Abstract
This article presents a critical analysis of social scientific research from the past fifteen years on the criminalization of immigration in the United States. Our review reveals three central themes. First, although immigration law is considered federal civil law, immigration enforcement has become substantially intertwined with criminal law enforcement in ways that have broad-ranging and radiating effects on immigrant communities of color. Second, race plays an important bidirectional role in the increasing criminalization of immigration. Specifically, criminalizing immigration results in the racialization of certain immigrant groups as dangerous outsiders, and such racialization has the effect of sustaining and promoting the policies and practices that target or have a disproportionate impact on certain immigrant groups. Third, given the increased externalization of US immigration enforcement programs, the effects of the criminalization of immigration are not limited to the United States. Each of these dynamics operates across multiple sites and in various components of the immigration enforcement system to disproportionately affect immigrants racialized as non-White in the United States. We discuss the important gaps in research and policy implications that follow from our review.
This article presents a critical review and analysis of social scientific research from the past fifteen years on the criminalization of immigration in the United States. The criminalization of immigration broadly refers to the proliferating intersections of criminal law and immigration law, including the increasing role of criminal legal system actors in immigration enforcement and the expanded reliance on purportedly civil systems and structures that are deployed and experienced as punitive (Chacón 2021a; García Hernández 2021; Stumpf 2006). This development has resulted in a dramatic rise in federal criminal prosecution of immigration violations and a significant expansion of the type of criminal offenses that can trigger detention, deportation, and bars to reentry. Our review and analysis focus on the immigration scholarship that illuminates the direct and indirect ways that race shapes and sustains current US detention and deportation policies and practices. This scholarship also explores various ways that these policies and practices have generated and reproduced racial disparities and inequalities across time and space. An important insight from this body of research is the bidirectional role that race plays in the increasing criminalization of immigration. By bidirectional role, we mean that criminalizing immigration results in the racialization of certain immigrant groups as dangerous outsiders, and such racialization has the effect of sustaining and promoting the policies and practices that target or have a disproportionate impact on certain immigrant groups, particularly those from Mexico and Central America.
Understanding the criminalization of immigration and its impacts on communities of color is essential when studying the nature and consequences of the operation of the contemporary criminal justice system. As Sara Wakefield and Kristin Turney (2025, this issue) emphasize in their discussion of the rise of the carceral state in the United States, one defining feature of the modern criminal legal system is the spread of its carceral logic and practices into institutions that appear seemingly unrelated to the criminal justice system. Katherine Beckett and Naomi Murakawa (2012) have described the policy and institutional innovations that underlie this phenomenon as the “shadow carceral state.” The contemporary immigration enforcement system, including the asylum system, which relies on and perpetuates the criminalization of immigration, is a paradigmatic example of the shadow carceral state (see Ryo 2024). Therefore, a more complete and expansive understanding of racial disparities in criminal case processing and dispositions (see Bushway et al. 2025, this issue), policing (Dovidio and Solomon 2025, this issue), collateral consequences of contacts with the criminal justice system (Lee et al. 2025, this issue), and reentry post-incarceration (Nahra et al. 2025, this issue) requires attending to complex issues related to the criminalization of immigration.
Our focus on research from the past fifteen years is timely and important for at least two reasons. First, the past decade and a half are a critical period of study for our understanding of the role of race in shaping the experiences and outcomes for immigrants in our criminal justice and immigration enforcement systems. Immigration law and policy in the United States have long histories of discrimination and exclusion based on race (Chacón 2022a; Chin 1998; Molina 2014; Ngai 1999). Yet, legal scholars and social scientists have critiqued migration scholarship as failing to grapple adequately with the influence of race on immigration law (Johnson 2000). In response, recent scholarship across multiple disciplines has examined how contemporary US immigration laws and policies continue to subjugate immigrants of color—particularly Latinos—even as these laws and policies have become facially race-neutral or “color-blind” (Aranda and Vaquera 2015; Douglas et al. 2015; Kibria et al. 2013; Vazquez 2015).
Another reason for our focus on the past decade and half is that it represents a unique moment in US enforcement policies and practices. This time period is defined by three key trends that have become the focus of a growing body of empirical scholarship. The first is the rise of immigration federalism, which refers to the devolution of decision-making authority from federal to state and local governments in the construction and enforcement of immigration law (Chacón 2019; Gulasekaram and Ramakrishnan 2015; Rodriguez 2017). Under this system, state and local law enforcement actors and institutions have come to play a central and ever-expanding role in policing and incarcerating immigrants in service of federal detention and deportation policies (Ryo and Peacock 2020). This development has raised important questions about the potential for bias and discrimination in interactions between the police and immigrants. For example, some studies find that federal-local law enforcement partnerships such as the 287(g) program promoted the rise of anti-immigrant ideology among local law enforcement officers (Donato and Rodríguez 2014). Other studies explore the difficulty of proving racial profiling in this context and suggest that scholars focus on the problem of “racially discrepant” policing (Coleman and Kocher 2019).
The second important trend defining the past fifteen years is the increasing externalization of the US border. By externalization of the border, we mean the outsourcing of border enforcement to “sending and in transit countries at the request of the (more powerful) receiving states (e.g., the United States or the European Union) for the purpose of controlling the movement of potential migrants” (Menjívar 2014, 357). Although Mexico has been the key country through which the United States has expanded its border enforcement strategies beyond its territorial border (FitzGerald 2020; Frelick et al. 2016), in more recent years, US externalization policies and practices have extended into Guatemala, El Salvador, and farther south (Álvarez Velasco 2020; McGuire and Coutin 2013). These policies and practices have focused largely, and until quite recently almost exclusively, on deterring Central American migration, and the empirical studies that we review illuminate how US border externalization strategies shape, and are shaped by, the racialization of Central American migrants. By racialization, we mean the social process of attributing racial identities to a given group to exploit, exclude, and exert power over that group; this social process also entails efforts to construct racial categories to mobilize political responses to this exploitation and exclusion (see Gans 2017; Mora 2014; Omi and Winant 2014).
The third significant trend that characterizes this time period is the proliferation of immigration enforcement technologies that enable new forms of governmental surveillance, discipline, and control of immigrants. As Anil Kalhan (2014, 1) has noted, immigration control has rapidly become an “information-centered and technology-driven enterprise.” The type of technologies that have transformed the regime of immigration control in the past decade and a half include algorithmic risk assessment tools used in immigration custody determinations (Frankel 2023); electronic monitoring devices such as ankle bracelets that track immigrants released from immigration detention (Marouf 2017); the Enforcement Integrated Database (EID), a case management system shared by several agencies within the Department of Homeland Security (DHS) (Muñiz 2020); and E-Verify, a biometric system that identifies whether an immigrant is eligible to work legally in the United States (Goldstein and Alonso-Bejarano 2017). Studies of technological advances that facilitate immigration enforcement show that they are often an invisible but powerful form of racialized control that intensifies the social and economic precarity of immigrant communities.
This article will begin with a background section that provides a historical overview of the criminalization of immigration in the United States. We will then briefly contextualize the US case by outlining parallel trends and developments in other parts of the world, including Canada, Australia, and the European Union. Next, the article will proceed with our critical analysis of the existing empirical research on the criminalization of immigration by focusing on five key areas. The first three sections will review and analyze social scientific studies that investigate the origin and operation of each of the three key trends we have noted: immigration federalism, externalization of the border, and immigration enforcement technologies. Taken together, these trends have generated a dramatic rise in the criminal prosecution of immigration crimes in criminal courts, as well as the civil prosecution of immigration law violations in immigration courts. Both phenomena are characterized by significant racial disparities. Thus, the fourth section examines empirical scholarship that investigates racial disparities in criminal and immigration court processes and outcomes. In addition, the impact of these trends on families—particularly families in Latino communities—has been devastating. Accordingly, the fifth section considers empirical studies that examine the effects of immigration enforcement on families. We conclude each of these five sections by highlighting the key policy implications that follow from our review of empirical studies in each of these bodies of research. Although our discussion of the literature is organized into five distinct sections, we hasten to note that there are many intersections across these areas of research given that each relates to various dimensions of the broader enforcement landscape.
A few notes on terminology are in order before we proceed. First, we recognize that scholars use different terms, such as Latino/a, Latinx, and Hispanic, to describe individuals of Latin American origin. These terms have varying historical origins associated with shifts in the US government classification system, changes in self-identification preferences, and the rise of political activism by different groups (for a recent review, see Martínez and Gonzalez 2021). For consistency, we use the term Latinos throughout this article, except when we are discussing specific studies; in the latter instances, we follow the convention that the study authors use. For the purposes of discussing racial disparities and racial biases in this article, we treat Latinos as a non-White racialized minority.1 We hasten to note, however, the continuing debate about whether Latinos constitute an ethnic category or a racial category and the analytical implications of deploying these terms in research. Some scholars construe Latinos as a racial category to draw attention to the racialization process that results in their subordinated status in the US racial hierarchy (see, for example, Vélez and Peguero 2023). Other scholars argue that treating Latinos as a racial category can render groups such as Afro-Latinos invisible (see, for example, Katerí Hernández 2021) and obscure the continuing salience of skin tone stratification within Latino communities (see, for example, Monk 2021). Throughout this article, we underscore the importance of understanding both the racialization process and the ways that racial labels can homogenize diverse populations and further subordinate marginalized subgroups.
