sector occupations.Footnote 2
By taking occupational socialization as our starting point for examining working-class inclusion, we get a clearer understanding of whether representatives will advocate for and promote the policy interests of working-class citizens. As Manza and Brooks (2008, 204) explain, “[w]orkplace settings provide the possibility of talking about politics and forging political identity, and work also provides a springboard for membership in organizations where class politics are engaged: unions, professional associations, business associations, and so forth.” Phillips (Reference Phillips1995) further explains in her theory of the politics of presence, shared life experience is an anchor for understanding the representation that constituents receive from their legislators. In other words, an occupation-based conceptualization of class is critical for evaluating how class identity shapes representation.
Following this increasingly accepted convention for conceptualizing and operationalizing class, we consider legislators as belonging to the working class if they previously earned a living by working in skilled and unskilled manual labor, as small artisans, in service-industry occupations, in entry-level secretarial positions, as rural laborers/small farmers (campesinos), and as union officers prior to entering politics.Footnote 3 Due to the wide variety of unionized occupations, we classify all deputies with backgrounds in union leadership as workers, even if they led teacher unions or public sector unions. Deputies who are teachers, bureaucrats, or nurses who belong to unions, but are not in union leadership, are not classified as workers. In order to distinguish between farmers who may also be large landowners, and rural laborers or poor farmers on small plots of land, we classify all deputies in the rural sector with ties to peasant organizations as workers, and all others as farmers. Throughout this study, we use the terms “worker” and “working class” interchangeably when we refer to legislators with working-class occupational backgrounds.
Region-wide data on working-class representation are available from the PELA surveys. We rely on these surveys in our cross-national analyses, but these data are limited in two respects: first, in their time span, and second, in only including information on legislators’ most recent occupation. These two limitations pose a risk of biasing our results towards the null if legislators with working-class backgrounds are not coded as such. Thus, we bolster our cross-national investigation with an in-depth look at Argentina and Mexico. Here sufficient resources are available to accurately estimate the numeric representation of workers in legislative office across time using official biographical data of federal legislators. These data allow us to engage in more detailed time-series analyses of the effect of working-class representation on citizens’ evaluations of their political institutions.
1.1.4 Observing Citizens’ Evaluations of Representatives
The primary purpose of this book is to understand how the exclusion/inclusion of working-class legislators in office shapes citizens’ perceptions of representative institutions. Since we are interested in citizens’ perceptions about the quality of political representation they receive, we must rely on some public opinion measures that capture these perceptions. Conceptually, we are interested in measures that can be clearly interpreted as evaluations of the quality of representation citizens receive from their elected representatives in legislatures and political parties. There is, however, little consensus among scholars about how to measure the extent to which citizens feel represented.
Our strategy is to rely on multiple evaluative measures of political parties and legislatures, such as the level of trust individuals have in legislatures and political parties, congressional job approval, and how well parties represent or listen to voters. Our original survey directly asks respondents the extent to which they feel working-class/white-collar legislators understand their problems and promote projects to improve the quality of life of all citizens, followed by open-ended questions. The qualitative responses allow us to further probe why the respondents feel the way they do about representation. Throughout the book, we refer to this collection of measures as evaluations of representatives or perceptions of representation interchangeably. If we find that the class backgrounds of legislators have similar effects across multiple evaluative measures of representation, we can be more confident in our findings.
Other work often refers to perceptions of representation as “symbolic representation,” a term derived from Pitkin (Reference Pitkin1967), and which is generally concerned with the symbolic effects that representatives have on the represented. While our work is similar to much of the symbolic representation literature, we shy away from this term here, since it has been used to encompass a wide range of outcomes.Footnote 4
1.2 A Closer Look at Argentina and Mexico
As previously noted, we fielded original surveys in Argentina and Mexico and collected an extensive new dataset of national deputies’ occupational backgrounds in these two countries. Argentina and Mexico are excellent cases that reflect general trends of economic inequality in the region. In both cases, class identity is highly salient, as it is across much of Latin America. Equally important, both cases are characterized by a crisis of representation where large parts of the population are disillusioned with representative institutions. Given that in both Argentina and Mexico we observe trends that are taking place across Latin America, these two cases are useful for drawing more general conclusions about the influence of working-class descriptive representation on citizens’ attitudes.
