Immigration and the Roots of Nationalism: Evidence

from a Sudden Demographic Change in Chile*

Giancarlo Visconti† Catalina Vega-Mendez‡
Pennsylvania State University Purdue University

May 5, 2023

Abstract

Do demographic changes caused by migration make people more nationalistic? An-

swering this question is difficult because demographic transformations can boost far-
right parties using nationalistic rhetoric, and therefore people’s new attitudes could
be elite-driven. Using administrative data and a three-wave panel survey from Chile
before the emergence of a far-right party, we show that immigration shocks (i.e., sud-
den demographic changes) make people prouder of being Chilean and more likely to
identify with Chile. Additionally, we find that municipalities that are more exposed to
immigration exhibit an increase in crimes associated with hostility toward foreigners
and an improvement in the electoral performance of far-right parties (when formed).
We also provide survey evidence from eight Latin American countries to enhance the
study’s external validity. Consequently, these findings reveal that people’s national-
istic attitudes after demographic changes are not purely elite-driven, which can have
important implications when studying the origin of far-right parties.

Keywords: Nationalism, Immigration, Far-Right Parties, Panel Data, Chile.

* We thank James McCann, Ilona Lahdelma, Amy Liu, Virginia Oliveros, Peter Siavelis, Logan Strother, Cara
Wong, Tianhong Yin, Elizabeth Zechmeister, and seminar participants at Illinois Urbana-Champaign, Migration Policy
Institute, Purdue, LASA, and APSA for helpful comments and suggestions. All errors are our own.
† Assistant Professor, Department of Political Science, Pennsylvania State University, University Park, PA 16802,

email: gvisconti@psu.edu.
‡ Ph.D candidate, Department of Political Science, Purdue University, West Lafayette, IN 47907, email:

cvegamen@purdue.edu.
1 Introduction

In the last decade, the world has experienced deep demographic transformations due to mi-
gration, with people leaving unsafe and impoverished countries for safer and more stable places
(Holland and Peters, 2020). In this context, the use of nationalistic rhetoric in politics has risen
globally in the last few years (Bremmer, 2017). In particular, large migration flows changing the
population structure of a country might affect native-born citizens’ identity and self-image, making
them more prone to nationalistic attitudes. This change could have important consequences since
strong attachments to a national identity may push people to engage in prejudice, discrimination,
and even violence against out-group members (Druckman, 1994; Leyens et al., 2003).
While extensive literature has studied the political consequences of demographic changes, this
research agenda has mainly explored how immigration affects economic and cultural grievances.1
However, while these factors have relevant political implications, a large migratory wave can also
change people’s political ideas, attitudes, and sentiments beyond these traditional grievances. Con-
sequently, in this paper, we study whether immigration shocks (i.e., sudden demographic changes)
affect individuals’ nationalistic orientations, their behavior toward migrants, and their support for
far-right parties.
However, it is difficult to answer these questions because demographic transformation, in ad-
dition to increasing nationalistic attitudes, can also foster the (re)emergence of far-right parties
(Lubbers et al., 2002; Abrajano and Hajnal, 2015; Hangartner et al., 2019; Reny et al., 2019; Di-
nas et al., 2019).2 These political organizations promote ideas such as the congruence between

1 For example, studies have shown that exposure to immigration can increase economic distress
(Dehdari, 2019) and spark a backlash against cultural and social transformations (Inglehart and
Norris, 2016).
2 However, other studies have not found a connection between immigration and far-right par-
ties, which could be explained by the size of the immigrant community (Arzheimer and Carter,
2006; Lucassen and Lubbers, 2012), by the existence of deeper rather than superficial interactions

1
state and nation (Mudde, 2007; Golder, 2016) and have fueled the public discourse with nation-
alistic ideas (Rydgren, 2006). In this context, people’s political attitudes could be driven by the
elite (Vrânceanu and Lachat, 2018; Luttig, 2020; Smith et al., 2021), and as a result, not directly
explained by demographic changes but by the rise of far-right parties. To solve this problem, we
study whether demographic changes make individuals more likely to have nationalistic attitudes
before the emergence of a far-right party. If so, individuals’ new attitudes might make far-right
candidates more attractive to voters.
Past research has overlooked how people change their political attitudes after significant demo-
graphic transformations. As Hangartner et al. (2019) hold, "little is known, however, about how
refugee migration shapes the attitudes, policy preferences, and political behavior of European citi-
zens." We know even less about the political consequences of demographic changes in the Global
South. The lack of research on this region is striking since developing countries have been the
"most impacted by the recent wave of migration, which has been driven largely by refugees and
asylum seekers" (Alrababa’h et al., 2021, p.35). While a large proportion of migration occurs
within the Global South (UNHCR, 2017), almost all the research on this topic has focused on
receiving countries located in the Global North.3
The case of Chile provides an opportunity to study the political effects of immigration. Since
2015, Chile has experienced a rapid and sudden demographic transformation due to migration from
other Latin American and Caribbean countries, and certain parts of the country experienced drastic
demographic changes, but others did not (Bellolio and Valdés, 2020). We take advantage of rich
administrative data on the visas requested at the municipality level from 2014 to 2017, focusing

between natives and immigrants (Andersson et al., 2020; Schaub et al., 2021), or by immigrant
characteristics (Hainmueller and Hiscox, 2010; Valentino et al., 2019).
3 There are a few examples of research studying the political consequences of migration in
Latin America, such as Malone (2019), Vega-Mendez and Visconti (2021), and Holland et al.
(2021). Also, there is research studying attitudes and preferences of Latin American emigrants
such as Escobar et al. (2015) and McCann and Jones-Correa (2021).

2
on how these rates have changed from year to year. During this period, the country allowed people
from other Latin American countries to travel without a tourist visa, and then they could request
a (non-tourist) visa while in Chile as visitors. As a result, visa requests can work as an accurate
measure of actual demographic changes (before the 2018 immigration reform).
Additionally, the Chilean case provides a unique chance to address the endogeneity problem
between the emergence of far-right parties and political attitudes since a far-right party was formed
in 2019 (Luna and Rovira Kaltwasser, 2021). In 2021, this newly formed party was able to secure
multiple seats in Congress and had a historical result in the presidential election, making it into the
second round and obtaining 44% of the votes. By using panel data from 2016 to 2018, it is possible
to leverage the timing of both the surveys and the rise of a national far-right political organization
in Chile to rule out the possibility of political attitudes being driven by the launch of this political
party.4
To conduct the analysis, we combine the administrative data (to measure exposure to immigra-
tion) with a three-wave panel study (to capture the outcomes). To be able to draw inferences about
the effects of an immigration shock on nationalistic attitudes, we use a dynamic or event-study
difference-in-differences design (DiD), which provides the effects of different lengths of exposure
to an immigration shock.
When examining the three waves, we find that immigration shocks make people more likely
to be proud and to identify with Chile. However, an increase in national pride and identity after a
demographic change can have two interpretations. On the one hand, it could be read as an iden-
titarian reaction to newcomers, which could lead to discrimination, prejudice, and even violence
(i.e., nationalism). On the other hand, it could be related to being proud of your country for helping
people in need and not leading to exclusionary attitudes and behaviors (i.e., patriotism).

4 Before the emergence of this party, its leader, José Antonio Kast, ran as an independent
presidential candidate in 2017. He obtained less than 8% of the votes and did not make it into the
second round. He had a minor role in that election with no party, national political infrastructure,
or legislative candidates.

