A Multilevel Analysis of Support for Immigrants’ Social Rights in Latin America

Abstract

Western theories and empirical comparative research on attitudes toward immigrants and their rights have largely overlooked Latin America. To address this gap, we conducted multilevel ordered logistic regression analyses on Latinobarómetro surveys from 17 countries (N = 19,004). The findings show that support for immigrants’ social rights is more contingent on immigration-related benefits—especially cultural enrichment—than on perceived threats. When threats do mobilize opposition, the perceived fiscal burden emerges as the sole significant driver, overriding both concerns about labor market competition and fears of rising crime. Furthermore, right-wing individuals were no less supportive of immigrants’ social rights than left-wing individuals. Instead, the most welfare-chauvinist attitudes were found among the politically disengaged. At the macrosocial level, the results provide evidence that contextual factors not only exert a direct statistical effect on public support for immigrants’ social rights but also moderate the influence of perceived micro-level threats. In particular, the national unemployment rate and the immigrant stock exacerbate the exclusionary effect of the perceived fiscal burden on levels of support among citizens. Ultimately, these findings challenge some theoretical assumptions derived from intergroup threat theory, provide novel evidence for the Threat-Benefit Model, and further suggest a distinct political dynamic in the region.

1. Introduction

As the number of international immigrants has increased, migration has become one of the most salient socio-political issues in many countries around the world (Allen et al. 2025; Dražanová 2022; Gu et al. 2022; Schmidt-Catran and Czymara 2023; Talani and Rosina 2025; Van Hauwaert 2023). The debate has centered on the number of immigrants who should be admitted, the rights to which they should be entitled, and particularly the extent to which they should have access to state social benefits (Bell et al. 2023; Burgoon and Rooduijn 2021; Knotz et al. 2025; Lutz 2025; Murard 2022). Accordingly, research on attitudes toward immigrants has attracted growing academic interest in recent years (Bell and Valenta 2025; Davidov et al. 2020; Dražanová 2022; Dražanová et al. 2024), although “much remains contested” (Crepaz 2022, p. 1).

While most scholars agree that a key driver of opposition to immigration and pro-immigrant policies is the perception that immigrants pose an economic or cultural threat (Alesina and Tabellini 2024; Dražanová 2022; Kleider 2022; Valentino et al. 2019), the debate is still open (Dražanová et al. 2024). This is partly because statistical analyses tend to use only one dimension of the threat or a single composite indicator (Tartakovsky and Walsh 2020), which does not allow for the comparative evaluation of their relative importance (Alesina and Tabellini 2024; Hainmueller and Hopkins 2014; Miller 2023). Consequently, scholars overlook that “different types of threat indeed have divergent causes and consequences” (Heizmann and Huth 2021, pp. 57–58; see also Kwon et al. 2024). Crucially, this traditional threat-centric paradigm almost entirely neglects the positive appraisals citizens may hold. The recent literature suggests that understanding public opinion requires moving beyond mere hostility and incorporating perceived benefits—such as cultural enrichment and economic contributions—through comprehensive frameworks, such as the Threat-Benefit Model (Tartakovsky and Walsh 2020; Valenzuela et al. 2024).

According to a recent meta-analysis, “despite a sizeable theoretical and empirical literature, no firm conclusions have been drawn regarding the factors affecting immigration opinion” (Dražanová et al. 2024, p. 317). Disagreement persists over what drives public attitudes, making it difficult to identify which factors are most influential (Dražanová 2022). Furthermore, political ideology has received limited attention as a key explanatory variable (Alesina and Tabellini 2024; Dražanová 2022; Koning and Kaushal 2024; Thomsen and Rafiqi 2019), particularly outside Western Europe (Leykin and Gorodzeisky 2024; Bell and Valenta 2024). Moreover, although it is assumed that attitudes toward immigrants are determined by individual- and country-level factors (Davidov et al. 2020; Dražanová 2022), the specific impact of national contexts—such as economic conditions or immigrant population size—and their interaction with individual perceptions remains significantly underexamined (Boateng et al. 2021; Dražanová et al. 2024; Esses 2021; Ziller and Careja 2022), and the results are inconclusive (Burgoon and Rooduijn 2021; Gugushvili et al. 2021; Laurence and Kim 2023; Ziller and Careja 2022). Notably, few studies have sought to explain citizens’ preferences for the provision of immigrants’ rights (Lutz 2025), especially outside Western Europe (Bell and Valenta 2024; Careja and Harris 2022; Niedzwiecki 2026).

This geographical gap is highly problematic when translating theories from the Global North to the Global South. As most studies on citizens’ attitudes have focused on Western developed societies (Alesina and Tabellini 2024; Bell and Valenta 2025; Dražanová 2022), they leave unclear whether their findings apply to low- and middle-income countries (Dempster et al. 2020; Young et al. 2018). This means that we know very little about the determinants of attitudes toward immigrants and their rights outside the developed world (Dražanová et al. 2024; Meseguer and Kemmerling 2018; Bessen et al. 2025; Ziller and Careja 2022), whether the main findings hold in different social, cultural, and political contexts (Bell and Valenta 2025; Dražanová 2022; Kosic et al. 2023), and how such contexts affect these attitudes (Sipinen et al. 2020; Dražanová 2022; Umansky et al. 2025). The omission of developing countries (Cooray et al. 2018; García-Muñoz and Milgram-Baleix 2021; Meseguer and Kemmerling 2018) is unjustified, given the considerable South–South migration in recent years (Argote and Perelló 2024; Bessen et al. 2025; Meseguer and Kemmerling 2018).

This study addresses these gaps by examining two key questions. First, how do immigration-related threat and benefit perceptions, alongside left–right political ideology and political disengagement, contribute to explaining levels of support for immigrants’ social rights in Latin America? Second, to what extent do contextual factors—such as a country’s economic conditions and the size of its immigrant population—moderate the association between threat perceptions and support for immigrants’ social rights? To answer these questions, multilevel ordered logistic regression models were applied to the Latinobarómetro survey data from 17 countries. We chose Latin America because it is an understudied region that provides a crucial testing ground for dominant theoretical frameworks on attitudes toward immigrants and their rights.

In recent years, the region has experienced the highest growth rate of international migrants globally (McAuliffe and Triandafyllidou 2021). Between 2010 and 2020, the foreign-born population practically doubled across most Latin American countries (Harris et al. 2023). This profound transformation has been characterized by a strong South–South mobility. According to the World Migration Report 2024, in 2020, there were around 11 million intra-regional migrants originating from LAC residing within the same region, whereas the number of migrants from other regions of the world living in LAC had remained relatively stable at around 3 million (McAuliffe and Oucho 2024). Notably, of the more than 7 million Venezuelans who have left their country due to humanitarian crises since 2015, approximately 85% have settled within another LAC country, representing the largest displacement crisis in the history of the Americas (Harris et al. 2023). These intra-regional migratory flows have been associated with a significant decline in public acceptance of immigration in almost every country (Esipova et al. 2020). However, with few exceptions (Bessen et al. 2025; Marroquin and Saravia 2022; Meseguer and Kemmerling 2018), we still know very little about the individual and contextual factors that explain support for immigrants’ social rights in Latin America. By examining these dynamics, we test whether Western-centric theories hold in a region characterized by high levels of migration and shared cultural proximity but also economic precarity, limited welfare provisions, and widespread political disaffection and weak party institutionalization.