In addition, for our research, we needed to identify cases where (1) data were available on working-class representation, (2) there was a sufficient number of working-class representatives in office to study, and (3) there was variation in our primary explanatory variables of interest. Our theory focuses on the combined effect of working-class inclusion and working-class policy representation on how citizens evaluate government. To study the effects of both inclusion and policy representation, we need to select cases where these two factors vary (i.e., variation on our two independent variables: working-class representation and policy representation). Both Argentina and Mexico address these needs better than other cases in the region.Footnote 5
Few Latin American legislatures have extensive accessible databases on the biographical characteristics of their members. The advantage of focusing on Argentina and Mexico is that individual-level biographical information exists in both countries for roughly a twenty-year period, and other scholars have been studying these institutions for decades. The availability of secondary sources that also focus on the occupational background of legislators – allowing us to code the class status of deputies – is an important supplement to our cross-national analysis. We constructed original datasets on the occupational backgrounds of national legislators in Mexico and Argentina over a period comprising more than twenty years.Footnote 6 Drawing on candidate biographies and more extensive professional profiles means we do not have to rely exclusively on the job they held immediately prior to running for office, which helps us overcome one of the limitations of the cross-national PELA data. Instead, we are able to use information from the course of a legislator’s career before entering politics.
In Argentina, we draw on data compiled by the Directorio Legislativo – a nongovernmental organization that began cataloguing national deputies’ self-reported professional backgrounds in 2000 (Barnes and Holman Reference Barnes and Saxton2019; Reference Barnes and Holman2020a; Carnes and Lupu Reference Carnes and Lupu2015; Franceschet and Piscopo Reference Franceschet and Piscopo2008; Micozzi Reference Micozzi2018). Specifically, legislators are asked to report their prior political offices, professional occupations, and party experience, as well as educational background and some personal information.
In Mexico, we rely upon information from the Sistema de Información Legislativa – a government website maintained by the Interior Ministry, which archives legislative profiles for national deputies. For Mexico, these data were originally compiled in 2010 and have been maintained since (Kerevel Reference Kerevel2015; Reference Kerevel2019). The website reports deputies’ previous political, public, and private careers; educational attainment; and some personal information. For both Argentina and Mexico, we also relied upon other published biographies and publicly available data when there was no other information recorded for a given deputy.
Having adequate data on the occupational background of legislators in Argentina and Mexico would not be particularly useful for this study if there were hardly any working-class deputies in office. Fortunately, in both countries, there is a substantial history of workers in office, a sufficient number of identifiable working-class representatives in recent years, and substantial variation in working-class representation across space and time within cases. While working-class representation is slightly higher in Mexico and Argentina than in several other countries in Latin America, the percentage of workers in office in both countries is similar to the average rate (5 percent) of working-class representation across the region (see Figure 1.1).
Figure 1.2 shows how the percentage of legislators with working-class backgrounds has changed over time in Argentina and Mexico from the early twentieth century.Footnote 7 The first thing to notice in Figure 1.2 is that with the brief exception of the 1940s and 1950s, workers have been present in higher numbers in the Mexican Congress than in the Argentine Congress. The decline in working-class representation in Mexico’s Chamber of Deputies parallels an increase in the representation of those from the business sector, and an increase in professional politicians and government bureaucrats.
Figure 1.2 Percentage of working-class deputies
Unfortunately, available data on the class backgrounds of Argentine legislators prior to 2000 is less reliable than in the Mexican case due to the lack of systematic sources of biographies of elected legislators.Footnote 8 However, based on the best available estimates, workers’ presence in the Argentine Congress has steadily declined since the 1970s. In recent decades, the representation of the working class in the Argentine Chamber of Deputies ranges from 2 to 4 percent over the 2002–2018 period compared to over 70 percent in the population; in the Mexican Chamber of Deputies, working-class representation ranges from 6 to 12 percent from 1997 to 2018 compared to making up over 80 percent of the population (Carnes and Lupu Reference Carnes and Lupu2015).
In addition to variation in working-class representation over time in Argentina and Mexico, both countries also currently exhibit variation in the share of national deputies that hail from different state/provinces across the country. Figure 1.3 illustrates this spatial variation in working-class representation. For each subnational unit, we calculated the average percentage of national deputies from the working class who were elected during each legislative session in our dataset. Darker colors indicate states/provinces where more workers were historically elected, and lighter colors correspond to fewer (or no) workers elected to the national legislature. Figure 1.3 reveals that in Argentina, some provinces have elected either no or few workers to the National Congress over the last two decades, whereas others such as Buenos Aires have led the pack, electing about 10 percent of workers on average into the legislature. In terms of spatial patterns in Argentina, the highest concentrations of working-class deputies are in the middle of the country, including the province of Buenos Aires which is where most of the country’s population and industry are located (Buenos Aires province accounts for nearly 50 percent of the country’s total gross domestic product [GDP]). The other large concentration of workers is in southern Argentina, particularly in some of the Patagonian provinces where oil and natural gas production dominate the local economy.