3
 To address these two possible interpretations, we study whether immigration shocks increase
hostility against migrants. While patriotism should not increase exclusionary behaviors, national-
ism may increment hostility toward foreigners. Prior research has shown how negative framings
about migration can result in severe forms of hostile behavior, including a rise in ethnic violence
and hate crimes against migrants (Igarashi, 2021; Dipoppa et al., 2023). We find that crimes as-
sociated with violence toward foreigners increased due to recent migration. Additionally, we use
secondary sources to document exclusionary attitudes and behaviors towards migrants in Chile.
As a result, all this evidence points to an increase in nationalism (rather than patriotism) after a
sudden demographic change.
Additionally, we show that areas more exposed to demographic changes were more likely
to support the newly formed far-right party in the 2021 presidential election. Finally, we also
provide survey data from eight Latin American countries and compare attitudes before and after
the immigration wave generated by Venezuela’s 2015-2016 political and economic crisis. This
extra analysis allows us to improve the external validity of the main findings.
The debate about why immigration contributes to the rise of far-right parties will likely remain
a vibrant discussion in comparative politics. The effects of electing far-right leaders on democratic
norms and institutions make this a timely and concerning topic to be studied. Traditional expla-
nations point to people’s economic and cultural grievances after immigration shocks. However,
we provide a complementary argument by showing that native residents’ fundamental political
attitudes can change in multiple ways making far-right alternatives more attractive to them. In par-
ticular, an increase in nationalistic attitudes can make voters more likely to support parties that will
strongly base their electoral platforms on the use of nationalistic and anti-immigration rhetoric.
Consequently, the rise of far-right political options might have more dimensions than previously
thought.
These findings also contribute to research on the political consequences of immigration influxes
and to the literature about the formation and evolution of political attitudes, as well as help reduce
geographic imbalances in the research about demographic changes due to migration. In addition,

4
the rise of nationalistic attitudes shows that even if mainstream parties adopt more strict measures
to curb immigration, they might not necessarily receive an electoral boost. Native-born residents
might be seeking something else than the curtailing of demographic changes.

2 Theoretical Background

Prior research has documented how demographic changes can trigger economic and cultural
grievances in the native-born population (Hainmueller and Hopkins, 2014). However, by only pay-
ing attention to these factors, we may overlook a crucial dimension of this phenomenon: people
exposed to immigration may not just be concerned about labor market effects and cultural trans-
formations. Still, they may also change their fundamental political ideas such that far-right parties
become a viable electoral option for them. More specifically, demographic changes may increase
feelings of loyalty or attachment to in-group members, which could translate into nationalistic at-
titudes. Research into these possible effects can contribute to a more complete understanding of
the connection between immigration and the rise of the far-right.

2.1 Nationalism

When people categorize themselves and others, in-group and out-group biases can emerge. In-
dividuals are prone to classify individuals and objects into social categories for cognitive reasons
(Tajfel and Turner, 1982). In particular, people will affirm their self-concept by being a member of
a group and making social comparisons between out-groups (Dunne, 2018). In this context, group
membership "leads people to favor that group and see other less worthy in comparison” (Druck-
man, 1994, p.48). More specifically, native-born citizens can enhance their group self-image and
identity as a preservation mechanism when confronted with an out-group such as migrants. This
process can foster nationalism as part of the “us” versus “them” conflict. People’s nationality
can work as an easy heuristic for in-group bias in the context of large migration and low ethnic
polarization, as in Latin America. Consequently, our first empirical expectation is the following:

5
 Hypothesis 1a: Demographic changes due to immigration will increase citizens’ national
pride and identity.

If people have become more nationalistic after an immigration shock, these new attitudes could
lead to hostility toward out-groups, such as discrimination, exclusion, and violence. Previous
research has shown how nationalism is connected with exclusionary attitudes and behaviors against
migrants (Pedersen et al., 2005; Gijsberts and Hagendoorn, 2017). By definition, nationalism refers
to people’s belief in the superiority of their country and that dominance can be expressed through
hostile behavior. Therefore, if national pride and identity are related to nationalism, we could
expect that an immigration shock will increase hostility against foreigners.

Hypothesis 1b: Sudden and rapid demographic changes due to immigration will increase
hostility toward immigrants.

2.2 Alternative Expectations

A first alternative expectation is that immigration shocks do not affect people’s attitudes. Na-
tionalistic orientations have been called “self-referential” by part of the literature (Jeong, 2013),
meaning that they might not be conditioned to the presence of others. Taking into account that
political attitudes tend to be explained by slow-moving variables such as ideological stances, par-
tisanship, and socialization processes (Campbell et al., 1980), it is possible to expect that a contex-
tual factor is not going to change people’s ideas about their nation. Following this logic, we could
expect the following:

Hypothesis 2a: Sudden and rapid demographic changes due to immigration will NOT increase
citizens’ national pride and identity.

However, if there is an increase in national pride and identity, that could be explained by patrio-
tism rather than nationalism. So, let’s compare these two concepts. Nationalism has been tradition-
ally conceptualized as an ideological feature promoting that the nation should be congruent with

6
the state and perceptions of superiority and dominance against other countries or nations (Druck-
man, 1994; Golder, 2016; Mudde, 2007),5 which translates into seeing immigrants as threats and
competitors (Osborne et al., 2017; Zhai and Yan, 2022). In contrast, patriotism refers to affection
toward your own nation and comes with a sense of responsibility and pride (Zhai and Yan, 2022),
which is often related to more tolerant attitudes toward immigrants (de Figueiredo Jr and Elkins,
2003; Jeong, 2013). Therefore, while nationalism encourages a sense of superiority and preju-
dice of one’s own group over “others,” patriotism seems to lead to a strong attachment to a nation
without the negative evaluation of “others" (Druckman, 1994).
As a result, patriotism is not connected to hostility. In fact, the definition of patriotism is
compatible with certain kinds of cosmopolitanism (Audi, 2009). Prior evidence has shown how
patriotism can be a predictor of positive behaviors such as civic activity (Marz˛ecki, 2020), tolerance
toward newcomers (Green et al., 2011), and pro-immigrant sentiments (Jeong, 2013; Huddy et al.,
2021). Consequently, if an increase in national pride and identity after an immigration shock
is explained by patriotism, these new attitudes should not be attached to more hostility toward
migrants.

Hypothesis 2b: Sudden and rapid demographic changes due to immigration will NOT increase
hostility toward immigrants.

Figure 1 summarizes all the hypotheses. A sudden demographic change due to an immigration
shock can affect people’s attitudes differently. First, exposure to migration can increase national
pride and identity (H1a) and hostility against migrants (H1b), suggesting that people adopt nation-
alist attitudes due to demographic changes. Second, exposure to migration can increase national
pride and identity (H1a) but without affecting behavioral outcomes, suggesting that people are
5 One of its most important varieties is ethnic nationalism, which refers to an exclusive notion
of nationhood where membership in the ethnic nation is the only way people can gain citizenship
rights, and this is conditional on certain hereditary traits such as native birth, religion, race, culture,
or language (Brubaker, 2009).

7
adopting patriotic attitudes (H2b). Finally, exposure to migration can have no effect on national
pride and identity and therefore have no impact on people’s nationalism and patriotism (H2a).

Increase Evidence of
Hostility against Nationalism
Increase Migrants
National Pride and
Not affect
National Identity Evidence of
Hostility against
Migrants Patriotism
Demographic
change

Not affect No evidence of

National Pride and Nationalism or
National Identity Patriotism

Figure 1: Empirical Expectations

3 Immigration and the Far Right in Chile

As previously mentioned, the significant geographical imbalances in knowledge about the po-
litical effects of demographic changes are particularly concerning given that a large proportion
of migration patterns currently occurs between countries in the Global South (since refugees and
economic migrants commonly relocate to places close to their countries of origin). For example,
in Africa, Uganda has been receiving refugees from neighboring countries experiencing civil war
and political violence for decades. Since 2011, the country has received over one million refugees
from South Sudan alone (UNHCR, 2017). In the Middle East, Lebanon has received close to
two-thirds of the Syrians displaced by the civil war that that country has suffered in the last decade
(Corstange and York, 2018). In South America, Venezuela’s socioeconomic collapse has generated
large-scale displacements to other countries in the region (Holland et al., 2021). In the Caribbean,
the 2010 earthquake in Haiti produced massive numbers of people who migrated to the US and
Latin America (Esnard and Sapat, 2011). A country that became a major recipient of migrants in
the last decade was Chile, receiving people mainly from Venezuela and Haiti, a process that started
in the mid-2010s (Bellolio and Valdés, 2020).