Second, we are mindful that individuals working in media, government, and academia, as well as members of the public, have used different terms with varying degrees of frequency to describe individuals without legal documentation in the United States. These terms have generated debate and controversy (see, for example, De Genova 2002; King 2021; Kopan 2018; Guskin 2013; Núñez 2013), as well scholarship on how the use of these terms might shape people’s attitudes toward immigrants and immigration (see, for example, Callister et al. 2021; Hoops and Braitman 2019; Merolla et al. 2013; Nelson and Davis-Wiley 2018). For the purposes of this article, we generally refer to non-US citizens lacking legal authorization to be in the United States as undocumented immigrants.2
Finally, we underscore the importance of recognizing the power of classification systems that generate and naturalize certain group labels. A growing number of scholars have highlighted the centrality of state-created categories in producing and reproducing inequality and stratification, and the ways that inequality and stratification in turn normalize and legitimate the existence of those same state-created categories (see, for example, Massey 2007; Menjívar 2023). Likewise, there can be a cyclical and mutually reinforcing dynamic between state-created categories and social-scientific inquiries. As Cecilia Menjívar (2024, 935) has argued, “Using state categories at face value to make our work legible to others, especially as we collect data, amplifies the power of state classifications. When we use state-created categories to collect and organize our data and then compare the profiles of the people the state classifies as undocumented, refugee, or ‘regular’ immigrants, we provide scientific evidence that the individuals the state sorts into those categories are, in fact, inherently different.” Thus, we emphasize the need to adopt a critical perspective on not only the group labels themselves but also the state-created categories that give rise to those group labels and how those categories might be shaping the scholarship that we review in this article.
BACKGROUND
To set the broader context for our review of empirical scholarship on the criminalization of immigration, we begin by briefly tracing the historical developments and contemporary trends relating to the increasing intersection of criminal law and immigration law that some scholars have labeled “crimmigration” (see, for example, Beckett and Evans 2015; Fan 2013; Mathis and Androff 2021; Stumpf 2006). The racialization of certain groups as “illegal” immigrants and criminal threats has been central to this process (Brown et al. 2018; Menjívar 2021). As Ruben Rumbaut and colleagues (2019, 476) note more generally, the criminalization of immigration is “rooted in a long history of racialized immigrant exclusion, containment, and disposal.” In particular, the racialization of Latinos, which has a long history in the United States (Baca Zinn and Wells 2023; Massey 2013), underpins the US government’s growing reliance on the criminal law apparatus to achieve its immigration-enforcement aims.
As mentioned earlier, one of the core substantive features of the criminalization of immigration is the shift to criminal prosecution of immigration law violations that used to be classified and addressed as civil violations. While unauthorized entry and unauthorized reentry have been classified as federal crimes since 1929 (García Hernández 2018), Congress significantly expanded the list of immigration-related violations that constituted federal crimes in the 1980s (Chacón 2021a; Medina 1997; Stumpf 2006). Among others, these violations included the hiring of undocumented immigrant workers and committing various acts of fraud in the immigration process. Congressional expansion of federal immigration crimes coincided with the dedication of significant federal resources toward the criminal prosecution of unauthorized entry and reentry, and procedural innovations designed to expand and streamline these prosecutions (Chacón 2009; Eagly 2010). These combined developments led to a dramatic rise in the number of cases involving immigration offenses in criminal court dockets across the United States. For example, the number of suspects prosecuted in a US district court with an immigration offense as the lead charge more than tripled from 1998 to 2018 (Motivans 2021). Between fiscal years 2018 and 2020, prosecution of immigration crimes constituted the largest share of all federal criminal prosecutions (Schmitt and Russell 2021).
The second core substantive feature of the criminalization of immigration is the expansion of the array of criminal offenses that triggers immigration detention and deportation. Scholars have traced this legal development to the late 1980s. The Anti-Drug Abuse Act of 1988 established a new category of federal crimes called “aggravated felonies” that could trigger immigration detention and deportation (Miller 2003). Beginning in 1990, Congress broadened the definition of aggravated felonies (Stumpf 2006). Over time, the list of “aggravated felonies” expanded so much that these offenses now need not even be “aggravated” nor “felonies” under criminal law (Morawetz 2000). The other key category of offenses for which an immigrant may be deported is “crimes involving moral turpitude” (CIMT). Prior to 1996, an immigrant was deportable for a single CIMT committed if, among other things, one year or more of confinement was actually imposed; this was changed in 1996 so that the mere possibility of a sentence of one year or more could render an immigrant deportable (Lerner 2021; Miller 2003). In sum, US immigration law and enforcement have become heavily focused on detaining and deporting immigrants with criminal records. In fiscal year 2008, immigrants without past criminal convictions accounted for 69 percent of those removed after an arrest by US Immigration and Customs Enforce (ICE); by fiscal year 2020, immigrants with past criminal convictions or pending criminal charges accounted for 92 percent of removals resulting from ICE arrests (Hausman 2022; US ICE 2016, 2020).
The criminalization of immigration has culminated in the development of what some scholars refer to as an “immigration industrial complex” (see, for example, Díaz 2012; Golash-Boza 2009). There are varying definitions of the immigration industrial complex, but the central idea is that the modern US immigration enforcement system is a massive carceral infrastructure driven by profit motive and capital accumulation through social control and the economic exploitation of racially subordinated groups (see, for example, Chacón 2022b; Doty and Wheatley 2013; Douglas and Sáenz 2013; Trujillo-Pagán 2014). Building on this theoretical framework, Emily Ryo and Chris Levesque (2025) argue that the US immigration detention system functions as a racialized wealth extraction process. They explain that the word racialized draws attention to how the costs of immigration detention are borne disproportionately by communities of color (particularly Latinos) that have been socially constructed as dangerous and undeserving; the word wealth highlights how the harms of immigration detention are multiplicative and compound over time across generations; and the word extraction underscores the violent and predatory nature of immigration detention. In this and other ways, race and racialization have played a central role in giving rise to, and shaping the consequences of, the criminalization of immigration.
Finally, although this article focuses on the US context, the criminalization of immigration is a global phenomenon that has taken hold in other key receiving countries as well, including members of the European Union, Canada, and Australia (Bosworth et al. 2018; FitzGerald 2020; Frelick et al. 2016). As in the United States, popular and state discourses in other receiving countries often connect immigration to “fears … based on racialized moral panics” and construct immigrants as criminals or security threats (Menjívar 2014, 356). For instance, although European research “tends to tie immigration enforcement and exclusion to a lack of formal citizenship rather than racism” (Armenta 2017, 83), other scholars contend that discussions of migration “crises” in Europe are inextricably tied up with a European racial project wherein race-neutral immigration policies hinge on racialized logics (De Genova 2016). For example, David FitzGerald and colleagues (2018, 356) note that “Muslims in Europe are increasingly treated as a quasi-ethnic group subject to widespread discrimination.” This and related issues in other parts of the world beyond the United States are outside the scope of our review, but they underscore the importance of conceptualizing the criminalization of immigration as a widespread social process that impacts the lives of millions of individuals on the move around the globe (see, for example, Menjívar 2014; Parmar 2018).
REVIEW OF EMPIRICAL SCHOLARSHIP
In this section, we offer a critical analysis of the existing empirical research on the criminalization of immigration. Our discussion is organized around the following five key themes: immigration federalism, externalization of the border, immigration enforcement technologies, racial disparities in criminal and immigration court processes and outcomes, and the effects of immigration enforcement on families.
Federalism, Interior Enforcement, and Policing
Although immigration law as codified in the Immigration and Nationality Act is federal law and within the exclusive purview of the federal government, in the past couple decades cities, counties, and states increasingly have become involved in regulating immigration and the lives of immigrants (Chacón 2019; Gulasekaram and Ramakrishnan 2015; Reich and Scott 2023). Scholars describe this sharing of responsibility over immigration-related matters by multiple tiers of government as immigration federalism (Motomura 1999; Varsanyi et al. 2012). Broadly defined, immigration federalism refers to “the engagement by national, state, and local governmental actors in immigration regulation” (Elias 2013, 707). Immigration federalism includes a diverse array of exclusionary and inclusionary rulemaking by state and local governments that seeks to govern immigrants’ ability to work and live in their jurisdictions. For example, many state and local jurisdictions across the United States have enacted laws and regulations that restrict or facilitate immigrants’ access to public benefits and services, education, housing, and employment (see Chavez and Provine 2009; Perreira and Pedroza 2019). These types of rulemaking aimed at immigrants’ social and economic integration (or exclusion) are an important aspect of immigration federalism, but they are beyond the scope of this section.3
This section focuses more narrowly on one specific dimension of immigration federalism that we refer to as immigration enforcement federalism. A defining feature of immigration enforcement federalism is that it relies on local criminal justice agencies and systems to facilitate and expand the arrest and removal of immigrants from the interior of the United States. The nature and extent of these types of partnerships vary. For example, some jurisdictions allow ICE access to their local jails to arrest immigrants, while others receive payment from the federal government to allow ICE to use their facilities to detain immigrants (Ryo and Peacock 2020). Our review examines two main categories of studies in the research on immigration enforcement federalism. The first category of studies seeks to uncover local jurisdictions’ motivations for participating in immigration enforcement or to identify the predictors of such participation. The second category of studies in this body of research evaluates the consequences of immigration enforcement federalism on policing and interactions with the police.