Figure 1.3 Percent of national deputies from each state who hail from the working class, Argentina and Mexico
Note: This figure shows the average percentage of national deputies from state/province who hail from the working class during the period included in our dataset.
In Mexico, many elected workers are from teachers unions and public sector unions, which are active in every state. Thus, the representation of workers across Mexico is higher than what we observe in Argentina. However, we do observe larger concentrations of workers in some states over others for two principal reasons. One, the agricultural sector and the oil sector both concentrate in certain states and lead to more elected peasant and oil worker representatives. Agricultural states such as Sinaloa and Nayarit on the Pacific coast (in the darker shades), and Veracruz and Puebla (the two darkest-shaded states, with 14 and 13 percent working-class deputies, respectively, in Figure 1.3), elect a higher percentage of workers. In addition, the oil sector concentrates in the Gulf of Mexico. There are higher percentages of workers in nearly every state along the Gulf, such as Tamaulipas, Veracruz, and Campeche. Two, we observe more workers elected in states where the Institutional Revolutionary Party (PRI) has historically performed well. In the previously mentioned states with higher working-class representation, Sinaloa, Nayarit, Veracruz, Puebla, Tamaulipas, and Campeche, anywhere from 55 to 63 percent of all elected deputies were from the PRI. In contrast, Tabasco, along the Gulf Coast and involved in the oil sector, has seen greater representation from the Party of the Democratic Revolution (PRD) and other parties. Only 42 percent of deputies from Tabasco are from the PRI, which decreases opportunities for oil workers affiliated with the PRI-aligned oil workers union. Other states, with proportionally fewer workers, such as Zacatecas, Baja California, Yucatan, and Quintana Roo, are all much more competitive for the PRD and the National Action Party (Partido de Acción Nacional [PAN]), with large majorities of deputies from these states coming from these two parties.
The final reason we focus on Argentina and Mexico is that both cases present variation in the policy representation of the working class. Drawing on an extensive literature (e.g., Bensusán Reference Bensusán2016; Burgess Reference Burgess1999; James Reference James1988; Levitsky Reference Levitsky2003; McGuire Reference McGuire1997; Middlebrook Reference Middlebrook1995; Murillo Reference Murillo2001; Posner, Patroni, and Mayer Reference Posner, Patroni and Francois Mayer2018), which we discuss in detail in Chapter 5, both cases exhibit variation in how well the working class is represented in the policymaking process. This variation exists not only over time within Argentina and Mexico but also across cases. Generally the working class is better represented in the policymaking process in Argentina compared to Mexico. Had we selected cases where workers are consistently well-represented or consistently poorly represented in the policymaking process, we would be unable to say much about the importance of working-class policy representation due to the lack of variation on this key independent variable.
Specifically, we explain in Chapter 5 that several features of union organization across the two cases structures unions’ abilities to effectively extract policy concessions that advance workers’ rights (Chambers-Ju Reference Chambers-Ju2021; Collier and Collier Reference Collier and Collier2002; Collier and Chambers-Ju Reference Collier, Chambers-Ju, Kingston and Yashar2012; Murillo Reference Murillo2001). In addition, the larger set of institutions that govern union–state relations and linkages between political parties, labor leaders, and workers shapes unions’ incentives to either represent workers’ policy concerns or abandon workers in favor of representing party, state, and/or business interests. This variation in incentives to represent the working class is necessary to address how variation in the policy responsiveness of working-class legislators influences citizen perceptions of representation.
1.3 The Importance of Studying Working-Class Representation in Latin America and Beyond
Latin America is an ideal setting to investigate our questions related to representation of the working class and citizens’ evaluations of political institutions. Political institutions draw their members from the upper strata of society in democracies the world over, but the gap between the haves and the have-nots is particularly evident in Latin American politics. Indeed, the political consequences of persistently high economic inequality are magnified in this region where class is one of the most politically salient social cleavages. Despite the widespread incorporation (cooption) of organized labor in Latin American party systems (Collier and Collier Reference Collier and Collier2002) and thriving civil societies that tether poor citizens to the political system (Boulding and Holzner Reference Boulding and Holzner2020; Reference Boulding and Holzner2021), policy outcomes continue to fail at remedying massive economic inequalities in the region. Moreover, the policy shift towards neoliberalism in the 1980s and 1990s weakened unions and party ties to organized labor (Levitsky Reference Levitsky2003; Silva and Rossi Reference Silva and Rossi2018), leading to declines in the election of workers to public office. In this section, we elaborate on why Latin America is an excellent region in which to situate our study, while also highlighting how the questions we ask are relevant beyond the region.