8
 The case of Chile is particularly interesting. Historically, this country has had a small propor-
tion of foreign-born residents that did not reach more than 1 percent of the inhabitants in 2002.
However, this has changed drastically in recent years, and immigration rates have grown faster in
Chile than in any other country in Latin America (Doña Reveco, 2018). One of the consequences
of this rapid and sudden demographic transformation has been the politicization of immigration,
as in the United States or Europe (Arostegui, 2018).
An excellent proxy of actual immigration in Chile is using visa requests before the year 2018.
Nine out of ten migrants in Chile come from other countries in Latin America (Bellolio and Valdés,
2020), and this is relevant because Latin American citizens could enter as visitors Chile without a
tourist visa,6 and a large part of them did not need a passport but just a valid ID.7 After entering the
country as a visitor, Latin American citizens could request a (non-tourist) visa following a simple
bureaucratic process. Migrants needed to present an employment contract to get a temporary visa,
and then after two years, they could obtain permanent residency (Stefoni, 2011). Employment
contracts were not limited to working in a company or an industry, and they could take less formal
agreements such as working for an individual taxpayer or a household as a nanny or contractor
(Fernandez, 2017).
Taking into account that Chile had a strong focus on securing borders before 2018 (Aedo, 2017)
and that obtaining a regular status allowed migrants to receive social benefits and have access to
key public services (Ministry of Interior, 2017);8 then it is not surprising that the official numbers
of irregular crossings were very low, with an average of three people arrested per day (Vedoya,
2017), and that the most common way to migrate to the country was to request a visa while in
Chile (Severino and Visconti, 2023).
However, these conditions dramatically changed in 2018 with the administration of then-

6 Only with the exception of people from the Dominican Republic.

7 Argentina, Bolivia, Brazil, Colombia, Ecuador, Peru, Uruguay, and Paraguay.
8 For example, to be part of the public health system FONASA or to receive financial support
to rent a home.

9
president Sebastian Piñera. He passed a bill reforming the immigration law, and after that change,
people from both Venezuela and Haiti had to request a tourist visa before traveling to Chile as
visitors, and these were the two countries with the largest percentage of migrants entering Chile
(Doña Reveco, 2018; Bellolio and Valdés, 2020). Consequently, the strategy of migrating as a
tourist and using the 90 days window to get an employment contract becomes infeasible for a large
proportion of migrants. Then, after that reform, the numbers of irregular crossings skyrocketed,
which means that using visas as a proxy for migrations stopped being an accurate and informative
measurement approach. As a result, we rely on administrative data covering the period before the
immigration law adopted in 2018.
Regarding Chile’s political configuration, after the transition to democracy in 1990, Chilean
politics was dominated by a center-left and a center-right-wing coalition. The former won four
presidential elections in a row, and the latter broke that pattern in 2009. Both coalitions traded
control of the executive power until 2022. Over time, these two large political alliances have con-
verged toward the center (Visconti, 2021),9 which can be explained by the center-left government’s
decision to keep the market-based reforms introduced by Pinochet during the military dictatorship
(Maillet, 2013), and by the moderation of the center-right parties that have appropriated elements
of social welfare policies (López and Baeza, 2011; López et al., 2013).
However, the ideological landscape has changed in recent years. In 2017 a left-wing coalition
called "Frente Amplio" emerged. They obtained seats in Congress in the parliamentary election
of 2017, and they won the presidential election of 2021. Meanwhile, in 2019, a far-right party
was created called the Republican Party (Meléndez et al., 2021).10 In 2022, they won seats in
the Lower and Upper Chambers of Congress, and they made it into the second round of the pres-
idential election, obtaining 44% of the votes. The Republican Party became a party of national
relevance in only a few years. Consequently, Chile provides a perfect opportunity to learn about

9 At least until the second election of Michelle Bachelet in 2014.

10 See Rovira Kaltwasser (2022) and Zanotti and Roberts (2021) for reviews on the emergence
of the (populist) right in Latin America.

10
the consequences of demographic changes in the Global South. This country has experienced un-
precedented immigration waves in the last decade, and the pre-2018 immigration law allows us to
use visa requests as an accurate strategy to measure local demographic changes.

4 Empirical Strategy

Several methodological challenges must be addressed to study the effects of demographic

changes. One of the most important is the non-random nature of migration (Dancygier and Laitin,
2014). Specifically, areas that experience the arrival of a significant number of newcomers might
differ in terms of observed and unobserved covariates from places that do not experience these
transformations. To address this concern, we use a three-wave panel survey and detailed adminis-
trative data from Chile to implement a dynamic difference-in-differences (D-DiD) and estimate the
impact of immigration on nationalistic attitudes. Additionally, we take advantage of the rapid de-
mographic transformation in Chile since the country started to receive large numbers of immigrants
in 2014–2015.
A difference-in-differences design takes advantage of pre-and post-treatment longitudinal data
to estimate the effects of a given intervention. This design relies on comparing outcomes over time
between a treatment and a control group. These groups do not need to be similar but do need to
follow a similar trajectory before the treatment (i.e., a parallel trend assumption). As a result, any
difference in their trajectories after the treatment can be attributed to exposure to the treatment. A
DiD is a particularly useful design when the treatment cannot be randomized but when it is possible
to exploit the longitudinal nature of the data to provide more credible inferences than studies that
only adjust for cross-sectional variables.
In particular, we use dynamic or event-study difference-in-differences, which is ideal when the
treatment occurs across multiple time periods (Callaway and Sant’Anna, 2020). When relying on
a dynamic DiD, effects are aggregated by the length of exposure. For example, we can learn the
impact of being exposed to an immigration shock just once or more than once. This approach helps

11
to facilitate the interpretation and analysis of results when having multiple time periods. When
using a dynamic DiD, the treatment follows a staggered adoption, meaning that when subjects
are treated, they will remain in the treatment group, and the control group is composed of only
never-treated units (Callaway and Sant’Anna, 2020).
Measuring exposure to immigration is not an easy task. One option is to use perceptions of de-
mographic changes, but previous studies have shown that people’s perceptions tend to be endoge-
nous to their political attitudes (Evans and Andersen, 2006). A possible solution to this problem is
to use administrative data to calculate immigration rates. However, demographic changes can be
explained by both regular and irregular migration, and administrative data can only inform us about
the former. The case of Chile provides an opportunity to address this concern since, as explained
before, people from Latin America could easily request a (non-tourist) visa while in the country as
visitors. As a result, demographic transformations before 2018 were mainly explained by regular
migration. We take advantage of rich and detailed administrative data that provides all of the visa
requests made in Chile between 2014 and 2017 and includes information about the municipality
where the migrant was living in. This data makes it possible to compute immigration patterns at
the municipality level.
Another potential issue when studying the impact of immigration is that native residents from
areas highly exposed to immigration might get used to these demographic changes and, as a result,
not change their political attitudes after the arrival of foreigners. Previous studies have shown the
limitations of using immigration rates to measure perceived immigration (Newman and Velez,
2014) and that people focus on changes rather than absolute levels (Kahneman and Tversky,
2013). To address this concern, we pay attention to immigration shocks or substantive demo-
graphic changes due to migratory waves that occur in a short time period (Severino and Visconti,
2023).
In particular, we compute the change in immigration year to year to measure demographic
transformations. For example, for the survey implemented in 2017, we used the number of visas
requested in 2016 and 2015 to estimate the demographic change. We calculate the change in