2. Theoretical Background

2.1. Intergroup Threat Theory (ITT)

Models based on intergroup threat theory are among the most prevalent approaches to explaining negative attitudes toward immigration and immigrants (Bell and Valenta 2025; Blinder and Lundgren 2019; Kosic et al. 2023). The theory posits that the stronger the perception of immigrants as a threat, the more negative the attitudes toward them (Benoit 2021; Ford and Mellon 2020; Kustov 2019). Those who perceive immigrants as a threat—whether cultural, economic, or security-related—are more likely to oppose social policies that favor them (Gootjes et al. 2021). This has given visibility to the concept of welfare chauvinism, understood as the belief or sentiment that welfare benefits should be restricted to a country’s own citizens (Allen et al. 2025; Bell and Valenta 2024; Careja and Harris 2022). This concept was introduced to explain the success of populist right parties in Western Europe in the 1990s and is often linked to viewing immigrants as a threat in a zero-sum game over welfare benefits (Bell and Valenta 2024; Careja and Harris 2022).

According to ITT, perceived economic threat is driven primarily by competition for scarce resources (Ford and Mellon 2020; Kwon et al. 2024; Meuleman et al. 2018; Vogt Isaksen 2019). This framework encompasses two primary arguments from the literature. The first is the labor market competition threat (LMC), which holds that immigrants threaten job security and wages (Dražanová and Gonnot 2023; Valentino et al. 2019). The differences in skills between citizens and immigrants are the key aspect. Immigrants with a skill set like that of citizens are a major concern, but low-skilled immigrants may also be willing to take low-paid jobs (Kocijan and Kukec 2022). In this way, perceived economic competition at the group level translates into prejudice toward immigrants (Esses 2021; Gu et al. 2022). In contrast, higher-skilled citizens see lower-skilled immigrants as complementary and tend to favor their arrival. Thus, individuals’ attitudes toward immigrants will vary depending on whether they fear displacement in the job market, which is linked to both citizens’ and immigrants’ qualifications (Meseguer and Kemmerling 2018). More broadly, this differentiation operates along socioeconomic lines: citizens who “objectively find themselves in precarious economic positions (e.g., welfare-dependent or unemployed)” are more likely to perceive immigrants “as competitors for the resources they themselves need, such as jobs or benefits” (Ziller and Careja 2022, p. 177), a logic that operates not only across the labor market dimension but also across the fiscal burden dimension discussed next.

The second argument is the fiscal burden threat (FB). Citizens, particularly native-born taxpayers, may perceive immigrants as a fiscal burden, fearing erosion of their benefits or higher taxes (Allen et al. 2025; Dražanová et al. 2024; Lutz 2025). The concern is that immigrants could strain the welfare system due to their vulnerability and dependence on social programs (Alesina et al. 2023; Alesina and Tabellini 2024; Burgoon and Rooduijn 2021; Miller 2023; Koning 2022; Valentino et al. 2019), which leads to opposition to social policies that grant immigrants equal access to welfare provisions (Brady and Finnigan 2014; Gugushvili et al. 2021; Mau and Burkhardt 2009). These concerns often manifest as welfare-chauvinist attitudes, reflecting “that many citizens prefer immigrants to be excluded from welfare benefits” (Lutz 2025, p. 132). According to this view, immigrants are less deserving of social solidarity, and citizens, as primary contributors, should have priority in accessing social protection and scarce public services (Bell et al. 2023; Breznau and Eger 2016; Lutz 2025; Norris and Inglehart 2019). However, it has also been pointed out that “much of the perceived fiscal burden stems from lower contributions to welfare systems, rather than from disproportionately higher benefit usage” (Allen et al. 2025, p. 72). In the Latin American context, where labor informality is widespread, immigrants’ integration into unregulated economic sectors structurally limits their tax contributions.

A third critical dimension is the security threat (ST). A common threat perception is also that immigration leads to higher crime rates and potentially terrorist activities (Abbondanza 2025, p. 19; Baranauskas and Stowell 2025; Boateng et al. 2021). The rhetoric of populist and anti-immigration parties has centered around crime and stereotypes of immigrants as law-breaking (Alesina and Tabellini 2024). This criminalization of immigrants has fueled what Stumpf (2006) has termed “crimmigration”—the progressive merger of criminal law and immigration enforcement—a discourse that actively manufactures public hostility by equating mobility with illegality. Hence, states seek to defend their borders against irregular migration, driven by this (discursive) nexus between immigration and crime (Abbondanza 2025). Ultimately, this crimmigration trend represents a broader transition toward the institutional convergence of immigration control and criminal justice (Franko 2026; Stumpf 2006).

In this study, we re-examine such arguments in the Latin American context, a critical step for testing their generalizability beyond Western societies. Based on ITT, we hypothesize the following:

H1a. 

Perceived labor market competition (LMC) is associated with weaker support for immigrants’ social rights.

H1b. 

Perceived fiscal burden (FB) is associated with weaker support for immigrants’ social rights.

H1c. 

Perceived security threat (ST) is associated with weaker support for immigrants’ social rights.

2.2. Extending ITT: The Threat-Benefit Model (TBM)

While ITT is the most prevalent approach to explaining attitudes toward immigrants, it has some drawbacks. The main weakness is related to “its focus on the exclusively negative aspects of the perception of immigrants” (Walsh and Tartakovsky 2021, p. 2). This emphasis on threats almost totally overlooks the fact that many people also have positive attitudes toward immigrants (Tartakovsky and Walsh 2016, 2020; Walsh and Tartakovsky 2020, 2021). To address these theoretical limitations, the Threat-Benefit Model (TBM) is included to reinforce the overall theoretical framework.

As an extension of ITT, the TBM posits that immigrants “evoke not only perceptions of threat but also of benefits” (Valenzuela et al. 2024, p. 1502). Incorporating this dual approach is critical, given recent empirical findings demonstrating that perceived benefits can be even more pivotal than perceived threats in shaping citizens’ attitudes (Valenzuela et al. 2024). This two-factor perspective represents an advance on previous theoretical understandings by examining both the threats immigrants may pose and the benefits they may bring to a society (Tartakovsky and Walsh 2016; Walsh and Tartakovsky 2020, 2021). The conceptualization of threats and benefits as distinct dimensions allows us to consider whether immigrants may be perceived as simultaneously threatening in one dimension and beneficial in another (Walsh and Tartakovsky 2021, p. 2; Tartakovsky and Walsh 2020, p. 3956).