Some scholars have used the terms sanctuary and anti-sanctuary to describe subfederal jurisdictions’ engagement in immigration regulation, including immigration enforcement. Excluded from our review are studies that examine sanctuary policies or practices without a comparative analysis of cooperative subfederal policies or practices that enforce federal immigration law (see, for example, Amuedo-Dorantes et al. 2022; Lasch et al. 2018).4 The decision to limit our discussion in this way is informed by several considerations. The term sanctuary has different meanings and connotations, and its usage has been contested on various grounds (see Su 2024). This makes it difficult for us to establish a consistent standard for identifying and evaluating these studies. For example, some studies define sanctuary jurisdictions as those that have enacted a specific sanctuary legislation, while others define them as those that decline immigration detainers (Ascherio 2022). Furthermore, sanctuary policies and practices are not necessarily the flip side of our focus in this section—namely, cooperative subfederal policies and practices that seek to enforce federal immigration law. As Daniel Martínez and colleagues (2018, 2) have explained: “Although local communities adopt ‘sanctuary’ policies to limit cooperation with the federal government, the implementation of these policies does not imply that noncitizens are protected from federal immigration enforcement action. … In fact, many ‘sanctuary’ jurisdictions allow for local assistance and cooperation in the enforcement of federal immigration laws when individuals have been convicted of violent or serious felony offenses.”
Legal Context
We begin by providing a brief description of two policy initiatives illustrative of immigration enforcement federalism that have received a great deal of scholarly attention: the 287(g) program and Secure Communities. The 287(g) program has its basis in section 287(g) of the Immigration and Nationality Act, which Congress enacted to allow federal immigration enforcement agencies to directly delegate certain immigration enforcement functions to state and local government agencies (Kolker 2021). Currently, the three 287(g) models in operation are the jail enforcement model, the warrant service officer model, and the task force model.5 The jail enforcement model allows deputized officers to interrogate suspected noncitizens about their immigration status when they have been arrested on state or local charges and to place immigration detainers on those thought to be subject to removal.6 An immigration detainer is a request from ICE to a state or local law enforcement agency to notify ICE as to when a noncitizen will be released from criminal custody and to maintain custody of the noncitizen for an additional period of up to forty-eight hours to allow ICE to assume custody for removal purposes (US DHS 2022). The warrant service officer model allows ICE to train, certify, and authorize state and local law enforcement officers to execute ICE administrative warrants on noncitizens in their agency’s jail. Finally, the task force model allows local law enforcement officers “to enforce limited immigration authority with ICE oversight during their routine police duties” (US ICE 2025).
Secure Communities is another key cooperative immigration enforcement initiative. It was rolled out in 2008 during President George W. Bush’s administration as an information-sharing partnership that links an immigration background check to the standard criminal background check process. When state or local law enforcement agencies arrest or take individuals into custody, those agencies traditionally have shared the individual’s fingerprints with the Federal Bureau of Investigation (FBI) for the purposes of a criminal background check (US DHS 2022). Under Secure Communities, the FBI then automatically sends the fingerprints to DHS, allowing that agency to check against its immigration databases and issue immigration detainer requests to the relevant state or local agency (US DHS 2022). President Barack Obama’s administration replaced Secure Communities with Priorities Enforcement Program (PEP)—in part to address the growing concern that Secure Communities was eroding trust between local law enforcement and immigrant communities. In brief, PEP maintained the same data-sharing infrastructure as Secure Communities but established enforcement priorities that focused on noncitizens “convicted of significant criminal offenses or who otherwise pose a threat to public safety,” purportedly taking into account local enforcement priorities (US DHS 2015). President Donald Trump, however, reinstituted Secure Communities through Executive Order 13768 and re-expanded enforcement priorities to include all undocumented immigrants (Chavez et al. 2023). In January 2021, President Joe Biden revoked Executive Order 13768, but the practice of local law enforcement sharing biometric data on arrested individuals with DHS continues (Chavez et al. 2023).
Motivations for Participating in Immigration Enforcement
Why do some jurisdictions participate in immigration enforcement? Government officials have justified federal-local immigration enforcement partnerships as a public safety measure aimed at reducing crime by targeting “criminal aliens” (see Kandel 2016). Scholars have reasoned that if this public-safety rationale were, indeed, the motivation for these partnerships, then crime rates should be a significant predictor of where, when, and to what extent these partnerships are established. However, studies have not found empirical evidence in support of this expectation (see, for example, Chand and Schreckhise 2015; Cox and Miles 2013; Wong 2012). Thus, scholars have proposed and tested other possible motivations. There are two sets of hypotheses implicating racial bias and racial disparities that are of particular relevance for our discussion. The first is the racial threat hypothesis, and the second is what we refer to as the racial diversity hypothesis. In the rest of this section, we examine in turn empirical research testing each of these hypotheses.
Racial Threat Hypothesis
The racial threat hypothesis posits that as racial minority groups in a given community grow in size, the White majority will view them as an economic, social, political, or criminal threat and subject them to increased discrimination and social control (Feldmeyer and Cochran 2018). Given the racial threat hypothesis’s focus on social control, a long-standing body of research in criminology has examined whether racial composition as a proxy for racial threat is linked to wide-ranging criminal justice outcomes, such as the size of police force and expenditures, minority arrests, convictions, incarceration, sentence length, and death penalty decisions (see Feldmeyer and Cochran 2018). More recently, scholars have applied racial threat hypothesis to understand immigration enforcement federalism by conceptualizing federal-local partnerships on immigration enforcement as a form of social control of immigrants—in particular, of Latinos. Felicia Arriaga (2017, 420) goes even further to argue that federal-local immigration enforcement partnerships are “a form of racialized social control meant to affect all Latinos, regardless of immigration status, age, or gender” (emphasis added).
Overall, empirical findings of quantitative studies are mixed—and conflicting at times—as to whether, and in what ways, racial threat might be associated with subfederal jurisdictions’ participation in immigration enforcement. These diverging findings are difficult to square because the studies use different measures of racial threat to examine varying types of participation in immigration enforcement. For example, Emily Farris and Mirya Holman (2017) find that the percentage of Latino residents in a county is not a significant predictor of the number of immigration-status checks conducted by police officers in the county, nor of whether the county has a formal policy allowing police officers to conduct immigration-status checks in certain situations (see also Chavez and Provine 2009). On the other hand, Adam Cox and Thomas Miles (2013) find that percent Hispanic in a given county is a significant predictor of early activation of Secure Communities in that county. Similarly, Daniel Chand and William Schreckhise (2015) find that percent Hispanic is positively related to the number of deportations under Secure Communities. It is worth noting, however, the findings in Chand and Schreckhise (2015) are sensitive to different modeling strategies, as Jillian Jaeger (2016) has shown.
Other studies find that, contrary to the expectations of the racial threat hypothesis, policy and demographic composition are negatively related. For example, Tom Wong (2012) finds that an increase in the share of the Hispanic/Latino population reduces the likelihood that a county will enter into a 287(g) agreement. Likewise, Heather Creek and Stephen Yoder (2012) examine predictors of state decisions to enter into 287(g) agreements and find that the greater the changes in percent Hispanic, the lower the likelihood of a 287(g) agreement. This result leads Creek and Yoder to speculate that political elites are reluctant to antagonize a growing electorate of Hispanics.
What might explain these divergent findings on the racial threat hypothesis? Alexandra Filindra (2019) offers explanations that relate to both theoretical and methodological problems in this body of research. For example, Filindra (2019, 534) identifies one of the key issues as “ad hoc” theorizing: “In a context of polysematic indicators, researchers may be tempted to adapt their theory to fit their findings.” Juan Manuel Pedroza (2019) suggests that part of the explanation may relate to the faulty assumption that the relationship between policy and demographic composition is linear rather than curvilinear. According to Pedroza (2019, 627), an important part of the racial threat hypothesis is that “once minorities comprise a formidable enough presence, their influence over legislative priorities translates into less restrictionist or welcoming policies.” This idea finds support in a study by Emily Ryo and Ian Peacock (2020) that investigates why local jurisdictions enter into Inter-Governmental Service Agreements (IGSAs) with the federal government to receive payment for the use of their facilities to detain immigrants for ICE. Ryo and Peacock find that percent Latino is positively associated with the likelihood of counties entering into IGSAs but that this relationship attenuates as percent Latino continues to increase after a certain threshold point. At the end of this section, we suggest several other possible explanations for the divergent findings in this body of research and discuss the importance of investigating those possibilities from a policy standpoint.