12
percentage points between each time period (i.e., changes in visa requests from one year to the
other).11 We define exposed municipalities as those where the change in visa requests is equal to
or greater than one standard deviation above the mean (considering all of the differences between
years in a given survey wave). Control municipalities are those in which demographic changes
were lower than one standard deviation. Survey respondents living in exposed municipalities are
considered exposed subjects, and when living in control municipalities, they are considered control
subjects. The use of a one standard deviation increase is a common strategy when studying the
impact of different types of shocks (see Bazzi and Blattman (2014) and Carreras and Visconti
(2022) for examples). Importantly, we also use a continuous version of the exposure indicator in
section 5.2 and find the same results. Therefore, the study’s conclusions are not conditional on
using a binary or a continuous indicator of demographic changes.
The panel study is based on a national representative sample and was implemented face-to-face
by the Centre for Social Conflict and Cohesion Studies (details about sampling can be found in the
supplementary appendix). We use two questions from this study to learn how demographic changes
(measured using administrative data) affect nationalistic attitudes (measured using panel survey
data). Both outcome questions follow the same structure: "To what extent are you in agreement or
disagreement with each one of the following statements?" The five-point scale answer goes from
1: strongly disagree to 5: strongly agree. The statements are: "I feel proud to be Chilean" and "I
identify with Chile." We standardized all the outcomes to facilitate their interpretation.
In summary, the final dataset has the following structure: the units of analysis are survey re-
spondents embedded in panel data, and exposed individuals live in an exposed municipality where
immigration increased by more than one standard deviation in the last year. Meanwhile, control
subjects live in municipalities that were never exposed to an immigration shock. The outcomes
measure national sentiments toward Chile.
Finally, the panel survey data includes 25 percent of all municipalities in Chile, but these 92
municipalities cover 67 percent of the population. Figure 2 depicts the municipalities in the control

11 Percentage change: (Visas year 1 – Visas year 2)/(Visas year 2).

13
group (i.e., never treated) and in the exposed group (i.e., regardless of time of exposure). Both
groups do not exhibit large differences regarding their population size, poverty levels, and electoral
behavior (pre-treatment characteristics). However, difference-in-differences designs do not require
covariate balance but parallel trends as the key identification assumption.

Figure 2: Map of Chile depicting municipalities that: (i) were exposed to an immigration shock,
(ii) were not exposed to an immigration shock, and (iii) were not included in the panel survey data.

14
5 Results: Proud and Identity

5.1 Dynamic Difference-in-Differences

Because it is not easy to interpret a design based on multiple time periods, a common approach
is to aggregate group-time effects into an event-study plot (Callaway and Sant’Anna, 2020). This
analysis will provide the average treatment effects at different lengths of exposure. Since there are
municipalities that were never exposed to an immigration shock, and the rest were exposed at a
different time, the dynamic DiD allows us to compute the effects of an immigration shock based on
the length of exposure. We report results for units treated for the first time (i.e., an instantaneous
treatment effect) and the second time to an immigration shock (i.e., the second survey where units
are considered exposed).12
Figure 5 reports the main results of immigration shocks on nationalistic attitudes. The dots rep-
resent the average effects, and the lines the 95% confidence intervals (computed using bootstrapped-
based standard errors). Results in grey correspond to the pre-exposure analysis; in particular, this
is the comparison between never treated and eventually treated. The lack of evidence of a signifi-
cant effect will work as evidence to support the parallel trend assumption (i.e., both groups would
be following a common trend before exposure). Results in black correspond to the post-exposure
analysis or to the effects of an immigration shock by the length of exposure.

12 As expected in a dynamic DiD, we do not use subjects exposed in the first wave or exposed
three times (i.e., always treated) because they do not provide pre-treatment information (Callaway
and Sant’Anna, 2020).

15
 (a) Proud of Chile (b) Identify with Chile

Figure 3: Average effect of immigration shocks on political attitudes by the length of exposure. A
length of exposure of -1 refers to the period before first exposure, of 0 to first exposure, and of 2
to second exposure. 6,249 observations.

The results for both outcomes are not significant before first exposure (proud: 95% CI: [-0.371,
0.023], identify: 95% CI: [-0.286, 0.089]), which provides evidence for the parallel trend assump-
tion (i.e., both groups follow the same trajectory in the pre-treatment period, which is expressed
by the null results). Also, there is no evidence of an effect during first exposure (proud: 95% CI: [-
0.103, 0.169], identify: 95% CI: [-0.170, 0.102]). However, an effect becomes clear one year from
first exposure. In particular, a second exposure to an immigration shock increases national pride
by 0.34 standard deviation points (95% CI: [0.153, 0.537]) and national identity by 0.20 standard
deviation points (95% CI: [0.010, 0.383]). In the supplementary appendix, we include an analysis
with covariates, and the study’s conclusions remain.
These results show that people do not change their political attitudes immediately after ex-
posure to an immigration shock. However, their new attitudes crystallize in only one year from
exposure suggesting that some contact with migrants is needed to affect their political orientations.

16
5.2 Generalized Difference-in-Differences

In the main analysis, we use a binary version of exposure to migration. This section uses a
continuous measure of demographic changes as a robustness check. We use the percentage-point
change between years 1 and 2 before the survey wave to measure demographic changes. Since our
previous design can only be implemented with a binary treatment (Callaway and Sant’Anna, 2020;
Callaway and Sant’Anna, 2021), we rely on this section on a generalized difference-in-differences
design or two-way fixed effects. We use waves to capture time-fixed effects and respondents or
municipalities for unit-fixed effects. Therefore, we use this robustness check to evaluate whether
the results are robust to using a different estimation approach (generalized rather than dynamic
DiD) and a different exposure indicator (continuous rather than binary). Tables 1 and 2 provide
the results of the impact of a demographic change on the two outcomes of interest. Both the
outcomes and the demographic change indicator are standardized to facilitate the interpretation of
the analysis.

Table 1: Generalized DiD using a continuous exposure indicator and proud to be Chilean as the
outcome
I feel proud to be Chilean
(1) (2) (3) (4)
Demographic change 0.0413∗∗ 0.0405∗∗ 0.0412∗∗ 0.0408∗∗
(0.0143) (0.0143) (0.0144) (0.0144)
Controls No Yes No Yes
Wave fixed effects Yes Yes Yes Yes
Respondent fixed effects Yes Yes No No
Municipality fixed effects No No Yes Yes
Observations 6,214 6,211 6,214 6,211
Note: ∗ p<0.05; ∗∗ p<0.01; ∗∗∗ p<0.001

Both results show that a one standard deviation increase in demographic changes due to mi-
gration increases national pride by 0.04 and national identity by 0.03 standard deviation units. It
is important to note that effect sizes cannot be compared with the dynamic DiD because of the
different structures of their exposure indicators. However, the direction and significance of the

17
Table 2: Generalized DiD using a continuous exposure indicator and identify with Chile as the
outcome
I identify with Chile
(1) (2) (3) (4)
Demographic change 0.0338∗ 0.0334∗ 0.0334∗ 0.0329∗
(0.0138) (0.0138) (0.0139) (0.0139)
Controls No Yes No Yes
Wave fixed effects Yes Yes Yes Yes
Respondent fixed effects Yes Yes No No
Municipality fixed effects No No Yes Yes
Observations 6,229 6,226 6,229 6,226
Note: ∗ p<0.05; ∗∗ p<0.01; ∗∗∗ p<0.001

effects are the same, which illustrates that the study’s main conclusions are robust to different
specifications.