Realistic (Economic) and Symbolic (Cultural) Benefits

Beyond the competition for limited resources emphasized by ITT, as discussed above, the TBM also suggests that immigrants can provide economic benefits (EBs). They reflect “the immigrants’ potential to contribute to the economic development of the receiving country,” related to their “readiness to take jobs that local people do not want or lack the skills to do,” as well as bringing “valuable skills, language knowledge, and international connections” (Tartakovsky and Walsh 2020, p. 3959; Walsh and Tartakovsky 2020, p. 77; 2021, p. 3).

While the ITT has also highlighted perceived cultural threat, from the TBM perspective, the symbolic dimension encompasses cultural benefits (CBs): “cultural diversity benefits are related to the new cultural elements (food, clothes, music, etc.) that immigrants bring with them, which may be perceived by some local people as culturally enriching the receiving society” (Tartakovsky and Walsh 2016, p. 77; 2020, p. 3959; Walsh and Tartakovsky 2020, p. 75; 2021, p. 3). Because most intra-regional migrants and host populations covered by the Latinobarómetro surveys share a cultural proximity, these cultural benefits may be more readily recognized by the native population, acting as a potential buffer against hostility (Valenzuela et al. 2024).

The TBM holds that perceiving immigrants as threatening or beneficial to the receiving society shapes opinions on immigration (Tartakovsky and Walsh 2016). While threat perception often leads to support for restrictive policies, as citizens perceive immigrants more positively (i.e., consider them beneficial for receiving society), they “will tend to support immigration policy directed at defending the immigrants’ rights” (Tartakovsky and Walsh 2016, p. 76), including “the wish to grant extensive rights” (Walsh and Tartakovsky 2020, p. 92). Based on the TBM framework, we hypothesize the following:

H2. 

The perception that immigrants benefit society (EB/CB) is associated with stronger support for immigrants’ social rights.

So far, we have argued that attitudes toward immigrants are shaped by the extent to which citizens view immigrants as threats or benefits. Within traditional threat frameworks (ITT), Latin America’s structural conditions, characterized by economic precarity and limited welfare provisions, could make perceptions of threat (LMC/FB) linked to immigration more salient to citizens, leading to the expectation that perceived threats would exert a stronger influence than perceived benefits.

H3a. 

The perception of immigrants’ realistic economic threats (LMC/FB) shows a stronger association with support for immigrants’ social rights than the perception of their positive contributions (benefits: EB/CB).

According to the TBM, however, perceptions of immigrants’ contributions to society are pivotal (Mepham and Verkuyten 2017; Tartakovsky and Walsh 2020), predicting citizens’ acculturation preferences better than perceived threats do (Valenzuela et al. 2024). Moreover, in contrast to most Western studies, which see immigrants’ cultural differences as a problem for citizens’ acceptance (Dražanová et al. 2024; Schmidt-Catran and Czymara 2023), the dynamics of intra-regional migration and the comparative cultural proximity between sending and receiving populations in Latin America (especially in countries included in the Latinobarómetro surveys) may imply a very different situation, one of greater acceptance. A form of regional solidarity with migrants may emerge among Latin American countries (see, for example, FitzGerald and Cook-Martín 2014, p. 76), even in a context of resource competition, fostered by cultural proximity based on shared language (Spanish, except in Brazil), religiosity (mostly Catholic), and history (common Iberian colonial heritage, parallel post-independence trajectories of state-building, and legal systems derived from civil-law traditions; see Acosta 2018). In alignment with the TBM framework, the expectation is as follows:

H3b. 

The perception of immigrants’ positive contributions (benefits: EB/CB) shows a stronger association with support for immigrants’ social rights than the perception of their realistic economic threats (LMC/FB).

2.3. Political Ideology

In recent years, political ideology has emerged as a key variable in understanding attitudes toward immigrants in Western countries (Alesina and Tabellini 2024; Leykin and Gorodzeisky 2024). People’s views are shaped by their broader ideological orientations (Koning and Kaushal 2024), which encompass core beliefs and values (Thomsen and Rafiqi 2019). Typically, “Right-wing political ideologies are associated with a lack of tolerance and openness towards cultural diversity, […] and acceptance of social inequality” (Leykin and Gorodzeisky 2024, p. 6). The empirical evidence shows that “rejection of immigrants and minority rights is stronger among political conservatives, or the right, than among political liberals, or the left” (Verkuyten et al. 2022, p. 1; see also Thomsen and Rafiqi 2019, 2020).

From the perspective of right-wing ideology, the increased societal diversity resulting from immigration is seen as leading to cultural disintegration, less compliance with social norms, and excessive social spending (Thomsen and Rafiqi 2020). It has also “promoted the idea of ‘national preference’, that is, giving ‘natives’ priority when it comes to jobs, housing, health care, and so on” (Rydgren 2018, p. 27). Opposition to immigrants’ social rights is a key ideological feature of right-wing populism (Knotz et al. 2025). Accordingly, right-wing individuals are more likely to “object to equal rights for immigrants” (Verkuyten et al. 2022, p. 1). The political left, conversely, is commonly associated with universalism, diversity, and cultural openness, which foster more tolerant and inclusive attitudes toward “others” (Leykin and Gorodzeisky 2024). As such, “broadly speaking, the left has emerged as a defender of immigration and multiculturalism, whereas the right is critical of both” (Dancygier and Margalit 2020, p. 735).

More importantly, by exacerbating the perception of threat, right-wing ideology mobilizes negative attitudes toward immigrants (Davidov et al. 2020; Eger and Bohman 2016; Norris 2005; Vargas-Maia and Pinheiro-Machado 2023). The most conservative segments of the political right feel threatened by the loss of society’s traditional values, which they associate with the arrival of immigrants with different cultural backgrounds (Leykin and Gorodzeisky 2024; Norris and Inglehart 2019). According to radical-right ideology, various dangers threaten national identity. Among them, however, immigration is considered the most important (Rydgren 2018). Because anti-immigration views are a core component of the political right’s ideology, a right-wing orientation has become an important predictor of anti-immigrant attitudes (Schmidt-Catran and Czymara 2023). Accordingly, from this perspective, we hypothesize the following:

H4a. 

Right-wing individuals express weaker support for immigrants’ social rights than left-wing individuals.

However, the association between left–right political orientation and attitudes toward immigrants is under scrutiny. Leykin and Gorodzeisky (2024) challenge the prevalent assumption in political and social-scientific discourses that there is a link between right-wing political orientation and anti-immigrant sentiment. The evidence showed that what applies to Western Europe does not apply to Eastern Europe. Furthermore, Bell and Valenta (2024, p. 1289) find evidence that right-wing individuals are not more welfare chauvinistic, indicating a “broad agreement across the political spectrum in these societies on the exclusion or minimisation of immigrants’ right to social benefits”. In the Latin American case, it has been pointed out that left-wing politicians have been as likely as right-wing leaders to increase barriers to immigrants’ access to means-tested policies. In the case of social pensions, for example, “political elites (both from the left and the right) want to restrict their access to immigrants as a way to cut costs” (Niedzwiecki 2026, pp. 236, 238).