Racial Diversity Hypothesis
The racial diversity hypothesis proposes another way in which race might influence subfederal jurisdictions’ involvement with immigration enforcement. Unlike the racial threat hypothesis, which focuses on the racial composition of immigrants residing in a given community, the racial diversity hypothesis focuses on the racial composition of authorities in power in government agencies engaged in decision-making about immigration enforcement. The racial diversity hypothesis emerges from research on representative democracy, which suggests that minority administrators in office or bureaucracies are more likely to implement practices or use procedures that benefit minority groups (Bishu and Kennedy 2020; Sowa and Selden 2003). A later wave of research on representative democracy has suggested that for this dynamic to be triggered or become detectable, there must be “a critical mass” of the relevant personnel within an organization that embody the characteristics of the minority group (Bishu and Kennedy 2020).
Empirical applications and tests of the racial diversity hypothesis to understand federal-local immigration enforcement are scarce. However, the few studies that exist do provide some empirical support for this hypothesis.7 Paul Lewis and colleagues (2013) analyze data from a national survey of police chiefs, and they find that the presence of a Hispanic police chief is associated with significantly less aggressive immigration enforcement by local police. Emily Farris and Mirya Holman (2017) find in their survey of over 550 sheriffs that the presence of a Latino sheriff is not associated with the county’s immigration enforcement practices and policies but that the percentage of Asian or Hispanic deputies is associated with a lower likelihood of counties adopting a policy that advises policy officers on conducting immigration-status checks. Similarly, Daniel Chand (2020) analyzes survey data from over 610 county sheriff offices and finds that offices with larger percentages of Hispanic officers—and to a lesser degree, percentages of Black officers—produce fewer removals and fewer submissions to ICE for immigration background checks. Notably, each of these studies controls for the size of the Hispanic population in estimating the effect of the racial identity of policy chiefs and police officers, which provides assurance that these race measures are not mere proxies for the strength of the minority electorate.
Effects of Federal-Local Partnerships
What are the consequences of immigration enforcement federalism on policing and interactions with the police? We start by highlighting the key study findings on the “intended” consequences. Briefly summarized, studies that have evaluated whether and to what extent federal-local partnerships have achieved, or failed to achieve, their goal of reducing crime and enhancing public safety generally find either no relationship or only a limited one. For example, Elina Treyger and colleagues (2014) find in their city-level analysis of crime rates that the activation of Secure Communities did not result in significant changes in crime rates. Songman Kang and B. K. Song (2021) show that the activation of Secure Communities resulted in a reduction in local crime only when neighboring jurisdictions also activated it. Kang and Song theorize that whereas the activation of Secure Communities in one county may cause some of the local crime to be merely displaced to nonactivated counties nearby, joint activation by neighboring counties may result in reductions in the overall crime in the region by triggering immigrants who are at high risk of offending to leave the region altogether.
In terms of the “unintended consequences” of immigration enforcement federalism, empirical evidence is more consistent and robust. As used by scholars in this area of research, the term unintended does not imply that the effects were unforeseen or unforeseeable; instead, the term simply refers to consequences that are beyond the stated policy goals. The rest of this section focuses on two types of unintended consequences that relate to policing: racially discrepant policing and racial disparities in crime reporting and victimization. By racially discrepant policing, we mean contacts between law enforcement and civilians that have disproportionate impacts on racial minorities. Following Mat Coleman and Austin Kocher (2019), we use this term to distinguish it from racial profiling, another term that is sometimes used in this literature but with varied meanings. The use of the term racial profiling can create confusion insofar as racial profiling implies or requires the existence of discriminatory intent or racial animus, but most studies are not designed to—and do not claim to—identify or measure such intent or animus. Racial profiling, as defined in this narrow way, fails to capture complex dynamics of racialized policing on the ground (see, for example, Armenta 2017; Chacón and Coutin 2018; Taylor et al. 2014).
Racially Discrepant Policing
In her qualitative study of policing of Latinos in Nashville, Tennessee, Amada Armenta (2017, 82) concludes that “most [law enforcement officers] do not see themselves as active participants in immigrant removal and they explain their behavior through a colorblind ideology.” Yet, investigations by government agencies and nongovernmental organizations, as well as lawsuits filed by immigrant rights organization and Latino advocacy groups have documented the rise of racially biased policing that subjects immigrants—particularly Latinos (both immigrant and U.S.-born) to heightened surveillance, stops, and arrests (Chacón and Coutin 2018; Provine and Sanchez 2011; Provine et al. 2016). An emerging body of scholarship and journalistic accounts (see Russakoff and Sontag 2018) also suggests that federal-local partnerships in immigration enforcement can lead to racially discrepant policing, especially given the wide discretion that police officers have in making stops and arrests. We focus our discussion on research concerning two related, but distinct, inquiries on racially discrepant policing of immigrant communities. The first line of inquiry centers on perceptions and experiences of Latinos concerning racially biased policing in light of local law enforcement’s involvement in immigration enforcement. The second line of inquiry focuses on changes in the behavior of police officers considering their increasing involvement in immigration enforcement. Prominent in this line of inquiry are questions about whether Latinos are targeted by the police for stops and arrests, and whether policing practices have a racially disparate negative impact on Latinos.
First, we begin by considering empirical research that centers on perceptions of racialized policing. Empirical studies show that Latinos—both immigrant and U.S.-born groups—report that they are targeted by the police or that they feel they are stopped more frequently by the police in different contexts. For example, Guadalupe Vidales and colleagues (2009) examine the impact of a highly publicized controversy in Costa Mesa, California, about whether and how to involve Costa Mesa police in enforcing immigration law. They use a survey design that captured the residents’ views of, and interactions with, the police both before and after the public controversy. In this study, Latinos reported being stopped by the police at a significantly higher rate after the public controversy compared to before the controversy, whereas no such increase was reported by non-Latinos. Using a cross-sectional survey design, Renee Rocha and colleagues (2015) find similar results. Their analysis of a national study of Latinos shows that a quarter of the respondents reported knowing someone who was detained or deported and that this report is more common in areas where the Secure Communities removal rate is high. Qualitative studies based on interviews with Latinos also highlight their firsthand accounts of how federal-local partnerships in immigration enforcement have promoted racially biased stops and arrests (see, for example, Aranda and Vaquera 2015; Lemon et al. 2024).
Next, turning to research on the behavior of police officers given their increasing involvement in immigration enforcement, we find that the study results are more complex. A study by Katharine Donato and Leslie Rodríguez (2014, 1697) is unique in that they analyze data on how police officers justify arrests. The study examines changes in how police officers in Nashville, Tennessee, reported reasons given for arrests of immigrants and nonimmigrants after the implementation of 287(g) agreements, and it finds that markers of foreignness (for example, country of origin, language use, and legal status) have taken on new salience in these arrest reports. In brief, Donato and Rodríguez (2014) conclude that the implementation of 287(g) agreements fostered “an ideology of anti-foreignness” among local law enforcement officers.
In addition to exploring law enforcement officers’ use of language to describe their actions in the wake of federal-local partnerships in immigration enforcement, scholars have also examined how such partnerships impact patterns of stops and arrests at the aggregate level. For example, Ashley Muchow (2024) finds that intergovernmental service agreements (IGSAs) that counties adopt to lease ICE jail space for immigration detention increased rates of Hispanic arrests but not those of White or Black residents. Muchow also examined whether post-IGSA increase in Hispanic arrest rates might be conditional on police officers’ perceptions of minority threat, as measured by shares of Hispanic population shares. Her analysis shows that the post-IGSA increase in Hispanic arrest rates is more pronounced in counties with moderate shares of Hispanic population (as compared to counties with lower shares of Hispanic population), and that this relationship becomes attenuated for counties with the highest Hispanic shares. According to Muchow (2024, 1–2), this curvilinear relationship suggests that IGSAs may “trigger minority threat concerns that increase arrests” of Hispanics.
Two additional studies are worth highlighting, in part to underscore the complexity of disentangling changes in the behavior of immigrants at risk of stops and arrests, and changes in the behavior of the police. Michael Coon (2017) investigates the effects of the implementation of the 287(g) agreement in Frederick County, Maryland, on the arrests of Hispanics. Examining changes in arrests made by both the Frederick County Police Department and the Frederick County Sheriff’s Office, Coon (2017) finds that Hispanic arrests fell, whereas no changes occurred in the arrests of Whites, and the arrests of Blacks increased. Coon suggests that the drop in Hispanic arrests is likely due to the “chill effect,” whereby the Hispanic community avoided interactions with law enforcement. Coon (2017), however, also finds that the implementation of a 287(g) agreement led to a significantly higher number of arrests of Hispanics by the Sherriff’s Office than would have occurred otherwise in its absence. Likewise, Britte van Tiem (2023) explores the impact of 287(g) agreements on traffic stops and arrests in North Carolina. Similar to Coon (2017), van Tiem finds that stops of Hispanic drivers fell in the wake of 287(g) agreements and shows that the most likely explanation for this decline is the decreased road use by Hispanics. Unlike Coon (2017), however, van Tiem finds no evidence that police officers increased pretextual stops and arrests of Hispanic drivers.