6 Results: Hostility toward Migrants

Hostility toward migrants can take multiple forms, such as discrimination, xenophobic speech,
exclusionary rhetoric, and violence, which is its most extreme version. We rely on two types
of evidence to elucidate whether these new attitudes come with changes in people’s behavior to-
ward migrants: (i) We utilize crime rates at the municipality level and a dynamic difference-in-
differences design to evaluate whether certain types of crimes associated with hostility against
migrants increase after an immigration shock. (ii) We use different secondary sources such as
surveys, interviews, and media.
As a first piece of evidence, we evaluate whether reports of crimes connected to hostility to-
wards migrants increased due to large demographic changes. Ideally, we would use hate crimes
against migrants as the outcome; however, the undersecretary of crime does not provide that
data. Nevertheless, previous studies and media reports show that hate crimes and hostility towards
migrants usually involve affrays, assaults, damages, and disorderly conduct (Arellano Calderón,
2022); and we have municipality-level data for these crimes allowing us to construct a proxy of

18
hostility towards migrants. Figure 4 uses a dynamic DiD to study the impact of demographic
changes on this type of crime. The design is the same as the main analysis used in section 5, but
now we use municipalities-waves as the level of analysis rather than panel survey respondents. The
number of observations is 276, which are all the municipalities-waves included in the panel study.

Figure 4: Average effect of immigration shocks on crimes associated with hostility toward mi-
grants. A length of exposure of -1 refers to the period before first exposure, of 0 to first exposure,
and of 2 to second exposure. 276 observations.

The findings are congruent with the previous analysis: there is support for the parallel trend
assumption when checking pre-exposure trends (95% CI: -0.197, 0.523), there is no immediate ef-
fect of immigration shocks (95% CI: -0.055, 0.364), and there is a significant change after one year
of exposure. In particular, crimes associated with hostility towards migrants increased by 0.230
standard deviation points (95% CI: 0.008, 0.442). We use a continuous version of demographic
changes in the supplementary appendix and also find a significant increase in crimes associated
with hostility against foreigners.
Of course, a natural concern is that these crimes and offenses can increase not as a result of
discrimination but because of migrants committing those crimes. To rule out this possibility, we
implement a falsification test for an effect that we expect to be absent to confirm the existence of
hostility toward migrants. In particular, we identify the impact of immigration shocks on crimes

19
not related to violence against migrants, such as incivility crimes (i.e., public intoxication and
disturbance of the peace), property crimes (i.g, burglaries, robbery, theft, robbery by surprise, failed
robbery, and handling of stolen goods), violent crimes (i.e., homicides, sexual abuse, domestic
violence against women, men, children, and the elderly) and weapon-related crimes (i.e., illegal
carrying of weapons and illegal possession of weapons). If immigration shocks are not increasing
these sixteen different types of crimes, we will have strong evidence to believe that the increase of
crimes associated with hostility towards migrants is not explained by migrants committing them.

Figure 5: Average effect of immigration shocks on placebo crimes. A length of exposure of -1

refers to the period before first exposure, of 0 to first exposure, and of 2 to second exposure. 276
observations.

20
 As a second piece of evidence, we use secondary sources to confirm these previous findings.
Survey evidence from a study implemented only for migrants in Chile in 2018 shows that half of
the respondents reported perceptions of discrimination from Chileans. Migrants felt discriminated
against in various contexts, such as when applying for jobs, when trying to rent a house, when using
public transportation, and when looking for health care, among other instances (CENEM, 2018).
Qualitative evidence from interviews with migrants supports these previous findings. Newcomers
reported differential treatment in private and public services, such as Chileans being attended to
before migrants regardless of the time of arrival (Rojas Pedemonte et al., 2015). Latin American
migrants in Chile had to encounter multiple obstacles in their integration process, and good ev-
idence of that was a large anti-immigration demonstration in the north of Chile in 2022, where
Chilean protesters attacked migrants and destroyed their property. This specific protest generated
a reply from the United Nations condemning violence against migrant communities as an erosion
of their human rights.13 Due to the high levels of exclusionary attitudes against foreigners in the
country, migrants have conducted their own protests asking for better working conditions, respect
for their human rights, and stopping racism.14 In 2021, many Haitians coming from Chile tried to
cross the US southern border. Different testimonials obtained by the New York Times show the
hostility towards them in Chile: “They tell us to go back home, that we are scum,” or, “We were
sold the ‘Chilean dream,’ but it turned out to be false.” All this evidence from surveys, interviews,
and media reports tells a story about discrimination and hostility against newcomers.
So far, we have provided evidence that an immigration shock increases both national pride and
national identity (hypothesis 1a) and increment crimes associated with hostility towards migrants,
which is confirmed by media, interviews, and survey evidence (hypothesis 1b). Therefore, we have
strong evidence to claim that Chileans are becoming more nationalistic (and not more patriotic)

13 Voz de America, "La ONU expresa preocupación por violencia contra venezolanos en Chile,"
February 2, 2022.
14 El Mostrador, "Inmigrantes marchan contra abuso laboral y discriminación," February 18,
2018.

21
after a large and sudden demographic change.

7 Support for Far-Right Parties

After showing that immigration shocks are making people more nationalistic, it is important to
evaluate whether these demographic changes are impacting the electoral performance of far-right
parties. In 2021, a far-right party competed nationally for the first time by presenting candidates
to the house, senate, and presidency. This party elected legislators in both chambers, and their
presidential candidate was the most-voted candidate in the election’s first round.
Because we have only one time period for the outcome since that party emerged in 2019, it
is not possible to implement a difference-in-differences design such as in the previous analysis.
Consequently, we use a different empirical strategy in this section. We rely on advances in optimal
matching and mathematical programming to construct a matched sample with two key charac-
teristics: (1) that the matched exposed and matched control groups are similar in key observed
characteristics, and (2) that the matched and unmatched samples are also similar in the same key
observed characteristics.
While traditional matching techniques such as propensity score matching do not guarantee co-
variate balance (King and Nielsen, 2019), we rely on cardinality matching where the researcher
defines the tolerances for imbalances beforehand, and then the largest matched sample that meets
those tolerances is found (Zubizarreta et al., 2014; Visconti and Zubizarreta, 2018). Also, it allows
constructing a representative matched sample where the matched sample looks similar to the un-
matched sample (Bennett et al., 2019; Kuffuor et al., 2022). This last point is particularly relevant
since a matched sample can be quite different from the unmatched sample since the commonly
optimized feature is covariate balance between exposed and control groups.
Regarding covariate balance, we define the standardized differences between the matched ex-
posed and control groups to be no greater than 0.2 standard deviations for all the covariates in
our study, a threshold commonly used in the literature to illustrate covariate balance (Silber et al.,

22
2013). As a result, our exposed and control groups will be similar between them in addition to
being similar to the unmatched sample. We use a mean balance constraint for all the pretreatment
covariates.15 We use municipalities as the level of analysis here. We utilize the vote share of the
center-right and center-left-wing candidates in the 2013 presidential election (the two candidates
who made it into the second round), turnout in 2013, the region’s number to constraint by geo-
graphic location, the population in 2011, and poverty in 2011. All the covariates are pretreatment
since the exposure indicator is constructed using immigration data between 2014 and 2017.
Importantly, we acknowledge that matching is not an identification by itself (Sekhon, 2009;
Keele, 2015). However, this design allows us to implement a sensitivity analysis afterward to
evaluate whether the findings are robust to hidden biases of different magnitudes (Rosenbaum,
2010). Consequently, we believe these results are more robust than relying on techniques that
adjust on observables but do not check for sensitivity to unobservables.
Table 3 compares the averages for all the covariates between the exposed and control groups
before matching. It shows how both groups are very different regarding their observed character-
istics. In fact, all of the standardized differences are greater than 0.2 (we report these numbers in
the supplementary appendix).

Table 3: Average values of covariates before matching

Covariates Exposed group Control group
Center-right vote share 2013 0.22 0.26
Center-left vote share 2013 0.55 0.44
Turnout 2013 27,874 56,992
Region 8.48 9.79
Population 2011 74,977 151,048
Poverty 2011 18.32 13.59
Observations 189 87

Table 4 illustrates what happens after implementing cardinality matching to construct a rep-

15 All of these covariates are continuous, ordinal, or binary, so the mean balance constraint is a
meaningful requirement, which would not be the case if we included nominal covariates.