Although the political salience of class varies to some extent across the region (Roberts Reference Roberts2002), Carnes and Lupu (Reference Carnes and Lupu2015) find substantial overlap between citizens’ class identity and political preferences in the vast majority of Latin American countries. Nevertheless, seats in Latin American legislatures are predominantly occupied by upper-class politicians whose class-based political preferences do not align with the majority of citizens. The average gap between the percentage of legislators with working-class backgrounds and the percentage of working-class citizens in a given country is around 70 percent (Carnes and Lupu Reference Carnes and Lupu2015). Though class is more salient in Latin America, a mismatch between citizens’ policy preferences and elites’ preferences is also observed in more developed countries. In the United States and much of Western Europe, economic policy also fails to reflect the interests of the “median voter,” because policy is not made by the median voter. Instead, policymakers are most responsive to constituents who share their own policy interests – members of the upper class (Bartels Reference Bartels2016; Cagé 2020; Gilens and Page Reference Gilens and Page2014; Piketty Reference Piketty2020).
In many ways, the political salience of class across the region makes Latin America an ideal case for examining the connection between economic and political inequalities (Boulding and Holzner Reference Boulding and Holzner2021). The vast majority of citizens in Latin America favor state intervention to correct for economic inequalities (Barnes and Córdova Reference Barnes and Córdova2016). LAPOP routinely asks respondents to indicate their agreement with the statement: “The state should enact firm policies to reduce inequality.” In the 2016/17 wave of the LAPOP survey, the average level of support for this statement was 71 percent. We argue that the near exclusion of the working class from legislatures in Latin America contributes to this disconnect between mass public opinion and redistributive policy outcomes. Indeed, prior research on attitudes about inequality in Latin America finds that the “masses” and “elites” have differing perceptions about the causes of inequality, as well as different policy preferences when it comes to combating inequality (Blofield and Luna Reference Blofield, Pablo Luna and Blofield2011; Reis Reference Reis and Blofield2011). This disconnect between public opinion about inequality and actual policy outcomes parallels trends in other democracies such as the United States (Bartels Reference Bartels2016; Gilens Reference Gilens2012; Hacker and Pierson Reference Hacker and Pierson2010) and across Europe (Kaltenthaler, Ceccoli, and Gelleny Reference Kaltenthaler, Ceccoli and Gelleny2008; Piketty Reference Piketty2020). It thus appears that citizens across a wide variety of countries would prefer governments to address economic inequality, while political elites often continue to promote policies that increase inequality.
In addition to enduring class divides and persistently high economic inequality, Latin America is suffering from what scholars have called a “crisis of representation” that is often characterized by widespread disillusionment with representative institutions (Boulding Reference Boulding2014; Boulding and Nelson-Núñez Reference Boulding and Nelson-Núñez2014; Hinojosa and Kittilson Reference Hinojosa and Caul Kittilson2020; Luna Reference Luna2016; Mainwaring Reference Mainwaring2006; Mair Reference Mair2013; Przeworski Reference Przeworski2019; Tanaka Reference Tanaka, Mainwaring, María Bejarano and Pizarro Leongómez2006). Institutional trust is vital for democratic legitimacy and stability (Cleary and Stokes Reference Cleary and Stokes2006; Linz and Stepan Reference Linz and Stepan1996). Although democratic governments can withstand short-term periods of dissatisfaction with specific political institutions (Hetherington Reference Hetherington1998), prolonged periods of distrust can erode citizens’ support for democratic principles (Easton Reference Easton1975). Recent research demonstrates political parties in Latin America are able to adopt their preferred policies while appealing to poor and working-class voters through the use of clientelism and “non-policy endowments,” such as access to patronage resources (Calvo and Murillo Reference Calvo and Victoria Murillo2019; Nichter Reference Nichter2018; Oliveros Reference Mood2012). While this strategy helps political parties appeal to working-class voters during elections, the inability or unwillingness of successive governments to actually address economic inequality results in widespread disillusionment with representative institutions (Mainwaring Reference Mainwaring2006).
Recent scholarship highlights declines in democracy in consolidated regimes from India to the United States, and around the globe more broadly (Levitsky and Ziblatt Reference Levitsky and Ziblatt2018; Przeworski Reference Przeworski2019). Przeworski (Reference Przeworski2019) explains, for instance, that in many countries, democratic declines are gradual, with elected officials slowly subverting democratic norms and institutions. In Latin America, these global trends have become especially acute. Between 2016 and 2018, eleven out of eighteen Latin American countries experienced declines in their “Electoral Democracy Index” score, as measured by the Varieties of Democracy (V-Dem) project. Following the completion of the 2018/19 LAPOP wave, researchers at Vanderbilt University published the Pulse of Democracy report and highlighted several other troubling trends (Zechmeister and Lupu Reference Zechmeister and Lupu2019). Particularly, the authors of the report explain that support for democracy declined significantly in 2016 and then continued to decline in the 2018/19 public opinion survey. In 2008, nearly 70 percent of the respondents in Latin America agreed that “despite its problems, democracy is the best form of government.” By 2018, that figure had reached a low of 58 percent. Similarly, less than 40 percent of Latin Americans in 2018 said they were satisfied with the way their country’s democracy functions (down from a high of nearly 60 percent in 2010), and trust in political parties and the legislature in 2018 were both at their lowest levels since LAPOP began collecting public opinion data.