Furthermore, traditional left-wing political ideology’s explanatory power may be more limited in Latin America, a region characterized by partisan dealignment and widespread political disaffection (Mainwaring 2018; Sánchez-Sibony 2024). Theoretically, these disaffected individuals are particularly susceptible to exclusionary and anti-establishment narratives that scapegoat marginalized groups, including immigrants, for systemic governance failures (Norris and Inglehart 2019). Consequently, and in contrast to the idea that attitudes toward immigrants are expected to differ along the traditional political spectrum (Heizmann and Huth 2021), we hypothesize the following:

H4b. 

Politically disaffected individuals (those who do not identify with any political ideology) express significantly weaker support for immigrants’ social rights compared to those individuals with defined ideological alignments.

2.4. Contextual Factors and Cross-Level Interactions

A core tenet of ITT is that people’s attitudes toward immigrants are affected not only by individual characteristics but also by the context in which they live (Dražanová 2022). Although “the number of studies researching contextual effects on attitudes toward immigration is quite limited” (Dražanová 2022, p. 94; Dražanová et al. 2024, p. 336), economic conditions and the relative size of the immigrant population are considered the two main contextual predictors of threat perception (Kosic et al. 2023) and animosity toward out-groups (Heizmann and Huth 2021).

On the one hand, it is argued that economic conditions are key to predicting anti-immigrant attitudes (Peresman et al. 2023). The argument is that higher levels of unemployment or decreasing levels of GDP per capita will lead citizens to view immigrants as an economic threat, as they are seen as competitors for scarce resources such as jobs and welfare benefits (Bartasevičius and Trunov 2024; Bell and Valenta 2025; Dražanová 2022; Laaker 2024; Van Hauwaert 2023), especially among society’s most vulnerable socioeconomic groups (Chang Kang and Look 2020; Dražanová 2022; Jaime-Castillo et al. 2016; Ruist 2016). From this point of view, “the impact of immigration matters most under conditions of high unemployment, with an interaction effect where people blame ‘foreigners’ for job insecurity, low wages, or loss of employment” (Norris 2005, p. 171). This, in turn, increases the desire to exclude immigrants from state welfare benefits (Bell et al. 2023, p. 304).

Additionally, some empirical studies have demonstrated that “in countries with a higher foreign population stock, the perception of immigrants as an economic threat tends to be more pronounced” (Heizmann and Huth 2021, p. 66). As the number of immigrants rises, perceptions of threat increase, in turn generating greater aversion to immigrants (Billiet et al. 2014; Eger and Bohman 2016; Kosic et al. 2023; Turner and Cross 2015). Although it is important to distinguish between the absolute size of the immigrant population (stock) and the pace of its arrival (growth rate) (Bessen et al. 2025; Schmidt-Catran and Czymara 2023), ultimately “what matters are not levels of immigration per se, but rather the belief that any influx of new minorities could take away public benefits such as housing, depress wages in low-skilled jobs, or exacerbate unemployment rates” (Norris 2005, p. 177).

From this perspective, the most pressing concern of citizens is their own protection and the well-being of their group. If they feel threatened, they will tend to support not just restrictive immigration control policies but also question the social rights of minorities (Blinder and Lundgren 2019). Therefore, we expect that macro-level factors have both a main effect and a moderating role in explaining citizens’ welfare-chauvinist attitudes.

H5a (Main effects at the country level).

The more unfavorable macroeconomic conditions are (as measured by GDP per capita growth rate and unemployment) and the larger the immigrant population is (as measured by immigration stock and immigrant population growth rate), the lower the level of support for immigrants’ social rights across countries.

H5b (Cross-level interaction effects).

The association between threat perceptions (LMC/FB/ST) and support for immigrants’ social rights is moderated by countries’ macroeconomic conditions and the size of their immigrant populations. Specifically, under unfavorable macroeconomic conditions or larger immigrant populations, the negative association between perceived threats and support for social rights will be more pronounced.

3. Data and Methods

3.1. Data and Sample

We used a quantitative, cross-sectional design with independent national samples from the 2020 Latinobarómetro survey. The survey was administered face-to-face in 17 countries between 26 October and 15 December 2020, and online in Argentina between 28 April and 16 May 2021, with a total of 20,204 anonymized interviews across the region. All national samples covered the adult population in both urban and rural areas. There were very few exceptions due to geographic inaccessibility, security conditions, or sparse density, for example the departments of Gracias a Dios and Islas de la Bahía in Honduras (approximately 1% of the adult population), the Emberá and Kuna Yala comarcas in Panama (approximately 2%), and Galápagos and undelimited zones in Ecuador (approximately 0.4%).

For the sampling design, based on national census data, a multistage probabilistic approach (three or four stages, depending on the country) was used that included the random selection of households. In the final stage, the selection of the ultimate respondent within the household was conducted using quotas for sex, age, and educational level. Sample sizes ranged from 1000 to 1200 respondents per country, yielding estimated margins of error of ±2.8% to ±3.1% at the 95% confidence level for univariate statistics at the national level (see further specifications at https://www.latinobarometro.org/latinobarometro-2020) (accessed on 18 May 2026).

The countries included in the 2020 Latinobarómetro survey were Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, Ecuador, El Salvador, Guatemala, Honduras, Mexico, Nicaragua, Panama, Paraguay, Peru, Dominican Republic, Uruguay, and Venezuela. However, Venezuela was excluded from our analyses as an outlier due to its ongoing migratory crisis and the lack of comparable macroeconomic data. The final sample comprises 19,004 cases. The database is publicly available and can be downloaded from https://www.latinobarometro.org/latContents.jsp (accessed on 3 June 2024).

3.2. Measurement

The dependent variable, support for immigrants’ social rights, was measured using a 4-point ordinal scale from the following survey question: “Tell me if you strongly agree (1), agree (2), disagree (3), or strongly disagree (4) with the following statement… Immigrants should have the same access to health, education, and housing as citizens of [COUNTRY].” The original response categories were reverse-scored, meaning that higher values indicate greater support. The 2020 wave of the Latinobarómetro was the only year this question was included to date.

To measure perceived threats, we used three measures: labor market competition (“Immigrants come to take our jobs away”), fiscal burden (an index constructed by averaging two items: “Immigrants are a burden on the state” and “Immigrants give more than they receive”, with the latter reverse-scored), and security threat (“Immigrants increase crime rates”). All final measures were reverse-scored on a 4-point scale, with higher values indicating a greater perception of threat. Perceived benefits were measured by economic contributions (“Immigrants are good for the country’s economy”) and cultural contributions (“Immigrants contribute new ideas to our culture and society”). Both were reverse-scored on a 4-point scale, with higher values indicating greater perceived benefits.