23
resentative matched sample. Now, the entire sample of eligible municipalities,16 the matched
exposed and the matched control have similar averages for the pretreatment covariates. In particu-
lar, all the differences between these three groups are lower than 0.2 standard deviation units (we
report these numbers in the supplementary appendix).

Table 4: Average values of covariates after matching

Covariates Entire sample Matched exposed Matched control
Center-right vote share 2013 0.24 0.24 0.24
Center-left vote share 2013 0.48 0.48 0.47
Turnout 2013 47,814 44,663 47,873
Region 9.38 9.50 9.25
Population 2011 127,069 138,712 120,935
Poverty 2011 15.08 15.55 14.52
Observations 276 33 33

Finally, to evaluate the impact of an immigration shock, we use a one-sided permutational t-test
in match pairs that incorporates a sensitivity analysis to hidden biases (Rosenbaum, 2015). The
outcome of interest is the vote share of the candidate representing the far-right party in the first
round of the 2021 election. Table 5 provides the point estimates, showing that an immigration
shock increases support for the far-right at the municipality level by six percentage points. For the
sensitivity analysis, the parameter Γ represents the odds of differential assignment to the immi-
gration shock due to an unobserved factor we call u. A Γ = 1 means that two municipalities with
the same observed characteristics have the same probability of getting exposed to an immigration
shock (i.e., there are no hidden biases). A Γ = 2.0 means that two municipalities with the same
observed characteristics have a different probability of getting exposed to an immigration shock; in
particular, one is two times more likely than the other to get exposed (i.e., because of the existence
of hidden biases). Therefore, since the p-values are still significant for a Γ = 2.1, we have strong
evidence that our results are robust to large hidden biases explained by failing to adjust for some
unobserved covariate u.17
16 Same 276 municipalities used in the previous analyses.
17 Results stopped being significant at a 0.05 level for a Γ = 2.2.

24
 Table 5: Effect of immigration shocks on the far-right vote share in 2021

Point estimate 0.060

p-value (Γ=1.0) 0.002
p-value (Γ=2.1) 0.044

8 Survey Evidence from Eight South American Countries

Are these results a consequence of the particularities of the Chilean case? In an attempt to in-
crease the external validity of this study, we test part of the argument using survey data from eight
South American countries. We use the World Values survey since it includes a proxy of nationalis-
tic attitudes. The question used as the outcome is: "Willingness to fight for one’s country," a binary
variable that has been already connected with nationalistic sentiments and national identity by the
literature (Shulman and Bloom, 2014; Torres, 2020; Kim, 2020). However, we acknowledge that
this question is different from what we used in the main analysis, so it should be interpreted with
that distinction in mind.
A limitation of a cross-case study is that we do not have access to high-quality administrative
data to measure demographic changes, as is the case in Chile. Therefore, we evaluate respondents’
nationalistic attitudes before and after Venezuela’s socioeconomic and political collapse in 2015.
This crisis generated the largest wave of regional migration in Latin American history. While five
million Venezuelans left their country looking for safer and more prosperous places, a vast majority
of them migrated to countries located in South America (Vega-Mendez and Visconti, 2021). We
use all the South American countries with survey data availability: Argentina, Bolivia, Brazil,
Chile, Colombia, Ecuador, Peru, and Uruguay.
We use two waves before the collapse of Venezuela (2013 and 2014) and two waves after (2017
and 2018) to estimate the impact of this demographic change. We generate a binary indicator of
whether the survey was implemented after 2015, which is the coefficient of interest. Since this
is not panel data, we cannot implement a dynamic difference-in-differences, such as in the main

25
analysis. Therefore, we use six different models to estimate the effects of the Venezuelan crisis
using: i) controls18 , ii) country-fixed effects with controls, iii) year-fixed effects, iv) year-fixed
effects with controls, v) country- and year-fixed effects, and vi) country- and year-fixed effects
with controls. Additionally, we implement an interrupted time series analysis in the supplementary
appendix.
Figure 6 provides the impact of the 2015-2016 Venezuela collapse on willingness to fight for
one’s country. We find a positive and significant impact for all our different estimations. For
example, when using the country-fixed effect and controls (i.e., the smallest effect we found),
willingness to fight for your country increased by four percentage points after 2015 (95% CI:
[0.02, 007]). We report all the other point estimates and confidence intervals in the supplementary
appendix.

Country FE

Country FE + controls

Year FE

Year FE + controls

Country and Year FE

Country and Year FE + controls

−0.3 −0.2 −0.1 0.0 0.1 0.2 0.3

Figure 6: Effect of 2015-2016 Venezuelan crisis on willingness to fight for your country. Results
for six different types of analyses. 13670 observations.

18 Subjects’ characteristics that should be affected by exposure to migration, such as education,

gender, and age.

26
9 Conclusions

Nationalistic sentiments are not generated spontaneously, and strong national loyalties might
contribute to exclusionary attitudes and behaviors against foreigners and out-group members. In
extreme cases, some groups have even pushed for ideas such as ethnic democracies (Bertrand
and Haklai, 2013), where the goal is that native residents establish ownership over democratic
institutions.
Far-right parties have found fertile soil to flourish in the context of unprecedented demographic
changes. Consistent evidence shows a connection between demographic changes and the rise of
far-right political alternatives. Traditional explanations have argued that economic grievances (i.e.,
labor market competition or fiscal burdens) and cultural concerns (i.e., conflict with out-groups or
cultural backlash) explain why native residents become more likely to vote for radical right-wing
parties. However, in this paper, we hold that this story could be more complex than economic
and cultural grievances. In particular, if immigration shocks make people more nationalistic even
before the emergence of far-right parties, then this type of political organization might find the
ideal scenario to outperform other parties.
Far-right parties can promote a type of nationalism that is not always compatible with demo-
cratic values such as individual rights and tolerance (Golder, 2016). In the case of the US, eth-
nocultural conceptions of identity have been associated with nativism and preferences for cultural
homogeneity, which translates to support for restrictions on immigration, language, and citizenship
(Wright, 2011; Schildkraut, 2014). Similarly, studies using survey data from Europe have shown
that “ethnic” conceptions of the nation are associated with radical right voting (Lubbers and Coen-
ders, 2017), an increase in perceived threats for immigration, and preferences for more restrictive
migration policies (Stockemer et al., 2021; van der Brug and Harteveld, 2021).
To study the impact of immigration on nationalistic attitudes, we use rich administrative data
with all the visa requests made at the local level for a four-year period, which allows us to construct
an accurate estimate of immigration in Chile. Then, we use three-wave panel data that asks a

27
series of questions about nationalistic attitudes, allowing for the construction of the outcomes of
interest. Using a dynamic difference-in-differences design, we show that people changed their
political attitudes before the emergence of a far-right party. In particular, native-born residents
have become more likely to be proud of and identify with Chile. Additionally, we use crime rates
at the municipality level and find that immigration shocks are increasing crimes that are associated
with hostility towards migrants but not other types of crimes. Also, immigration shocks increased
support for a far-right party when this competed for the first time in the 2021 presidential election.
Demographic changes can make native-born residents more likely to support ideas pushed by
radical or extreme right-wing parties. These findings have important political implications when
trying to understand identity formation and the emergence of far-right parties in recent years. This
is particularly relevant to the extent that the same parties could promote policies that deteriorate
democratic norms and lead to prejudice and discrimination against newcomers.

References

Abrajano, M. and Z. L. Hajnal (2015). White Backlash. Princeton University Press.

Aedo, A. (2017). Encarnando (in) seguridad. orden policial y política de la presencia en la frontera
norte de chile. Antípoda. Revista de antropología y arqueología (29).

Alrababa’h, A., A. Dillon, S. Williamson, J. Hainmueller, D. Hangartner, and J. Weinstein (2021).

Attitudes toward migrants in a highly impacted economy: Evidence from the syrian refugee
crisis in jordan. Comparative Political Studies 54(1), 33–76.