This erosion of support for democratic institutions is perhaps unsurprising. High levels of inequality are associated with decreased satisfaction with democracy not only in Latin America (Saxton Reference Saxton2019) but also in democracies more generally (Anderson and Singer Reference Anderson and Singer2008). Some scholars have even gone so far as to say that inequality acts as a “referendum” on democracy and thus erodes citizens’ attachments to democratic principles (Krieckhaus et al. Reference Krieckhaus, Son, Mukherjee Bellinger and Wells2014). We contend that issues around class politics can further help us understand the ongoing crisis of democratic representation in Latin America. When large segments of the population, such as the working class, are excluded from political representation, this signals that democracy is not functioning as it was intended and is “intrinsically unfair” (Williams Reference Williams1998). Moreover, when white-collar legislators are left to make policy for working-class citizens, it is less likely the interests of working-class citizens – or the average citizen for that matter – will be meaningfully represented (Mansbridge Reference Mansbridge2015).
For these reasons – the salience of class identity, high inequality, and a crisis of democratic representation – Latin America is an ideal setting to test our expectations about working-class inclusion and citizens’ evaluations of political institutions. Nevertheless, the trends we observe in Latin America are global trends. Class divisions exist in most societies, and economic inequality is increasing around the world. Moreover, disenchantment with the present form of liberal representative democracy exists nearly everywhere. Thus, while Latin America is an ideal setting to answer the questions we pose, the trends in the region are certainly not unique.
Our findings and conclusions are of relevance outside Latin America. There are several contextual factors unique to the region, however, that may limit the generalizability of our findings. First, Latin America is one of the most economically unequal regions in the world, with workers comprising the majority of the population. We expect our findings will extend to places where economic inequality is higher, while they may be weaker in countries with greater economic equality as there may be fewer unmet policy needs of the poor and working class. Second, class has historically been a central political cleavage in Latin America. In countries where other identities and cleavages may be more salient than class, the presence of workers in office may not have similar effects on citizens’ attitudes. Finally, Latin America has a unique political culture (Inglehart and Welzel Reference Ingelhart and Welzel2020), with high levels of support for government policies that promote economic redistribution (Barnes and Córdova Reference Barnes and Córdova2016). In cases where citizens are more skeptical of state involvement in the economy, there may be a weaker relationship between the class backgrounds of politicians and evaluations of democratic institutions. Future research will need to determine the extent to which our findings travel outside the region.
1.4 Our Contributions
In the face of rising global inequality, questions surrounding the link between class and unequal access to government representation are paramount. This book makes several contributions to our understanding of inequality and democratic representation. A growing body of work examines the link between political inequality and economic inequality (e.g., Bartels Reference Bartels2016; Cagé Reference Cagé2020; Hacker and Pierson Reference Hacker and Pierson2010; Page and Gilens Reference Page and Gilens2017; Piketty Reference Piketty2020; Solt Reference Solt2008). Although multiple studies demonstrate that the overrepresentation of economic elites in politics has important policy consequences, only recently has research specifically addressed the issue of whether the poor and working class want to be represented by other poor, working-class citizens (e.g., Barnes and Saxton Reference Barnes and Saxton2019; Carnes and Lupu Reference Carnes and Lupu2016a; Vivyan et al. Reference Vivyan, Wagner, Glinitzer and Eberl2020). Our findings provide new evidence that clearly demonstrates citizens know workers are dramatically underrepresented, and they would prefer to be represented by more workers, not by the upper class. Our findings also suggest citizens may favor policy interventions designed to improve working-class representation such as the adoption of socioeconomic quotas, as proposed by Julia Cagé (Reference Cagé2020).