Political ideology was measured by the following question: “On a scale where ‘0’ is ‘left’ and ‘10’ is ‘right’, where would you place yourself?” Values were recoded into a four-category variable: left (0–4), center (5), right (6–10), and none (97). The “left” category was used as the reference category. Only genuine nonresponses, “Don’t know” or “No answer” (DK/NA), were treated as missing data and imputed using Multiple Imputation by Chained Equations (MICE). Crucially, these nonresponses are conceptually distinct from the “none” category, which captures explicit political disaffection. By treating political ideology as a 4-level categorical variable during imputation, the MICE algorithm was able to probabilistically handle true missingness (Missing at Random [MAR] assumption) while preserving the structural variance of the disaffected group. Then, the categories were introduced into the statistical models as dummy variables, with the left as the reference category.

The following Level 1 variables were included in the statistical analysis as control variables: sex (1 = female), age (1 = 16–25 to 4 = 61 or more), education (1 = illiterate to 7 = complete university education), socioeconomic status (1 = low [very bad/bad] to 3 = high [very good/good]), and perception of the country’s economic situation (1 = very bad to 5 = very good). For socioeconomic status, we used the Latinobarómetro’s interviewer-rated classification, which draws on on-site observations of the dwelling (e.g., quality of the facade and building materials), the furniture, and the interviewee’s overall appearance. The original five-category Latinobarómetro variable was collapsed into three levels (low, medium, and high) to ensure adequate cell sizes across countries and to improve the reliability of the overall classification by reducing the difficulty interviewers face when distinguishing between adjacent, similar groups.

We opted for this observational indicator over self-reported household income for three main methodological reasons. First, income questions in face-to-face surveys systematically suffer from high item nonresponse rates, and evidence shows this missingness is not random but rather driven by question sensitivity and task complexity (Jabkowski and Piekut 2024; Moore et al. 2000). This nonresponse tends to concentrate at the extremes of the income distribution (Bollinger et al. 2019; Hlasny 2020). Relying on the interviewer’s observation circumvents this issue, granting practically a full classification coverage. Second, self-reported income is highly susceptible to systematic underreporting bias, which can severely distort socioeconomic rankings. This underreporting is particularly acute among independent or informal workers (Kukk and Staehr 2017; Kukk et al. 2020) and high-income households, whose capital incomes are systematically underreported in Latin American household surveys (Alvaredo et al. 2025; Larrañaga et al. 2022; Lustig and Vigorito 2025). Finally, indicators based on housing conditions and durable assets provide a more stable and reliable proxy for long-term socioeconomic status than the high volatility and measurement error associated with self-reported current income (Carranza et al. 2023; Poirier et al. 2020). This approach certainly introduces interviewer bias that can affect the accuracy of estimates during classification. However, addressing income nonresponse and underreporting in household and public opinion surveys is a complex task, especially in contexts of high market informality, which is beyond the focus and scope of this research.

Regarding Level 2 contextual economic variables, we included the annual GDP per capita growth rate (2019) and the unemployment rate (2019) from the United Nations Economic Commission for Latin America and the Caribbean (ECLAC). We used macroeconomic data from the year prior to the Latinobarómetro survey to allow for a time lag in their effects on individual attitudes (Bell et al. 2023, p. 306) and, fundamentally, to capture the true structural baseline of the economies prior to the anomalous macroeconomic shock triggered by the COVID-19 pandemic in 2020. The demographic impact of immigration was measured using the immigration stock as a percentage of the total population (2020) and the immigration growth rate (2015–2020), utilizing United Nations data.

3.3. Statistical Data Analysis Techniques

We estimated multilevel ordered logistic regression models, given that individuals are nested within countries and the dependent variable is ordinal (Rabe-Hesketh and Skrondal 2022), using Stata version 19. Missing data—which occurred exclusively at Level 1—were handled using Multiple Imputation by Chained Equations (i.e., MICE) with Fully Conditional Specification (FCS) (He et al. 2022). Imputation was stratified and executed independently within each country to accommodate random slopes and allow the covariance structure to vary freely across macro-contexts (Grund et al. 2018). This approach avoids intractable dimensionality and overparameterized matrices that are computationally unstable and fail to converge in multilevel imputation algorithms, particularly in models with few clusters (Grund et al. 2026; Rabe-Hesketh and Skrondal 2022).

To facilitate interpretation and avoid multicollinearity, especially in cross-level interactions, predictors were then standardized. Crucially, to rigorously isolate within-country compositional mechanisms from unobserved macro-level confounders, we applied Within-Cluster Centering (CWC) (Bell et al. 2018; Rabe-Hesketh and Skrondal 2022). With the exception of nominal variables, Level 1 predictors were within-cluster-centered (i.e., group-mean-centered) to capture a strictly relative positional effect, purge compositional bias, and alleviate collinearity (Finch et al. 2024). Subsequently, to preserve true cross-national differences in variance, these centered variables were scaled by their global standard deviation, while Level 2 predictors were grand-mean-standardized. Additionally, to treat macro-level units as exchangeable and prevent more populous countries from artificially dominating the macro-level variance, the Level 2 grand mean and standard deviation were calculated by assigning strictly equal weight to each of the 17 national units (Rabe-Hesketh and Skrondal 2022).

The analytical strategy utilized a dual-architecture modeling process to ensure computational stability and rigorous inference. First, the fixed-effect coefficients for the different models—encompassing Level 1, Level 2, and cross-level interactions—as well as the exact statistical significance (p-values) for the micro main effects (Level 1) were obtained through a multilevel ordered logistic regression with random intercepts and Cluster-Robust Standard Errors (CR-SE), pooling the results across the 20 imputed datasets using Rubin’s rules (Rubin 1987). Second, although fixed-effect coefficients are unbiased for the macro-level parameter even with very few Level 2 clusters (e.g., N < 10), when the Level 1 sample size is sufficiently large (Elff et al. 2021), standard errors are systematically underestimated in these small macro-samples (McNeish 2017). This underestimation renders asymptotic p-values unreliable, thereby inflating the Type I Error rate (Hox et al. 2018; Zhang and Lai 2024). To address this small-sample bias in our study, the p-values for all Level 2 predictors and the cross-level interactions were recalculated using a strict small-sample correction based on Student’s t-distribution, with 15 degrees of freedom—often referred to as the m-l-1 rule (Elff et al. 2021). In each cross-level model, a single perceived threat was introduced, interacting simultaneously with only one national context variable.

Variance explained is reported using marginal and conditional R² for mixed models (Nakagawa et al. 2017), which were averaged across the 20 imputed datasets (He et al. 2022). Finally, because interaction coefficients in nonlinear ordinal models assume proportional odds and do not intuitively reflect the absolute magnitude or direction of the effects, interpretations were conducted via simple slopes analysis and the visualization of cumulative predicted marginal probabilities. By projecting the cumulative probability of crossing the support threshold (categories 3 and 4 combined), we restored the monotonicity of the inverse logistic function. This geometric transformation ensured that cross-level moderation was accurately evaluated, free from the distortion caused by unimodal intermediate categories (Mize 2019; Rabe-Hesketh and Skrondal 2022).