Andersson, H., S. H. Dehdari, et al. (2020). Workplace contact and support for anti-immigration
parties. url: https://econpapers. repec. org/paper/crmwpaper/2006. ht m.

Arellano Calderón, D. A. (2022). Crímenes de odio en chile: una aproximación a la realidad

nacional desde los derechos humanos.

Arostegui, M. (2018). Waves of migrants head to chile, putting a strain on services. VOA.

28
Arzheimer, K. and E. Carter (2006). Political opportunity structures and right-wing extremist party
success. European Journal of Political Research 45(3), 419–443.

Audi, R. (2009). Nationalism, patriotism, and cosmopolitanism in an age of globalization. The

Journal of Ethics 13, 365–381.

Bazzi, S. and C. Blattman (2014). Economic shocks and conflict: Evidence from commodity
prices. American Economic Journal: Macroeconomics 6(4), 1–38.

Bellolio, A. and G. Valdés (2020). Gestión de la migración en el siglo xxi. el caso de chile.

Bennett, M., J. P. Vielma, and J. R. Zubizarreta (2019). Building representative matched samples
with multi-valued treatments in large observational studies. Working paper, Harvard University.

Bertrand, J. and O. Haklai (2013). Democratization and Ethnic Minorities: Conflict Or Compro-
mise? Routledge.

Bremmer, I. (2017). The wave to come. Time.

Brubaker, R. (2009). Ethnicity, race, and nationalism. Annual Review of Sociology 35, 21–42.

Callaway, B. and P. H. Sant’Anna (2020). Difference-in-differences with multiple time periods.

Journal of Econometrics.

Callaway, B. and P. H. Sant’Anna (2021). Difference-in-differences with multiple time periods.

Journal of Econometrics 225(2), 200–230.

Callaway, B. and P. Sant’Anna (2020). did: Difference in differences. R package version 2(0).

Campbell, A., P. E. Converse, W. E. Miller, and D. E. Stokes (1980). The american voter. Univer-
sity of Chicago Press.

Carreras, M. and G. Visconti (2022). Who pays for crime? criminal violence, right-wing incum-
bents, and electoral accountability in latin america. Electoral studies 79, 102522.

29
CENEM (2018). Haitianos en chile: Integración laboral, social y cultural. Centro Nacional de
Estudios Migratorios, Universidad de Talca.

Corstange, D. and E. A. York (2018). Sectarian framing in the syrian civil war. American Journal
of Political Science 62(2), 441–455.

Dancygier, R. M. and D. D. Laitin (2014). Immigration into europe: Economic discrimination,

violence, and public policy. Annual Review of Political Science 17, 43–64.

de Figueiredo Jr, R. J. and Z. Elkins (2003). Are patriots bigots? an inquiry into the vices of
in-group pride. American Journal of Political Science 47(1), 171–188.

Dehdari, S. H. (2019). Economic distress and support for radical right parties-evidence from
sweden. Available at SSRN 3160480.

Dinas, E., K. Matakos, D. Xefteris, and D. Hangartner (2019). Waking up the golden dawn:
does exposure to the refugee crisis increase support for extreme-right parties? Political analy-
sis 27(2), 244–254.

Dipoppa, G., G. Grossman, and S. Zonszein (2023). Locked down, lashing out: Covid-19 effects
on asian hate crimes in italy. The Journal of Politics 85(2), 389–404.

Doña Reveco, C. (2018). Amid record numbers of arrivals, chile turns rightward on immigration.
Migration Policy Institute 17.

Druckman, D. (1994). Nationalism, patriotism, and group loyalty: A social psychological perspec-
tive. Mershon international studies review 38(Supplement_1), 43–68.

Dunne, T. C. (2018). Friend or foe? a reversal of ingroup bias. Group Decision and Negotia-
tion 27(4), 593–610.

Escobar, C., R. Arana, and J. A. McCann (2015). Expatriate voting and migrants’ place of res-
idence: Explaining transnational participation in colombian elections. Migration Studies 3(1),
1–31.

30
Esnard, A.-M. and A. Sapat (2011). Disasters, diasporas and host communities: Insights in the
aftermath of the haiti earthquake. Journal of Disaster Research 6(3).

Evans, G. and R. Andersen (2006). The political conditioning of economic perceptions. Journal
of Politics 68(1), 194–207.

Fernandez, J. (2017). Las dificultades de los extranjeros para regularizar su situación migratoria.
Technical report, Diario el Dia.

Gijsberts, M. and L. Hagendoorn (2017). Nationalism and exclusion of migrants: Cross-national

comparisons. Taylor & Francis.

Golder, M. (2016). Far right parties in europe. Annual Review of Political Science 19, 477–497.

Green, E. G., O. Sarrasin, N. Fasel, and C. Staerklé (2011). Nationalism and patriotism as pre-
dictors of immigration attitudes in switzerland: A municipality-level analysis. Swiss Political
Science Review 17(4), 369–393.

Hainmueller, J. and M. J. Hiscox (2010). Attitudes toward highly skilled and low-skilled immigra-
tion: Evidence from a survey experiment. American political science review, 61–84.

Hainmueller, J. and D. J. Hopkins (2014). Public attitudes toward immigration. Annual review of
political science 17, 225–249.

Hangartner, D., E. Dinas, M. Marbach, K. Matakos, and D. Xefteris (2019). Does exposure to the
refugee crisis make natives more hostile? American Political Science Review 113(2), 442–455.

Holland, A. C., M. Peters, and Y.-Y. Zhou (2021). Left out: How political ideology affects support
for migrants in colombia. Available at SSRN 3803052.

Holland, A. C. and M. E. Peters (2020). Explaining migration timing: political information and
opportunities. International Organization 74(3), 560–583.

31
Huddy, L., A. Del Ponte, and C. Davies (2021). Nationalism, patriotism, and support for the
european union. Political Psychology 42(6), 995–1017.

Igarashi, A. (2021). Hate begets hate: anti-refugee violence increases anti-refugee attitudes in
germany. Ethnic and Racial Studies 44(11), 1914–1934.

Inglehart, R. F. and P. Norris (2016). Trump, brexit, and the rise of populism: Economic have-nots
and cultural backlash.

Jeong, H. O. (2013). Do national feelings influence public attitudes towards immigration? Journal
of Ethnic and Migration Studies 39(9), 1461–1477.

Kahneman, D. and A. Tversky (2013). Prospect theory: An analysis of decision under risk. In
Handbook of the fundamentals of financial decision making: Part I, pp. 99–127. World Scien-
tific.

Keele, L. (2015). The statistics of causal inference: A view from political methodology. Political
Analysis.

Kim, N. K. (2020). Territorial disputes and individual willingness to fight. Journal of Peace
Research 57(3), 406–421.

King, G. and R. Nielsen (2019). Why propensity scores should not be used for matching. Political
Analysis 27(4), 435–454.

Kuffuor, O., G. Visconti, and K. M. Young (2022). Constructing generalizable geographic natural
experiments. Research & Politics 9(3), 20531680221113763.

Leyens, J.-P., B. Cortes, S. Demoulin, J. F. Dovidio, S. T. Fiske, R. Gaunt, M.-P. Paladino,

A. Rodriguez-Perez, R. Rodriguez-Torres, and J. Vaes (2003). Emotional prejudice, essen-
tialism, and nationalism the 2002 tajfel lecture. European Journal of Social Psychology 33(6),
703–717.

32
López, M. Á. and J. Baeza (2011). Las elecciones chilenas de 2009-10:¿ se derechizó el país?
América Latina: política y elecciones del bicentenario (2009-2010), 277–302.

López, M. Á., N. Miranda, and P. A. Valenzuela-Gutiérrez (2013). Estimando el espacio político

del cono sur y brasil: las elecciones presidenciales en el eje izquierda-derecha. Postdata 18(2),
403–442.