Racial Disparities in Crime Reporting and Victimization
To detect and sanction criminal behavior, law enforcement officials must rely on community members’ willingness to report crime and to cooperate with their investigations. Heightened fears of deportation and distrust in the police can erode this willingness to report crime and cooperate with the police (Becerra et al. 2017; Messing et al. 2015). A growing body of empirical research indicates that federal-local partnerships in immigration enforcement have both types of attitudinal and behavioral impacts, especially in Latino communities. These changes, in turn, render Latinos more vulnerable to victimization. Thus, studies have concluded that, despite the partnerships’ aims of promoting community safety, their net effect is to make communities less safe, particularly Latino communities.
Some studies have focused exclusively on Latino communities. Mai Nguyen and Hannah Gill (2016), in a qualitative study that relied in part on interviews with Hispanic residents in two communities in North Carolina, examine the impact of the implementation of 287(g) agreements. They find that these federal-local partnerships led to an increased fear and distrust of law enforcement among the Latino immigrants that compromised public safety and security. Nik Theodore and Robert Habans (2016) conducted a survey of over two thousand Latinos in four US counties regarding their perceptions of, and interactions with, the police. The survey respondents included Latino residents with varying legal status: authorized immigrants, undocumented immigrants, and nonimmigrants. The study finds that both authorized and undocumented Latino immigrants are less likely than Latinos who are not immigrants to report crime and more likely to experience social isolation in light of the increasing police involvement in immigration enforcement; and the effects are greater for undocumented immigrants than for authorized immigrants. However, these results vary by context (Menjívar et al. 2018) and social position, such as gender (Simmons et al. 2021). A recent study by Tom Wong and colleagues (2021) is notable for its experimental design. In a survey experiment using a probability-based sample of undocumented immigrants in San Diego County, the investigators find that when the respondents were exposed to a prompt that made the federal-local partnership salient, there was a significant reduction in the respondents’ level of trust in the police.
The studies reviewed above are informative of how federal-local partnerships might affect Latino immigrants, but are the effects more pronounced for Latino immigrants than other racial minorities? Renee Rocha and colleagues (2015) address this question by analyzing survey data that included both Latino and Anglo (defined as non-Hispanic Whites) residents in Texas. Rocha and colleagues (2015) find that high-enforcement communities, as measured by the prevalence of removals carried out under Secure Communities, are associated with lower levels of trust in government and of political efficacy among both native-born and foreign-born Latinos, whereas the reverse is true for non-Hispanic Whites. They also find that these results are more pronounced when it comes to noncriminal removals. Rocha and colleagues (2015, 901) thus conclude: “immigration policy enforcement redistributes trust in government from resource-deprived immigrants to Anglos,” which perpetuates existing political inequalities.
Eric Baumer and Min Xie (2023) use longitudinal data from the National Crime Victimization Survey to show that the activation of Secure Communities and the adoption of 287(g) agreements had racially disparate impacts on victimization. Specifically, Baumer and Xie find these federal-local partnerships significantly increased the risk of violent victimization (including rape, sexual assault, robbery, and assault) among Latinos, but not among non-Latino Whites and Blacks. Reva Dhingra and colleagues (2021) reach similar conclusions in their study of crime reporting. They show that total reported crime fell in counties with higher shares of Hispanics and where local law enforcement had more cooperation with ICE. Conversely, Elisa Jácome (2022) examines PEP, the Obama-era policy initiative that we briefly described earlier, which modified Secure Communities to rebuild trust between law enforcement and immigrant communities by focusing immigration enforcement action only against individuals convicted of serious criminal offenses. Jácome (2022) analyzes data on police incidents from the Dallas Police Department to demonstrate that the number of violent and property crimes reported to the police by Hispanics increased significantly after the introduction of PEP compared to those reported by non-Hispanics.
Policy Implications and Directions for Future Research
Immigration enforcement federalism is a core feature of the contemporary immigration enforcement system. Contacts with state or local law enforcement can lead to federal immigration detention and deportation for immigrants. In this context, the role that race might play in shaping the expansion or curtailment of local law enforcement agencies’ powers demands special scrutiny by policymakers and scholars alike. Considering the existing studies’ mixed findings on the racial threat hypothesis, an important task for future research is to more systematically theorize the nature of racial threat and how it might change over time, as well as explore new ways of measuring racial threat. For example, demographic characteristics like the size of the minority population (whether measured in absolute or relative terms) may not be a good proxy for racial threat, which after all, is about the perceptions of threat by dominant group members (see Feldmeyer and Cochran 2018). In addition, the growth in the size of a minority population under certain circumstances can give rise to a social dynamic that is quite different than what is expected under the racial threat hypothesis. Specifically, the “contact hypothesis” suggests that the racial attitudes of White Americans tend to moderate or soften over time from increased contacts with racial minorities (Lewis et al. 2013).
There are also important policy implications to the findings that indicate that racial diversity of local law enforcement personnel is a significant predictor of whether a local jurisdiction will become involved in immigration enforcement.8 These findings suggest that the diversification of local law enforcement agencies might be a policy tool that can counter the negative impacts of immigration enforcement federalism. However, these findings also give rise to several important questions that warrant future investigation, including: What type and level of diversity is needed? If racial minority groups generally tend to have greater sympathy for other marginalized populations, including immigrants (see Fussell 2014), direct racial concordance between the law enforcement personnel and the targeted immigrant group would not be necessary to achieve the desired policy goal. What are the mechanisms through which the racial identity of the law enforcement personnel operates to counter or curb the involvement of local law enforcement agencies in immigration enforcement? Is the diversity of leadership within local law enforcement agencies relatively more or less important than the diversity of frontline police officers? Addressing these and related questions promise to generate policy-relevant insights.
There are also wide-ranging policy implications that follow from research on the impacts of immigration enforcement federalism. Understanding racial minorities’ perceptions of, and experiences with, racialized policing is important because they have wide-ranging downstream effects on immigrant’s health, their trust in the police and views about police legitimacy, and their willingness to report crime and cooperate with the police (Morales and Curry 2021; Nichols et al. 2018). Thus, more research is needed to better understand how and why racial minorities perceive and experience local immigration enforcement policies and practices as racially discriminatory and biased. Finally, future research on racially discrepant policing should go beyond a narrow focus on stops and arrests to uncover the many other ways that other ostensibly race-neutral policing practices can produce racially disparate outcomes when local criminal justice institutions become involved in immigration enforcement. As Amada Armenta and Isabela Alvarez (2017, 7) have argued, there are “numerous, and sometimes hidden, ways that law enforcement agencies contribute to immigration control.”
Externalization of US Border Enforcement
In this section, we review the scholarship on the externalization or expansion of US border enforcement strategies beyond the physical southern US border as these strategies intersect with racialization practices in the migration corridor. Externalization policies have been implemented throughout Mexico (Basok et al. 2015; FitzGerald 2019; Frelick et al. 2016) and have extended into Guatemala and El Salvador (McGuire and Coutin 2013) in efforts to stop asylum seekers from reaching the southern US border. Today US externalization has expanded farther south through Costa Rica, Panama, and Ecuador, where visas were reintroduced for citizens of China, Afghanistan, Bangladesh, Eritrea, Kenya, Nepal, Nigeria, Pakistan, and Somalia with the stated justifications of combatting smuggling networks and preventing migrants from using Ecuador as a steppingstone to journey by land to the United States (Álvarez Velasco 2020).9 However, in this review, we focus on US externalization strategies in Mexico, where they have a longer history and thus the scholarly research on their evolution and effects is more robust.
We identify two main areas of empirical research. First, we review works that reveal the expansion of US externalization policies through Mexico; these works draw attention to how key US policies have evolved over decades. Second, we focus on research on the effects of externalization policies on the ground, with attention to how such US policies have transformed the migration corridor in Mexico as a transit country. A key aspect of both strands of research is a nearly universal focus on Central American migration through Mexico, which reflects policy efforts that until recently have been mostly aimed at containing Central American migration. This body of work exposes how US border externalization strategies shape, and are shaped by, the racialization of Central American migrants. This literature illustrates the ways that US policies beyond the border operate to both racialize and criminalize individuals before they even come to the United States, and it also illuminates how policies beyond the US border ultimately contribute to racially discrepant policing practices and racial discrimination within US criminal legal systems. We begin with a brief overview of the historical context, followed by discussions of the existing empirical research on the expansion of externalization programs and their effects.