4. Results

Latin America has experienced remarkable intra-regional migration dynamics since 2010, mainly from Venezuela, Haiti, and Central America (Bessen et al. 2025; Meseguer and Kemmerling 2018; McAuliffe and Triandafyllidou 2021). These migratory flows were associated with a significant decline in public acceptance of immigration across almost every Latin American country (Esipova et al. 2020). Alongside this process, several countries faced new challenges in welfare provision, leading to highly contested access to social policy for immigrants (Niedzwiecki 2026). As the Latinobarómetro surveys showed in 2020 (illustrated in Figure 1), while support for immigrants’ social rights averaged 68% across the region, substantive disparities emerged between countries. Notably, support levels in Peru (46%), Argentina (47%), Bolivia (58%), and Chile (58%) were among the lowest, underscoring how demographic changes reshape public solidarity.

As differences across countries may be linked to various individual and contextual factors, we next conducted a multilevel ordered logistic regression to analyze the influence of individual and contextual variables and their cross-level interactions.

Table 1. Multilevel ordinal regression models of support for immigrants’ social rights.

	Model 1	Model 2	Model 3	Model 4
Individual-level variables
Sociodemographic characteristics
Sex [1 = female]	−0.015	−0.015	−0.015	−0.015
Age groups	−0.049 *	−0.049 *	−0.049 *	−0.049 *
Education	−0.016	−0.016	−0.016	−0.016
Socioeconomic status	0.000	0.000	0.000	0.000
Country’s economic situation
Perception of country’s economic situation	0.050	0.050	0.050	0.050
Perceived threats
Fiscal burden	−0.356 ***	−0.356 ***	−0.356 ***	−0.356 ***
Labor market competition	0.025	0.025	0.025	0.025
Security threat	−0.041	−0.041	−0.041	−0.041
Perceived benefits
Economic benefits	0.382 ***	0.382 ***	0.382 ***	0.382 ***
Cultural benefits	0.470 ***	0.470 ***	0.470 ***	0.470 ***
Political ideology (ref. category = Left)
Center	−0.050	−0.050	−0.050	−0.050
Right	−0.047	−0.047	−0.047	−0.047
None	−0.204 *	−0.205 *	−0.204 *	−0.205 *
Country-level variables
Annual GDP per capita growth rate	0.001
Unemployment rate		0.455 ***
Immigration stock			−0.379 ***
Immigrant population growth rate				−0.289 ***
Random effects variance (country-level)	0.703	0.577	0.575	0.633
Marginal R2 (fixed)	0.186	0.222	0.212	0.202
Conditional R2 (total)	0.330	0.338	0.329	0.331
Individual-level N	19,004	19,004	19,004	19,004
Country-level N	17	17	17	17

Note: Except for nominal variables, Level 1 predictors were within-cluster-centered (CWC) and subsequently scaled by the global standard deviation of the centered variable, while Level 2 predictors were grand-mean-standardized. Each coefficient represents the expected change in the log-odds of the dependent variable for a one-standard-deviation increase in the respective predictor. Results are pooled across 20 multiply imputed datasets using Rubin’s rules. Significance tests are adjusted using the m-l-1 small-sample correction (df = 15). * p < 0.05, *** p < 0.001.

presents the results of four multilevel ordered logistic regression models that explain support for immigrants’ social rights in Latin America. The inclusion of marginal and conditional R² metrics indicates robust model fit, accounting for approximately 33% of the total variance when both fixed and random effects are included. A primary observation is that individual-level predictors (Level 1) remain stable across the various models, independent of the countries’ macroeconomic or demographic contexts (Level 2).

Regarding Level 1, support for social rights is largely associated with perceived threats and benefits attributed to immigrants, rather than with sociodemographic characteristics (except for age) across all models. Furthermore, among the threats, the fiscal burden displays the strongest and sole significant negative association with support. Contrary to initial expectations derived from intergroup threat theory (which has been predominantly tested in the Global North), viewing immigrants as labor market competitors does not significantly correlate with lower support. Given the high levels of labor informality in the region, immigrants are often absorbed into unregulated economic sectors without directly displacing native formal workers. Similarly, security threat perception shows no statistically significant effect.

This indicates that Latin American welfare chauvinism is primarily rooted in the perception of immigrants as a fiscal burden, rather than fear of job loss or security concerns. As discussed in the theoretical section, when citizens perceive immigrants as a fiscal burden, they fear the erosion of their benefits or higher taxes, leading to opposition to social policies that grant immigrants equal access to welfare provisions. According to this view, citizens should have priority in accessing social protection and scarce public services. Therefore, Hypothesis H1a (fiscal burden) is corroborated, while H1b (labor market competition) and H1c (security threat) lack empirical backing as standalone predictors.

Conversely, the belief that immigrants contribute to the economy and, particularly, to culture predicts a substantially higher likelihood of support for their social rights. These results corroborate Hypothesis H2. It is noteworthy that perceived benefits, especially cultural benefits, display a stronger statistical association than threat perceptions (as evidenced by the larger absolute magnitudes of their coefficients). Thus, Hypothesis H3b is also substantiated, while hypothesis H3a is rejected. This finding corroborates the core premise of the TBM, empirically supporting the claim that perceived benefits rather than threats tend to better predict citizens’ attitudes toward immigration (Valenzuela et al. 2024). Yet, general perceptions of the national economy do not correlate with support, suggesting that concerns are fundamentally about strain on state resources, such as the allocation of taxpayer revenues or the imposition of higher taxes, rather than broader macroeconomic grievances.

From a political–ideological standpoint, the empirical evidence shows that, unlike established patterns in advanced democracies, right-wing individuals do not express weaker support for immigrants’ social rights than left-wing individuals, thereby failing to corroborate Hypothesis H4a. Importantly, when using the political left as the reference, the strongest and only significant exclusionary attitudes are found among those who do not identify with any political position, substantiating Hypothesis H4b. This suggests that, while traditional left–right ideologies do not strictly differentiate between levels of support, political disaffection is a stronger predictor of anti-immigrant sentiment in the region. In Latin America’s weakly institutionalized party systems, when citizens lack programmatic partisan anchors (Mainwaring 2018; Sánchez-Sibony 2024), they could become more susceptible to anti-establishment and xenophobic narratives (Norris and Inglehart 2019), leaving the politically disengaged as the most reactive and exclusionary group. Moreover, as discussed in the theoretical framework, the association between left and right political orientation and attitudes toward immigrants is under scrutiny. The evidence shows that what applies to Western Europe does not apply to Eastern Europe (Leykin and Gorodzeisky 2024; Bell and Valenta 2024). In the Latin American context, it has also been noted that left-wing politicians have been as likely as right-wing leaders to raise barriers to immigrants’ access to means-tested policies, such as social pensions (Niedzwiecki 2026).