Lubbers, M. and M. Coenders (2017). Nationalistic attitudes and voting for the radical right in
europe. European Union Politics 18(1), 98–118.

Lubbers, M., M. Gijsberts, and P. Scheepers (2002). Extreme right-wing voting in western europe.
European Journal of Political Research 41(3), 345–378.

Lucassen, G. and M. Lubbers (2012). Who fears what? explaining far-right-wing preference in
europe by distinguishing perceived cultural and economic ethnic threats. Comparative Political
Studies 45(5), 547–574.

Luna, J. P. and C. Rovira Kaltwasser (2021). A la derecha de la derecha:¿ hay espacio para una
fuerza populista radical de derecha. Fundación Friedrich Ebert: Montevideo.

Luttig, M. D. (2020). Reconsidering the relationship between authoritarianism and republican

support in 2016 and beyond.

Maillet, A. (2013). La construcción política de los mercados. variedades de neoliberalismo en el

chile post-dictadura. Santiago/Paris: PUC Santiago/SciencesPo.

Malone, M. F. T. (2019). Fearing the “nicas”: Perceptions of immigrants and policy choices in
costa rica. Latin American politics and society 61(1), 1–28.

Marz˛ecki, R. (2020). Constructive emotions? patriotism as a predictor of civic activity in poland.

Italian Political Science Review/Rivista Italiana di Scienza Politica 50(1), 33–51.

33
McCann, J. A. and M. Jones-Correa (2021). Political incorporation, interrupted? the effects of
loss and distress following the 2016 us elections on the civic attitudes and involvement of latino
immigrants. Political Behavior, 1–21.

Meléndez, C., C. Rovira Kaltwasser, and J. Sajuria (2021). Chile 2020: pandemia y plebiscito
constitucional. Revista de ciencia política (Santiago) (AHEAD).

Mudde, C. (2007). Populist radical right parties in Europe. Cambridge: Cambridge University
Press.

Newman, B. J. and Y. Velez (2014). Group size versus change? assessing americans’ perception
of local immigration. Political Research Quarterly 67(2), 293–303.

Osborne, D., P. Milojev, and C. G. Sibley (2017). Authoritarianism and national identity: Examin-
ing the longitudinal effects of sdo and rwa on nationalism and patriotism. Personality and Social
Psychology Bulletin 43(8), 1086–1099.

Pedersen, A., J. Attwell, and D. Heveli (2005). Prediction of negative attitudes toward australian
asylum seekers: False beliefs, nationalism, and self-esteem. Australian Journal of psychol-
ogy 57(3), 148–160.

Reny, T. T., L. Collingwood, and A. A. Valenzuela (2019). Vote switching in the 2016 election:
How racial and immigration attitudes, not economics, explain shifts in white voting. Public
Opinion Quarterly 83(1), 91–113.

Rojas Pedemonte, N., N. Amode, and J. V. Rencoret (2015). Racismo y matrices de “inclusión” de
la migración haitiana en chile: elementos conceptuales y contextuales para la discusión. Polis.
Revista Latinoamericana (42).

Rosenbaum, P. R. (2010). Design of Observational Studies. Springer.

Rosenbaum, P. R. (2015). Two r packages for sensitivity analysis in observational studies. Observ.
Stud 1, 1–17.

34
Rovira Kaltwasser, C. (2022). Rethinking the right in latin america. Latin American Politics and
Society, 1–11.

Rydgren, J. (2006). From tax populism to ethnic nationalism: Radical right-wing populism in
Sweden. Berghahn Books.

Schaub, M., J. Gereke, and D. Baldassarri (2021). Strangers in hostile lands: Exposure to refugees
and right-wing support in germany’s eastern regions. Comparative Political Studies 54(3-4),
686–717.

Schildkraut, D. J. (2014). Boundaries of american identity: Evolving understandings of “us”.

Annual review of political science 17, 441–460.

Sekhon, J. S. (2009). Opiates for the matches: Matching methods for causal inference. Annual
Review of Political Science 12, 487–508.

Severino, F. and G. Visconti (2023). Immigration shocks and unfounded concerns about crime:
Evidence from haitian migration to chile. Working paper.

Shulman, S. and S. Bloom (2014). Does nation building increase the strength of citizen loyalty?
theoretical and empirical investigations of ukrainian nationalism. Polity 46(3), 354–380.

Silber, J. H., P. R. Rosenbaum, R. N. Ross, O. Even-Shoshan, R. R. Kelz, M. D. Neuman, C. E.

Reinke, J. M. Ludwig, F. A. Kyle, and D. W. a. Bratzler (2013). Racial disparities in operative
procedure timethe influence of obesity. The Journal of the American Society of Anesthesiolo-
gists 119(1), 43–51.

Smith, A. E., M. Moseley, M. Layton, and M. Cohen (2021). Learning to rebel: When anti-
authoritarian parenting values are endogenous to politics. Presented at Southern Political Sci-
ence Association.

Stefoni, C. (2011). Ley y política migratoria en chile. la ambivalencia en la comprensión del

35
 migrante. La construcción social del sujeto migrante en América Latina. Prácticas, representa-
ciones y categorías, 79–110.

Stockemer, D., D. Halikiopoulou, and T. Vlandas (2021). ‘birds of a feather’? assessing the
prevalence of anti-immigration attitudes among the far right electorate. Journal of Ethnic and
Migration Studies 47(15), 3409–3436.

Tajfel, H. and J. C. Turner (1982). Social psychology of intergroup relations. Annual review of
psychology 33(1), 1–39.

Torres, D. J. (2020). The formation and solidarity of puerto rican national identity. Arcadia
University.

UNHCR (2017). Global trends - forced displacement in 2017. Technical report, The United
Nations High Commissioner for Refugees.

Valentino, N. A., S. N. Soroka, S. Iyengar, T. Aalberg, R. Duch, M. Fraile, K. S. Hahn, K. M.

Hansen, A. Harell, M. Helbling, et al. (2019). Economic and cultural drivers of immigrant
support worldwide. British Journal of Political Science 49(4), 1201–1226.

van der Brug, W. and E. Harteveld (2021). The conditional effects of the refugee crisis on immi-
gration attitudes and nationalism. European Union Politics 22(2), 227–247.

Vedoya, S. . (2017). Tres extranjeros al día, en promedio, han sido detenidos por ingreso clandes-
tino a chile en 2017. Technical report, La Tercera.

Vega-Mendez, C. and G. Visconti (2021). Does immigration increase concerns about the pro-
vision of public services? evidence from colombia. Revista Latinoamericana de Opinión
Pública 10(1), 79–103.

Visconti, G. (2021). Reevaluating the role of ideology in chile. Latin American Politics and
Society 63(2), 1–25.

36
Visconti, G. and J. Zubizarreta (2018). Handling limited overlap in observational studies with
cardinality matching. Observational Studies 4, 217–249.

Vrânceanu, A. and R. Lachat (2018). Do parties influence public opinion on immigration? evi-
dence from europe. Journal of Elections, Public Opinion and Parties, 1–21.

Wright, M. (2011). Policy regimes and normative conceptions of nationalism in mass public opin-
ion. Comparative Political Studies 44(5), 598–624.

Zanotti, L. and K. M. Roberts (2021). (aún) la excepción y no la regla: La derecha populista radical
en américa latina. Revista Uruguaya de Ciencia Política 30(1), 23–48.

Zhai, Y. and Z. Yan (2022). Political ideology, ingroup favoritism, and conspiratorial think-
ing: Patriotism, nationalism, and covid-19 conspiracy theories. Psychological Reports,
00332941221079727.

Zubizarreta, J. R., R. D. Paredes, and P. R. Rosenbaum (2014). Matching for balance, pairing for
heterogeneity in an observational study of the effectiveness of for-profit and not-for-profit high
schools in chile. The Annals of Applied Statistics 8(1), 204–231.

37