Until now, most research on working-class representation has examined how the near exclusion of working-class legislators shapes legislative behavior and public policy. Earlier work concluded that the dominance of the upper-class in legislatures had little influence on policy outcomes (e.g., Matthews Reference Matthews, Loewenberg, Patterson and Jewell1985; Norris and Lovenduski Reference Norris and Lovenduski1995; Putnam Reference Putnam1976). More recent work challenges this conclusion. For instance, Carnes (Reference Carnes2013) shows that legislators’ class background has profound consequences for the policymaking process in the United States, as working-class legislators vote differently on economic issues compared to those from other class backgrounds, even compared to legislators from the same district and party. Similarly, Bartels (Reference Bartels2016) demonstrates that economic policy outcomes in the United States disproportionately benefit upper-class citizens and have contributed to growing inequality since the mid-twentieth century. In the US state legislatures, Barnes, Beall, and Holman (Reference Barnes, Beall and Holman2021) show that legislatures with more women representatives from low-status occupations allocate larger shares of their budget to education and healthcare services. Conversely, Kirkland (Reference Kirkland2021) finds that white-collar mayors in the United States are more likely to reduce redistributive spending.
Outside the United States, O’Grady (Reference O’Grady2019) shows that even in the highly disciplined UK Parliament, working-class members of parliament (MPs) give different speeches and vote differently compared to co-partisans from other class backgrounds. Micozzi (Reference Micozzi2018) shows that union-affiliated legislators in Argentina sponsor more bills related to working-class interests. Alexiadou (Reference Alexiadou2022) finds the presence of cabinet ministers from working-class backgrounds, as well as those who were teachers and social workers, leads to increased generosity of social welfare benefits across parliamentary democracies. All told, most research on working-class representation focuses on legislative behavior and policy consequences.
We depart from this tradition to examine the previously overlooked implications of working-class representation for citizens’ evaluations of and attachments to democratic institutions. In doing so, our book also contributes more generally to literature on symbolic representation. Many studies have considered how descriptive characteristics of legislators such as gender, race, or ethnicity influence how citizens feel they are represented (e.g., Atkeson Reference Atkeson2003; Barnes and Burchard Reference Barnes and Burchard2013; Bobo and Gilliam Reference Bobo and Gilliam1990; Clayton, O’Brien, and Piscopo Reference Clayton, Josefsson, Mattes and Mozaffar2019; Dos Santos and Jalalzai Reference Dos Santos and Jalalzai2021; Gay Reference Gay2002; Hinojosa and Kittilson Reference Hinojosa and Caul Kittilson2020; Karp and Banducci Reference Karp and Banducci2008; Pantoja and Segura Reference Pantoja and Segura2003; Schwindt-Bayer and Mishler Reference Schwindt-Bayer and Mishler2005; Stauffer Reference Stauffer2021). To date, little research has considered how representatives’ class backgrounds influence citizens’ attitudes.
Earlier research asserted the exclusion of the working class from legislatures was likely to influence how citizens felt represented and the legitimacy of legislatures, but little empirical work was ever conducted (Norris and Lovenduski Reference Norris and Lovenduski1995, 224; Putnam Reference Putnam1976, 44). The rare exceptions come from recent work on the United Kingdom that suggests citizens there have perceived changes in the descriptive representation of the working class over time, and that these changes have impacted vote choice and the extent to which they perceive policy differences across parties (Evans and Tilley Reference Evans and Tilley2017; Heath Reference Heath2015).
Similarly, this research makes a key addition to work that seeks to understand citizens’ evaluations and trust of democratic institutions. There are a number of potential factors that may influence evaluations of democracy and democratic institutions, such as the presence of consensus or majoritarian political institutions (Lijphart Reference Lijphart1999), whether citizens support losing parties in previous elections (Anderson et al. Reference Anderson, Blais, Bowler, Donovan and Listaug2005; Conroy-Krutz and Kerr Reference Conroy-Krutz and Kerr2015), a government’s economic performance (Booth and Seligson Reference Booth and Seligson2009; Dalton Reference Dalton2004; McAllister Reference McAllister and Norris1999), perceptions of corruption (Anderson and Tverdova Reference Anderson and Tverdova2003; Dahlberg, Linde, and Holmberg Reference Dahlberg, Linde and Holmberg2015), and the level of economic inequality in a society (Carlin Reference Carlin2006; Córdova and Layton Reference Córdova and Layton2016; Saxton Reference Saxton2021; Zmerli and Castillo Reference Zmerli and Carlos Castillo2015). Previous studies found that greater representation of women and racial and ethnic minorities increases trust in democratic institutions (Alexander and Jalalzai Reference Alexander and Jalalzai2020; Badas and Stauffer Reference Badas and Stauffer2018, Reference Badas and Stauffer2019; Barnes and Taylor-Robinson Reference Barnes, Taylor-Robinson, Alexander, Bolzendahl and Jalalzai2018; Dos Santos and Jalalzai Reference Dos Santos and Jalalzai2021; Hayes and Hibbing Reference Hayes and Hibbing2017; Jalalzai and Dos Santos Reference Jalalzai and Dos Santos2015; Karp and Banducci Reference Karp and Banducci2008; Liu and Banaszak Reference Liu and Ann Banaszak2017; Tate Reference Tate2001; Ulbig Reference Ulbig2007). Despite a wide body of work dedicated to explaining variation in evaluations of democratic institutions, prior to our work, few studies have identified a link between the class backgrounds of representatives and trust in democratic institutions (Barnes and Saxton Reference Barnes and Saxton2019).