Historical Context
US border externalization is not a new phenomenon; it is a process that has been unfolding over the past several decades (Rojas-Wiesner 2022) in Mexico,10 with marked acceleration in the last two decades (Shull 2021). Externalization has rested on an increased emphasis on pre-border surveillance and interception, including immigration checkpoints along the migration route in Mexico since at least the 1980s (Menjívar 2000; Minian 2020). These surveillance and interception practices became widespread with the significant rise in Guatemalan and Salvadoran migration that resulted from the civil wars in Central America (Menjívar 2000; Rojas-Weisner 2022). Border controls along the migration corridor since the 1980s were, therefore, a response to increasing numbers of Central American migrants seeking to enter the United States, and the goal was to prevent them from reaching the US border (Minian 2020; Rojas-Wiesner 2022).11
As foreign policy goals and political concerns in the United States have changed over time, externalization measures have been subsumed under national security and international crime control strategies aimed at combatting human smuggling, drug trafficking, and terrorist threats (Frelick et al. 2016). However, the focus of US externalization measures continues to be on stopping, delaying, returning, and deterring asylum seekers and migrants as they journey to the United States (FitzGerald 2019). Although historically these efforts have focused almost exclusively on Central Americans, lately these efforts have been expanded to also target Cubans, Haitians, Nicaraguans, and Venezuelans as well (American Immigration Council 2023).
Over time, and in collaboration with the United States, Mexico has progressively instituted an increasingly broad array of legislative and enforcement programs, including visa requirements that have become gradually more stringent (especially for Central Americans), funneling migrants through certain paths, implementing policies of interdiction, and requiring temporary permits in transit (Hurrell 2006). Externalization schemes have expanded so deeply into Mexico that scholars have argued that Mexico has become a “vertical border” (Torre Cantalapiedra 2021; FitzGerald 2019).
Major US Externalization Programs
Existing empirical research suggests that US externalization programs have played a key role in the racialization and criminalization of Central Americans. Migrant detention in Mexico, which has grown significantly in the last twenty years (Castillo 2022), is an important case in point. In a three-year ethnography of detention facilities in Tapachula, Alethia Fernández de la Reguera Ahedo (2020) shows that despite promoting themselves as civil institutions of temporary confinement, detention in Mexico has followed the US pattern; thus, facilities in Mexico also operate as punitive mechanisms punctuated by protracted waits where due process and basic rights are denied (Fernández de la Reguera Ahedo 2020). Fernández de la Reguera Ahedo observes an othering of Central American migrants in detention, as Central Americans in Mexico (like in the United States) are racialized as gang members, dangerous, and associated with dirty work and criminality (see also Vogt 2020). She explains that conditions of filth inside detention facilities create an image of these migrants as dirty, which contributes to their othering.12
In addition, existing empirical research on externalization shows that externalization programs work unevenly and are implemented dissimilarly on the ground across different migrant groups. Studies show that race and racialization practices are central to understanding these disparities. As we will discuss in greater detail, the US Border Patrol focuses interdiction activities overwhelmingly on the southern US border to stop Mexican and Central American migrants who travel by land (Heyman 1999).13
Empirical studies have examined how Mexico and the United States have established multiple interdiction and deterrence collaborations, backed by significant US investment in military and security forces, such as the Smart Borders initiatives that provide technological infrastructure to control US-bound migratory flows through Mexico (Fernández De La Reguera Ahedo 2020; FitzGerald 2019; Minian 2020; Muñiz 2022). The multiple programs and initiatives have contributed to a significant expansion of enforcement along the border and interior migration routes (Solano and Massey 2022). For instance, the Border Partnership Agreement of 2002, which was aimed at disrupting human and drug-smuggling networks operating in Mexico and the United States, led to a marked increase in detention and deportation of migrants from Mexico (Solano and Massey 2022).
The two countries, along with Canada, also signed the Partnership for Prosperity, which became the Security and Prosperity Partnership of North America in 2005. The Partnership for Prosperity is aimed at promoting greater economic cooperation in the region (Delano 2012), but signatory states present the Partnership as a means of addressing security concerns and enhance border security through information sharing (Villarreal and Lake 2009). Plan Mérida, created in 2008, was a security-focused partnership that included Central American countries, and was intended to combat drug trafficking and human smuggling along the main migration routes. Plan Frontera Sur / the South Border Plan, which was implemented in 2014, was formed in response to increasing numbers of unaccompanied migrant youth from Central America (Castillo 2022). As part of these collaborations, Mexico created a National Guard in 2019 and deployed it to its southern border to break apart the “caravans” arriving from Central America.
Externalization also comprises bureaucratic measures that screen potential migrants, such as visa requirements before migrants leave the sending countries. Visa requirements for migrants in transit contribute to further narrowing the migratory flow. Mirroring US temporary protection programs, Mexican immigration law since 2011 has granted temporary stays for victims of crimes and other categories of vulnerable migrants. However, Mexico’s bureaucratic agencies often discourage migrants from applying for these permits (Basok and Rojas-Wiesner 2018). Externalization also includes strategies such as the Dangers Awareness Campaigns that are funded by Mexico and the United States and are aimed at warning Central Americans of the dangers of the trip as a way of deterring migration (Vogt 2017).
Externalization policies are also implemented at the US-Mexico border to prevent migrants (primarily Central American asylum seekers and increasingly asylum seekers from Venezuela, Nicaragua, and Haiti) from approaching the border. In recent years, these policies have included a metering policy (or “queue management”), implemented at US ports of entry in 2016, whereby US Customs and Border Protection (CBP) limited the number of migrants permitted to access the asylum process each day (American Immigration Council 2021b). They also included the Migration Protection Protocols (MPP), in effect between 2019 and 2022, which designated Mexico as a safe third country to which migrants and asylum seekers could be returned for the duration of their immigration proceedings (Smith 2023). During the COVID-19 pandemic, the Trump administration also invoked Title 42, a rarely used law that was enacted in 1944, to allow officials to categorically refuse entry to migrants at the US-Mexico border, including asylum-seeking migrants, on the stated grounds of preventing the spread of contagious disease (Gramlich 2022).
Effects of Externalization
Existing research has identified several important consequences of externalization. First, externalization strategies have transformed patterns of Central American migration to and through Mexican territory (Basok et al. 2015). Using census and nongovernmental organization data from Mexico, Jasso Vargas (2021) examines the effects of the geography of Plan Frontera Sur in 2014 and finds that after its implementation, Central American migrants spent longer waiting times in southern and northern border states of Mexico. Scholars have noted that Mexico is no longer simply a transit area but, instead, a space where migrants linger for uncertain periods (Anguiano-Téllez 2008) or simply end up settling (Rivas Castillo 2012). Indeed, “transit,” scholars argue, may obscure the circularity, unpredictability, and instability of the journey for Central Americans in the context of externalization programs in Mexico (Basok et al. 2015). With the roadblocks that containment programs have created, research now points to a back-and-forth movement (rather than a linear step-by-step process) as migrants consider their options and opportunities, a kind of mobility shaped by race, class, and gender (Arriola Vega 2021).
Externalization has also contributed to transforming “transit countries,” such as Mexico, into countries of emigration, transit, and destination simultaneously (Caso Raphael et al. 2006). Extended and uncertain periods in Mexico have transformed migrants’ intentions and migratory flows through Mexico. Scholars find that as migrants face more obstacles to journey through Mexico, they have been “stuck in place” (Frank-Vitale 2020). Thus, many migrants have sought and found employment in Mexico, mostly in low-paying jobs, either to earn cash to continue the journey or as a strategy of settlement. This has meant that many migrants are increasingly channeled to new destinations in urban areas such Mexico City (Nájera Aguirre 2022). Demographic data support these findings, showing that Central American migrants are spending increasing amounts of time in Mexican territory, trekking to places far from the traditional migrant route (Jasso Vargas 2021). Jéssica Nájera Aguirre (2022) finds that Mexico City’s informal economy has become an important entry point for Central American migrants to adapt to the country and navigate the increasing precarity resulting from enforcement measures (see also Castillo 2022; Jasso Vargas 2021). However, the racialization of Central Americans persists in the workplace in Mexico; Central American workers are often seen as inferior and become targets of mistreatment (Doering-White 2021; Jasso Vargas 2021).
Second, research shows that to respond to these transformations, Mexico has implemented laws and policies that extend legal status to those who stay there (Basok and Rojas-Wiesner 2018), but such laws and policies do not always protect migrants (Galemba et al. 2019; Nájera 2022). In addition, as border externalization has pushed migration flows to dangerous terrain and criminal networks, there has been an increased need among migrants to hire smugglers to traverse more perilous routes (see also Frank-Vitale 2020, 2023). To address this need, the “migration industry” (Hernández-León 2013) has flourished, with a wider array of actors who now participate in facilitating migration. This literature reveals how externalization policies actually fuel the creation of the very kinds of criminalized smuggling networks and organizations that they purport to address.