At Level 2, the contextual variables show mixed results. Contrary to our expectations anchored in traditional macroeconomic threat models, annual GDP per capita growth does not significantly predict support (Model 1). Notably, higher national unemployment rates have a statistically significant positive coefficient (Model 2), indicating that individuals in high-unemployment contexts exhibit a higher baseline likelihood of supporting social rights. This seemingly counterintuitive finding may reflect a broader societal demand for robust social safety nets in economically depressed contexts. Nevertheless, as subsequent cross-level interactions will demonstrate, this inclusive disposition is highly conditional and collapses sharply into exclusionary attitudes when immigrants are explicitly framed as a fiscal burden. Conversely, in countries with a higher immigration stock (Model 3) or a higher immigrant population growth rate (Model 4), individuals are less likely to support immigrants’ social rights in Latin America. These findings partially align with hypothesis H5a.

Next, we investigated the extent to which macrostructural characteristics condition the relationship between threat perceptions and support for immigrants’ social rights in the region. As summarized in

Table 2. Summary of cross-level interaction effects.

	Economic Context		Demographic Context
	GDP Growth per Capita	Unemployment Rate	Immigrant Stock	Immigrant Growth Rate
Perceived threats
Fiscal burden	0.106	−0.123 *	−0.079 *	−0.000
Labor market competition	0.111	−0.071	−0.049	0.039
Security threat	0.089	−0.090	−0.070 *	0.017

Note: All interaction terms were estimated in separate multilevel ordered logistic regression models with robust standard errors. Models included the full set of individual-level variables from Table 1. Significance tests are adjusted using the m-l-1 small-sample correction (df = 15). * p < 0.05.

, each model included a single perceived threat, interacting with only one national context variable.

While twelve models were estimated, robust cross-level interactions emerged in only three configurations, providing targeted evidence to test our hypotheses.

Table 3. Marginal effect of fiscal threat on support for immigrants’ social rights, moderated by macro-level contextual variables.

Structural Context	Marginal Effect (b)	Robust SE	t(15)	Exact p-Value
Unemployment Rate
Low (−1 SD)	−0.217	(0.045)	−4.768	<0.001 ***
Mean (0)	−0.337	(0.043)	−7.820	<0.001 ***
High (+1 SD)	−0.458	(0.095)	−4.837	<0.001 ***
Immigrant Stock
Low (−1 SD)	−0.275	(0.050)	−5.534	<0.001 ***
Mean (0)	−0.353	(0.049)	−7.194	<0.001 ***
High (+1 SD)	−0.430	(0.072)	−5.951	<0.001 ***

Note: Results are pooled across 20 multiply imputed datasets using Rubin’s rules. Significance tests are adjusted using the m-l-1 small-sample correction (df = 15). *** p < 0.001.

shows that the national context, operationalized through national unemployment rates and immigration stock, structurally moderates the exclusionary impact of perceptions of immigrants as a fiscal burden. The simple slopes analysis, evaluated with a rigorous small-sample t(15) correction, confirms that the negative association between perceived fiscal burden and support for immigrants’ social rights remains highly significant across all levels of unemployment (p < 0.001). However, its magnitude varies considerably. In countries with relatively low unemployment (−1 SD), perception of immigrants as a fiscal burden is associated with a moderate reduction in support (b = −0.217). In stark contrast, this negative relationship is more pronounced in high-unemployment contexts (+1 SD), where the slope becomes markedly steeper (b = −0.458), demonstrating that structural economic scarcity acts as a powerful catalyst for welfare chauvinism when fiscal threats are politically salient. A similar pattern emerges regarding demographic density. In countries with a high immigrant stock (+1 SD), the negative association of perceived fiscal threat with support reduction is particularly severe (b = −0.430), compared with a weaker negative association in low-stock environments (b = −0.275).

Analysis of the perceived security threat (narratives linking immigrants to criminality) requires more nuanced consideration. Unlike concerns about the fiscal burden,

Table 4. Marginal effect of security threat on support for immigrants’ social rights, moderated by immigrant stock.

Demographic Context	Marginal Effect (b)	Robust SE	t(15)	Exact p-Value
Immigrant Stock
Low (−1 SD)	0.028	(0.041)	0.694	0.498
Mean (0)	−0.043	(0.032)	−1.347	0.201
High (+1 SD)	−0.114	(0.047)	−2.424	0.029 *

Note: Results are pooled across 20 multiply imputed datasets using Rubin’s rules. Significance tests are adjusted using the m-l-1 small-sample correction (df = 15). * p < 0.05.

shows that the cross-level interaction reveals a clear “threshold pattern” linked to demographic exposure. In countries with low or average immigrant populations, the security threat has no statistically significant association with support for rights. This exclusionary mechanism is observed only in high-stock contexts (+1 SD), where the perceived security threat becomes statistically significant in predicting lower support levels (b = −0.114). In other words, narratives criminalizing migrants are associated with attitudes of welfare exclusion solely when the demographic presence of out-groups is high.

To interpret these findings in substantive probabilistic terms and account for the nonlinear transformation of ordinal logistic models, Figure 2 illustrates the cumulative predicted marginal probabilities. The first observable trend in panels (a) and (b) is that as the perception of fiscal threat rises, the predicted probability of support for immigrants’ equal social rights is consistently lower. Second, both panels corroborate that the macroeconomic and demographic context significantly moderates this individual-level relationship.

Looking at panel (a), a cross-level dynamic emerges, reflecting the conditional nature of the initial positive association with unemployment (as seen in Table 1). In countries with high unemployment rates (+1 SD), individuals who perceive no fiscal threat exhibit the highest predicted probability of supporting immigrants’ social rights (approximately 90%). However, as the perception of threat progressively intensifies, this probability drops precipitously to roughly 64%. This steep disparity confirms that macroeconomic vulnerability amplifies welfare-chauvinist attitudes, a dynamic considerably less pronounced among individuals residing in countries with lower unemployment rates (−1 SD).

The national immigrant stock also exerts a powerful moderating role. In the absence of a perceived fiscal burden (−2 SD), the probability of support for immigrants’ social rights among individuals living in countries with a high immigrant population (+1 SD) borders 80%. Conversely, when the perception of threat peaks (+2 SD), this probability is notably lower (near 47%), which contrasts with the flatter slope observed among individuals in countries with a low proportion of immigrants (−1 SD), where support ranges from 87% under no perceived threat to near 70% when the threat is highly perceived. Unlike the fiscal burden perception, the perceived security threat has no significant association with support for rights, except in high-stock contexts (+1 SD), where it corresponds to a lower predicted probability of support, from approximately 70% to 60%.