We contribute not only by expanding the types of characteristics that citizens are likely to consider in evaluating the quality of representation they receive but also by providing substantial evidence that class is a visible characteristic of which citizens are aware, in addition to gender, race, and ethnicity. In considering the link between descriptive representation and citizens’ attitudes and evaluations, recent work emphasizes the importance of a descriptive representatives’ visibility (Alexander Reference Alexander2015; Barnes and Taylor-Robinson Reference Barnes, Taylor-Robinson, Alexander, Bolzendahl and Jalalzai2018; Hinojosa and Kittilson Reference Hinojosa and Caul Kittilson2020; Stauffer Reference Stauffer2021). If citizens do not know descriptive representatives are in office, how can their presence shape citizens’ evaluations? We advance this conversation by first demonstrating that citizens are generally aware of working-class representation, and second showing that the link between working-class representation and citizens’ evaluations is strongest among a subset of citizens who are most likely to be aware of legislators’ class status.
Beyond bringing research on class representation and symbolic representation into dialogue, one of our core theoretical contributions is to argue that the combination of descriptive and policy representation will have the greatest effect on citizens’ evaluations of institutions, more so than either descriptive or policy representation alone. The literature on citizens’ evaluations of political institutions (and work on symbolic representation more generally) tends to focus primarily on descriptive representation, while ignoring the importance of policy representation. Others consider how policies shape citizens’ evaluations of the state (Campbell Reference Campbell2012). But empirical research rarely considers the combined effect of descriptive and policy representation on citizen attitudes (Schwindt-Bayer and Mishler Reference Schwindt-Bayer and Mishler2005), and when it does, the focus tends to be on descriptive characteristics other than class. This book is the first to consider this relationship with respect to the underrepresentation of the working class. We contend that the behavior of descriptive representatives matters for citizens’ evaluations, not just their presence.
This book also contributes to a growing body of work on the crises of representation and mounting dissatisfaction with democracy (Claassen Reference Claassen2020; Magalhães Reference Magalhães2014; Wike and Fetterolf Reference Wike and Fetterolf2018). Although some scholars express skepticism that disenchantment with democracy has implications for democratic stability (Przeworski Reference Przeworski2019), others empirically demonstrate public support for democracy does help democracy endure (Claassen Reference Claassen2020). While few scholars have linked this growing crisis of representation to the overrepresentation of the upper class in politics, we argue the continued exclusion of the working class is a crucial factor that erodes satisfaction with key agents of representation (such as the legislature and political parties) and contributes to the crisis of representation. Our findings show that more workers in office will help, in part, to increase support for democracy.
Finally, several scholars suggest that rising inequality around the globe has contributed to populist movements on both the political left and right as illiberal, and sometimes authoritarian politicians draw on ordinary citizens’ political dissatisfaction and economic insecurity (Inglehart and Norris Reference Inglehart and Norris2017; Mudde and Kalkwasser Reference Mudde and Rovira Kaltwasser2017; Norris and Inglehart Reference Norris and Inglehart2019). Whereas populism of the left in Latin America promises to redistribute resources from the top of the income distribution, populism of the right in the United States and Europe simultaneously attacks cultural changes and the political “establishment.” Moreover, some work suggests that it is primarily these new populist parties that have incorporated working-class candidates to attract voters, while traditional parties have increasingly nominated career politicians from middle-class backgrounds (Evans and Tilley Reference Evans and Tilley2017; Matthews and Kerevel Reference Matthews and Kerevel2022). Our work suggests incorporating more workers as candidates may help traditional parties draw support away from populist parties and movements, potentially arresting or reversing the crisis of representation that previously led many citizens to either abstain or vote for radical alternatives to mainstream parties.
1.5 Organization of the Book
Chapter 2 develops our theory of working-class inclusion. The chapter is structured around the three central questions that we tackle in this book: (1) Do citizens want to be represented by members of the working class? (2) Will any worker do? That is, how do citizens evaluate workers who do not represent working-class policy interests? (3) How do citizens know workers are in office? In answering these questions, we develop new expectations that we evaluate in the following chapters.