Finally, scholars show that increased externalization has contributed to the rise of overtly nativist discourse in Mexico that characterizes Central American migrants, especially those who are part of migrant “caravans,” as invaders. These views contribute to growing rates of violence against asylum seekers in Mexico (Ortega Velázquez 2020). The scholarship on migrant experiences in Mexico points to various forms of violence that externalization has made possible. Relying on oral histories with (mostly Indigenous) Guatemalans, Ana Raquel Minian (2020) compares their experiences on the migrant trail through Mexico before and after externalization programs were put in place. This study finds that after the mid-1980s, migrants were more likely to be exposed to violence while traveling through Mexico owing to the need to evade government agents and as they were pushed to ever-more treacherous routes. Through ethnographic and archival research, Rebecca Galemba and colleagues (2019) analyze complaints from Central American migrants in shelters to examine the violence these migrants experience in transit. These researchers observe that most of the reports involve violence at the hands of state actors, including those who act in concert with non-state actors like taxi drivers (see also Basok and Rojas-Wiesner 2018; Brigden 2018; Doering-White 2018; Frank-Vitale 2020; Vogt 2017).
However, existing research also notes the complex social dynamics in these spaces of transit that defy portraying transit experiences solely as violent. As migrants cross more difficult and dangerous terrain, state and non-state actors become increasingly involved either in rescuing migrants or in enforcing the various strategies of border externalization. Through in-depth participant observation in Mexico’s Central American migrant trail, John Doering-White (2018, 2021) and Amelia Frank-Vitale (2020, 2023) illustrate the development of Central American community and mobility along the Mexican migrant trail. These authors map the various social dynamics and the actors involved in facilitating or restricting the migrants’ mobility within the constraints that US externalization strategies create. Through ethnographic analysis, Doering-White (2018) observes that even migrant shelters are caught between their humanitarian goals of protecting migrants from smugglers and the need to acknowledge migrants’ reliance on smugglers to safely pass through the increasingly dangerous transit corridors in Mexico. This mirrors Mexican policy in general, as Mexico has attempted to reconcile the dual goals of migrant protection and legal enforcement, creating opportunities for violence along the routes (Galemba et al. 2019).
Although Central Americans in the research we review here experience abuses and violence in transit, some obtain assistance and support, including from state agencies. Using interview data, Amalia Campos-Delgado and Karine Côté-Boucher (2024) examine the control arrangements that characterize detention facilities in Mexico, where they sometimes observed empathic encounters between detained Central Americans and the officers in charge of containing their movement. More broadly, as conditions in the transit corridor have worsened, civil society organizations in Mexico have become key sources of direct aid, legal knowledge, and logistical advice to the migrants (Torre Cantalapiedra and Mariscal Nava 2020). As well, individuals acting independently have provided food and shelter to Central American migrants, such as the women who stand by the railroad tracks throwing bags of food for the migrants on the train (Montes and Paris Pombo 2019).
Policy Implications and Future Research
The scholarship on transit migration through Mexico we review here has focused on the experiences of Central Americans, as these migrants have been racialized as dangerous and criminals and thus have borne the brunt of the enforcement and administrative policies at the core of US border externalization. These studies illuminate racialization practices in institutions and in informal interactions that shape migration in the context of externalization. However, as migrants from around the globe now cross Mexico en route to the United States, there is a clear need for comparative analysis that examines whether and how racialization and the distinct forms of racism endemic to sending, transit, and receiving countries inform the expansion and implementation of externalization programs. This research would refine current understandings of how race works in externalization strategies that unevenly affect different migrant groups on the ground.
The literature on Central American transit in Mexico has paid insufficient attention to informal institutions, as most of this literature focuses on nongovernmental organizations (such as shelters), “third spaces” (such as churches), or governmental agencies and detention in Mexico (Doering-White 2021). As Central Americans and many other migrants today wait longer periods in Mexico, they form relationships with nonmigrants and come in contact with informal institutions as they enter precarious labor and housing markets (Vogt 2020). For example, migrants develop social relations with locals who might regard them with solidarity at times, but most commonly see them through the stigma of criminality. More research is needed on these informal spaces and social relations between migrants and nonmigrant Mexican nationals, with attention to intersections of race, gender, and social class. Research on informal relations where racialization processes play a central role would contribute theoretically to understandings of complex social dynamics in new enforcement contexts. This is necessary for developing sound policies to address the challenges migrants experience in transit. Finally, given the longer and more uncertain times that migrants experience in Mexico, more research is needed to better understand migrants’ social and economic lives in the places where they live and work, even if only temporarily.
Technologies of Enforcement
Over the past fifteen years, federal executive agencies have greatly expanded their use of technology in the service of immigration enforcement (Meissner et al. 2013). This is true both at the border and in the interior of the United States (Meissner et al. 2013). Technology has also facilitated a renewed focus on the workplace as a site of immigration enforcement (Stumpf 2012). Our review examines empirical studies of immigration enforcement technologies across these three broad categories: border enforcement technologies, interior enforcement technologies, and workplace enforcement technologies.
A growing body of legal scholarship on technological shifts in enforcement raises legal and policy concerns about these developments (Hu 2015; Kalhan 2013, 2014; Kritzman-Amir 2021; Stumpf 2012). And a large and growing body of theoretical work is dedicated to the problems raised by these technological developments, including the rapid growth of immigration enforcement databases, surveillance technologies, algorithmic risk assessment tools, and related privatization trends (Chacón 2022b; Skinner 2018). However, there are fewer empirical studies of these developments, perhaps in part because the government agencies and private contractors who design, operate, and maintain these systems share very little information about them with researchers or the public (Muñiz 2022). Of these studies, only a handful directly focus on how racial bias is reflected in the design of, or reproduced by, these technologies. Nevertheless, in the final subsection, we discuss policy implications drawn from existing studies.
Historical Context
Over the past two decades, Congress has authorized significant spending for technological tools that are purportedly designed to facilitate the policing of the US-Mexico border by US Customs and Border Protection (CBP) and other enforcement agents. This rise in spending has been part of a general expansion of CBP, which had a budget of $16.9 billion in fiscal year 2020 (Heyman 2022, 126). Much of this spending has been geared toward enforcement initiatives along the US-Mexico border. Of the nearly twenty thousand agents employed by CBP in fiscal year 2019, close to seventeen thousand were stationed at the US-Mexico border region (Heyman 2022). CBP agents and other federal law enforcement agents rely on a wide array of technologies in policing the border. At border checkpoints and other ports of entry, technological surveillance tools used to police migration include the use of cameras for license plate scans, facial recognition scans, scans of biometric data on travel documents, X-ray and gamma ray scans of some border crossers and their vehicles, and scans of preclearance passes that also rely on stored data and that facilitate cross-border movement in discriminatory ways (Heyman 2022).14 On the land border, between border checkpoints (where agents are scarce), immigration enforcement agencies rely on surface or near-surface sensors and cameras, balloons, drones, and satellites to monitor movement at the border. These devices often operate alongside physical fencing. (Heyman 2022).
This proliferation of technological tools used for immigration enforcement purposes at the border has been accompanied by similar developments within the interior of the country. Since 2001, the federal government has significantly expanded funding for immigration enforcement in the interior of the United States. The federal government has allocated these funds to increase not only federal immigration enforcement personnel dedicated to policing immigration within the United States but also the enforcement technologies—often developed and managed by private entities—that can be used to accomplish enforcement goals (Kalhan 2014; McCarroll 2020; Meissner et al. 2013). For many individuals, this surveillance begins even before border crossing, extends throughout the border-crossing process, and is ubiquitous within the interior of the country (Kalhan 2014). Indeed, some scholars have argued that the distinction between border and interior enforcement has become increasingly artificial, as many of the technologies used to identify, investigate, track, and monitor the targets of immigration enforcement efforts have produced a digital border that follows people wherever they go, irrespective of the locations of physical borders or of the citizenship of the monitored individual (Kalhan 2014; Muñiz 2022). Moreover, the intensification of enforcement and the proliferation of enforcement technologies in areas as far as one hundred miles from the border increasingly have blurred the categorical distinctions between interior and border enforcement (Boyce 2023).
Finally, and relatedly, the government has developed a database designed to facilitate the enforcement of federal laws that prohibit the employment of unauthorized workers.15 Since the passage of the Immigration Reform and Control Act of 1986 (IRCA), US law has required US employers to complete administrative procedures to confirm that their workers are authorized by the US government to work. The law imposes civil penalties on employers who fail to comply with these administrative procedures, and criminal penalties on employers who knowingly engage in a pattern or practice of hiring unauthorized immigrant workers (Wishnie 2007). In 1996, Congress passed the Illegal Immigration Reform and Immigrant Responsibility Act (IIRIRA) to facilitate more effective compliance with IRCA by refining relevant databases and creating the E-Verify program. Congress intended the E-Verify program (originally called the Basic Pilot program) to facilitate employers’ investigation of potential employees’ work-authorization status by allowing employers to use worker information compiled by the Social Security Administration (SSA) to check the workers’ identification information against a centralized database maintained by DHS (Stumpf 2012). The system became available to employers in November 2007 (Bruno 2018). Employer participation is voluntary under federal law, though some employers, such as federal contractors, are required to use E-Verify (Bruno 2018).