As expected, these findings reveal a nuanced picture of how national contextual variables moderate perceived threats in the region. Both the unemployment rate and immigrant stocks act as structural catalysts, amplifying exclusionary attitudes when interacting with subjective fiscal threat perceptions. The negative association of the perceived fiscal burden with support for immigrants’ social rights is substantially stronger in countries with (a) higher unemployment rates and (b) larger immigrant stocks. Moreover, the statistical analysis reveals that (c) the perceived security threat solely predicts a lower likelihood of support in countries with a higher immigrant population (+1 SD). However, statistical analysis also revealed that neither the annual growth rate of GDP per capita nor the most recent flow of the immigrant population moderates these associations. Consequently, hypothesis H5b is partially substantiated: the national context significantly moderates the relationship between threat perceptions and support for immigrants’ social rights, but this moderation is contingent upon the type of threat and the specific macroeconomic or demographic variable considered.

5. Conclusions

The present study analyzed the influence of immigration-related threat and benefit perceptions, alongside political ideology and contextual factors, on public support for immigrants’ social rights in Latin America. Drawing on multilevel ordered logistic regression models estimated from Latinobarómetro survey data, this research reveals a political dynamic distinct from that of Western democracies regarding the role of ideology. Furthermore, our findings challenge some theoretical assumptions derived from intergroup threat theory and provide novel evidence supporting the Threat-Benefit Model in the Global South.

At the individual level, the findings demonstrate that threat perceptions linking migration to heightened labor market competition or increased crime rates do not exert a statistically significant effect on support for immigrants’ social rights (operationalized as equal access to healthcare, housing, and education). In contrast, perceiving this out-group as a fiscal burden on the state constitutes the most decisive factor in explaining low levels of support for their social rights. In a region characterized by economic precarity and weak social safety nets, citizens tend to view migrants as an additional strain on the provision of scarce public services and the allocation of taxpayer revenues, a dynamic that links generalized economic anxieties with targeted welfare chauvinism, partially accounting for anti-migrant hostility.

Conversely, those who associate immigration with both economic and cultural benefits to society exhibit higher levels of solidarity toward immigrants. Notably, the data indicate that support for immigrants’ social rights is more strongly associated with perceived benefits—particularly cultural enrichment—than with threat perceptions (including the framing of this out-group as a fiscal burden). These results reinforce arguments for examining the influence of both threats and benefits on citizens’ attitudes toward minority groups, given the central role of perceived benefits (Tartakovsky and Walsh 2016; Valenzuela et al. 2024). This finding extends the theoretical and empirical scope of the Threat-Benefit Model to the Global South. We found that, even in a context characterized by structural scarcity, where traditional intergroup threat theory would predict material competition to dominate attitudes toward immigrants’ social rights, the recognition of cultural and economic contributions is more relevant than threat perceptions, underscoring the asymmetrical power of positive symbolic framing in overriding material anxieties.

From a political–ideological perspective, the analysis also shows that, when using the political left as the reference category, people with a right-wing orientation are no less supportive of immigrants’ social rights than those with a left-wing identity. Rather, the most exclusionary attitudes are found among the politically disaffected or those disillusioned with conventional politics. This finding holds particular analytical and normative relevance, given that this demographic represents a population segment highly susceptible to populist or nationalist discourses operating outside traditional ideological channels (Norris and Inglehart 2019). Amid the Latin American migration crisis, such widespread political disaffection provides fertile ground for the politicization of immigration and the spread of exclusionary narratives. Ultimately, the division that emerges between citizens over solidarity with immigrants is not between the ideological left and right but between individuals integrated into the formal political system and those who feel detached from it, an emerging systemic cleavage that warrants scholarly attention.

At the macrosocial level, the results also show that contextual factors not only directly correlate with public support for immigrants’ social rights but also moderate the influence of perceived threats. The national unemployment rate and the immigrant stock strongly amplify the exclusionary effect of the fiscal threat. In countries with high unemployment or a large immigrant population, the perception of a fiscal threat triggers an accelerated decline in support levels. Crucially, when fiscal concerns are absent, although widespread national unemployment may initially correlate with elevated solidarity toward immigrants (reflecting a broad demand for state protection), it becomes a powerful catalyst of welfare chauvinism the moment that immigrants are perceived as a fiscal burden for the state. Structural scarcity alone does not necessarily generate welfare-chauvinist attitudes; rather, it requires the political activation of a fiscal threat narrative within that environment. Furthermore, fears of an immigration-related increase in crime become salient and politically consequential only in contexts of high migrant concentration, being directly associated with diminished support for the social rights of this out-group.

From a public policy standpoint, particularly in a region characterized by historically weak party system institutionalization (Mainwaring 2018), these findings underscore the importance of moving beyond traditional left–right divides to safeguard the social rights of the foreign-born population. To lessen anxieties among citizens and promote greater acceptance of immigrants’ social rights, the positive impact of migrants’ perceived cultural contribution offers an opportunity for intervention—through the media and the educational system—focused on their symbolic contributions to society. Indeed, emphasizing the benefits that immigrants bring is a highly fruitful policy strategy, as fostering positive perceptions is structurally more viable and politically more feasible than dismantling deeply entrenched threat narratives (Valenzuela et al. 2024; Walsh and Tartakovsky 2020). Moreover, the robust association between perception of this out-group as a fiscal burden and opposition to its social rights suggests that governments must communicate more effectively how migrant integration can bolster long-term sustainability of the welfare state and tax revenues. Simultaneously, public institutions must assume a more proactive role in mitigating public anxieties by providing evidence-based information, thereby helping citizens disentangle empirically unfounded perceptions from the actual societal impacts of migration (Tartakovsky and Walsh 2020).

Finally, we identify four key areas for further investigation. First, to overcome the limitations of cross-sectional designs (Careja and Harris 2022) in establishing robust causal inferences, longitudinal panel data are needed (Kwon et al. 2024). Second, the findings regarding the limited explanatory power of ideology point out the need to examine other national-level contextual variables—such as the electoral success of anti-immigrant populist parties, political discourse, and media coverage—given their close association with public perceptions of this out-group (Kleider 2022; Koning and Kaushal 2024; Pari-Bedoya 2026). Third, future studies should try to replicate these findings across diverse geographical contexts to more clearly determine how perceived economic and cultural benefits shape citizens’ attitudes toward immigrants (Valenzuela et al. 2024). Fourth, further research should also address the potential bias that socioeconomic status estimates may introduce into statistical analyses, as they can distort relationships among variables (Jabkowski and Piekut 2024). This consideration is especially relevant when estimates rely on income or occupation, since the underrepresentation of certain social segments in national survey samples can affect the statistical significance tests intended to assess the incidence of socioeconomic status on attitudes toward immigrants. It is therefore essential to handle (i) nonresponse in income questions, which tends to cluster at the extremes of the distribution (Bollinger et al. 2019; Hlasny 2020), and (ii) the underreporting of income, particularly visible in households with self-employed or informal workers (Kukk and Staehr 2017; Kukk et al. 2020) and in high-income households (Alvaredo et al. 2025; Larrañaga et al. 2022; Lustig and Vigorito 2025). On this latter point, it is important to note that insofar as a substantial share of workers in Latin America belong to the informal sector of the economy, this can also affect traditional sociological estimates of social class based on occupation.