Chapter 3 empirically addresses the question: Do citizens want to be represented by members of the working class? We demonstrate, using novel survey data from Argentina and Mexico and publicly available cross-national data from LAPOP, that citizens do prefer to be represented by legislators from the working class. To do this, we first examine patterns of support for working-class representation using a series of original survey questions in Argentina and Mexico. We asked citizens about their preferences for working-class representation and show that the average citizen in both Argentina and Mexico wants more working-class deputies to occupy seats in congress. Then we introduce data on the class backgrounds of legislators obtained from elite survey data, and present descriptive information about the occupations, gender, and race/ethnicity of working-class deputies. Finally, using cross-national survey data and data on legislators’ class background, we demonstrate that citizens have better evaluations of representative institutions (e.g., more trust in congress and political parties, higher approval of congress, and stronger beliefs that parties listen to and represent people like them) when working-class deputies hold a higher share of seats in the national assembly. We also show that this relationship holds when accounting for a number of potential cofounders, for both men and women, and across racial/ethnic groups. Together, our original survey data combined with data from across Latin America demonstrate support for our argument that citizens do want to be represented by working-class legislators.
Chapters 4 and 5 address our second question: Will any worker do? Our theory argues that for working-class legislators to effectively improve evaluations of representative institutions, they should have strong ties to workers and a track record of advocating for their policy interests. Otherwise, the presence of working-class deputies is expected to trigger backlash. Chapter 4 examines working-class deputies’ propensity to engender positive evaluations of representative institutions conditional on the policy proposals of working-class deputies. We first show, using our original survey data, that the average citizen believes working-class deputies are more likely to understand the problems they face, and promote policies to address them compared to upper-class deputies. Then, consistent with our theory of working-class inclusion presented in Chapter 2, we show that citizens represented by working-class deputies who propose pro-worker policies are much more likely to positively evaluate their representatives and political institutions. To make this case, we use an original survey experiment fielded in Argentina and Mexico wherein we directly manipulate deputies’ class background and whether working-class numeric representation is or is not accompanied by policy representation designed to improve the lives of workers. Overall, we find that both the inclusion of working-class representatives, combined with policy representation of workers, has the largest effect on improving how well citizens feel represented. Numeric representation without policy representation, or policy representation without inclusion, has only minor effects.
Chapter 5 turns to an examination of the ties between working-class representatives and constituents, by taking an in-depth look at the relationship between labor unions, political parties, and workers in Argentina and Mexico. Drawing on prior research, we explain how the evolution of unions and parties throughout history lead to working-class deputies in Argentina having stronger ties to workers and a better track record of policy representation than working-class deputies in Mexico. Then, we leverage an original dataset of working-class representation over time and across states in Argentina and Mexico to show empirically, that whereas increases in working-class representation in Argentina are associated with citizens evaluating their representative institutions more positively, the increased presence of working-class legislators in Mexico leads to backlash and more negative evaluations of legislatures and political parties.
Chapter 6 is the first of two chapters to address our final question: How do citizens know workers are in office? Our theory argues that even though citizens are unlikely to know the exact share of seats workers occupy in office, they are generally aware of working-class representation. Drawing on campaign material, candidate websites, and social media websites, we show that both parties and individual politicians have an incentive to showcase politicians’ class status. Then, we present qualitative evidence from publicly available data, coupled with an inventory of government websites, to show that even absent these political incentives, information on candidates’ class background is publicly available and – at least some of this information – makes it into the hands of citizens, thanks to the popular press. Then, we turn to evidence from two survey experiments from Argentina and Mexico that were designed to evaluate whether citizens can infer information about deputies’ class status from facial images alone. We demonstrate that participants can correctly identify the class background of the national deputies depicted in photographs at a rate significantly better than chance. The results hold regardless of whether the respondent claims to recognize any of the images, for both black-and-white and color images, and even after controlling for the deputy’s skin tone.
Finally, Chapter 7 examines how the relationship between working-class representation and positive evaluations of representative institutions varies among citizens who are more or less likely to be aware of working-class representation. Even though voters can learn about working-class representation through political campaigns, news, and paying attention to politics, we show that levels of political interest and news consumption vary dramatically among citizens within the same country – implying that not all voters are equally likely to be aware of working-class representation. Then, using survey data from across Latin America, we demonstrate that the positive relationship between working-class representation and better evaluations of representative institutions is strongest among citizens with high levels of political interest and those who are avid news followers.
Chapter 8 summarizes our theory and central findings, and synthesizes findings from the cross-national analyses, in-depth case studies, and original survey data. We show how the different pieces of evidence from across the five empirical chapters fit together to explain that people do want to be represented by workers, that citizens are aware of working-class representation, and that both numeric and policy representation produced by working-class representation is necessary to engender trust in institutions. We conclude by discussing the broader implications of our theory for growing dissatisfaction with the way democracy functions across the region and the world, and for the political inclusion of marginalized groups.