Border Technologies
In the context of border enforcement, a very limited body of empirical studies has focused on several specific technologies, including cell phones and mobile applications, DNA testing, technologies like drones and sensors that monitor physical border entry, and the screening of individuals (and their documents) at points of entry, including airports. Studies of these border surveillance practices suggest that while, historically, border surveillance technologies targeted racialized Chinese and Mexican migrants, the proliferating systems aimed at the detection and surveillance of border crossers increasingly target Latinos broadly, as well as migrants identified as Muslim (Chaar López 2024). In turn, these technologies themselves become “racializing assemblages” that actively produce racialized targets of enforcement (Chaar López 2024). Exemplifying this phenomenon, one qualitative study found that the US-born adult children of Mexican immigrants living in the US-Mexico border region were collateral subjects of border enforcement technologies and policing who, during periods of increased border enforcement, experienced declines in emotional well-being that White residents did not experience (Flores-Gonzalez et al. 2024).
Cell Phones and Mobile Applications
Cell phones are an increasingly important tool of border surveillance and management. Migrants and the US government rely on cell phones as means of navigating and controlling border crossings, respectively. At the most basic level, cell phones provide location data that can be used by both enforcement agents and organizations seeking to provide humanitarian assistance to migrants (Walsh 2010). In October 2022, CBP introduced a mobile application for cell phones called CBP One (US CBP 2023; American Immigration Council 2021c).16 Immigrants without entry documents were instructed to use it to schedule appointments for interviews—including preliminary asylum screening (US CBP 2023)—prior to arrival at US ports of entry (US CBP 2023). One study analyzes the media and advocacy reports of the “glitches” that migrants encountered when trying to use CBP One (Kocher 2023). Among other things, the reported glitches suggest that the required use of CBP One created new barriers for migrants who were darker skinned or Black, as they reported experiencing greater difficulties than other migrants in uploading their photos to the application (Kocher 2023). But the study did not purport to be a systematic study of the CBP One app, or its racial impacts, and to our knowledge, no such study currently exists. The same can be said of cell phone tracking technologies.
DNA Testing
DNA testing is another important technological development used in policing international borders. Both globally, and within the United States, DNA testing is becoming an increasingly common feature of border inspection and admissions processes. Beginning in 2000, the Immigration and Naturalization Service (INS) provided agency guidance to its officers concerning the use of voluntary DNA testing by immigrants seeking to establish familial relationships for immigration purposes (Cronin 2000). Under President George W. Bush, the US government began mandating the DNA testing of certain overseas refugees seeking to unite with family members in the United States to confirm the “legitimacy” of the intending immigrant’s claim of familial ties (Dove 2013). Those efforts were not universal, instead focusing on refugees from Somalia and Ethiopia, and based on official assertions that this population had high rates of fraud (Dove 2013). After a hiatus in this testing, in 2012, US Citizenship and Immigration Services (USCIS) announced the resumption of DNA testing of refugees seeking family reunification in the United States (Dove 2013). By focusing on some nations and not others, these testing programs have been implemented in racially discrepant ways that predominantly affect Black migrants.
More recently, the Trump administration promulgated regulations for routine DNA testing of immigrant detainees (Makhlouf 2020). During the enforcement of a widely criticized “zero tolerance” policy aimed primarily at immigrants from Central America, children were separated from the adult caretakers with whom they had entered the United States, while many of those adult caretakers were criminally prosecuted and deported. Subsequently, the Office of Refugee Resettlement (ORR) under President Trump relied on DNA testing in efforts to reunify children with their family members (Makhlouf 2020). A federal court eventually ordered ORR to rely on DNA testing only when necessary, and under strict guidelines (Makhlouf 2020). Because access to mobility is itself differentiated along racialized national origin lines (Achiume 2022), Tally Kritzman-Amir hypothesizes that “there is at least a correlation between race and the chances of being required to take a DNA test” (Kritzman-Amir 2021, 104). The descriptive accounts of when and how DNA data are used in immigration enforcement at the border seem consistent with this hypothesis and suggest the need for empirical studies of racial disparities in, or racial biases stemming from or reinforced by, DNA testing in US border-enforcement processes.
“Virtual Border” Sensors
A third category of border enforcement technologies constitutes a “virtual border.” In addition to the controversial physical border wall promoted loudly by President Trump and constructed by his predecessors and his successor (Gulasekaram 2012; Mendoza 2023), the US government has adopted the use of several technologies that reinforce the border virtually, albeit quite imperfectly. These government efforts have involved a mix of radars, sensors, and cameras in various sectors of the US-Mexico border (US GAO 2017). Even the government agency charged with patrolling the border has not been able to accurately assess the efficacy of these technologies (US GAO 2017).
A small but growing number of empirical studies assess these physical surveillance technologies. One study applies geospatial analysis of landscape and human variables in a corridor of the US-Mexico border in Arizona-Sonora heavily traveled by undocumented immigrants to test whether the expansion of the border surveillance infrastructure, Secure Border Initiative Network (SBInet), shifted their routes toward areas that are more remote and difficult to traverse (Chambers et al. 2021). The authors find “a statistically meaningful difference in the distribution of human remains pre- and post-SBInet” (Chambers et al. 2021, 462), indicating that US border surveillance technologies worked in conjunction with the difficult desert terrain to increase migrant suffering and death. The impacted populations are most likely to be Latino immigrants, given the geographic siting of these surveillance tools (Heyman, 2022).
Screening Technologies
Finally, in addition to technologies designed to detect individuals entering without inspection, the US government has increasingly employed new technologies to screen the bodies and the documents of individuals crossing the border, whether by land, sea, or air (Hu 2015; NIF 2022; US GAO 2020). While theoretical accounts of these technologies attend to their likely racial impacts (Hu 2015; Skinner 2018), largely missing are systematic empirical studies of these technologies.17
One reason that the racial impacts of these border screening technologies are difficult to study is because of the classified or semi-classified status of many of the relevant technological tools, analytical methods, algorithms, and databases (Hu 2015). This may explain why there is so little empirical study of how the government constructs surveillance and enforcement tools like its “No-Fly list” (Hu 2015). One small-scale, qualitative account focused on the experience of Muslim women in hijabs at US airports, observing how these racialized travelers are excluded from the frictionless “trusted traveler” status purportedly made possible using databases in border management (Hegde 2019).18 Our review suggests a need for additional empirical studies of biometric border screening and its racial impacts.
Interior Enforcement
Several scholars have helped to map the proliferating technological tools used for immigration enforcement efforts in the interior of the United States. These studies use publicly available information to document the growth in border and interior enforcement technologies, strategies, and funding (Kalhan 2014; McCarroll 2020; Meissner et al. 2013). The technologies that they identify can be broadly divided into four categories: DHS databases used for retaining and accessing information about individuals, artificial intelligence tools that facilitate the integration of massive amounts of individual data obtained by data scraping and biometric recognition tools, surveillance tools, and risk assessment tools used in immigration enforcement efforts. There are virtually no empirical studies on the first two categories of interior enforcement technologies, but there are a handful of studies on surveillance and risk assessment tools, with some attention to the question of their racial impacts.
DHS Databases
DHS maintains several databases with information used for immigration and other law enforcement purposes. Earlier, we discussed the operation and effects of the Secure Communities program, which relies on the fingerprint screening of arrestees through DHS’s Automated Biometric Identification System (IDENT). But this is only one of the databases that DHS uses to manage enforcement efforts. The Enforcement Integrated Database (EID) also plays an important role in immigration enforcement efforts (Muñiz 2022). The EID stores and maintains information related to investigations, arrests, bookings, detention, and removal of noncitizens encountered during immigration and criminal law enforcement investigations and operations conducted by ICE, USCIS, and CBP. Ana Muñiz (2022) uses official government documents to trace the thirty-three-year evolution of EID as it evolved from a case management system to a “mass interior surveillance system.” She documents the ways that ICE agents and other immigration authorities have relied on the secondary information collected by DHS and other agencies and stored in EID in ways that have expanded both the number of people under DHS surveillance and the ability of DHS agencies to classify individuals as criminal or dangerous for enforcement purposes (Muñiz 2022). She also finds that “nonwhite immigrants may be particularly vulnerable” to the secondary ensnarement produced by long-lived dangerousness designations in the system (Muñiz 2022).
DHS also uses the data aggregating Immigration Case Management system (ICM) as the “core law enforcement case management tool for ICE Homeland Security Investigations” (US DHS 2023). The ICM is based on a system designed by the private company Palantir for police departments, modified for DHS’s immigration enforcement purposes (Woodman 2017). Users of ICM can access both government-owned databases and private databases and biometric data (Woodman 2017). The data of US citizens subject to criminal prosecutions are also available through ICM (Woodman 2017). Emma Knight and Alex Gekker (2020) reviewed documents concerning ICM and identified twenty-one different databases with individuals’ data, which are aggregated and centralized by ICM. They hypothesize that the ability of government agents to access a broad array of personal, location, and telecommunications information in the ICM system has enabled ICE to successfully target more immigrants for immigration enforcement efforts when those immigrants lack a criminal record (Knight and Gekker 2020). But they do not document this effect empirically, nor do they study the direct impacts of ICM’s operations on individuals or racial groups.
Artificial Intelligence in Immigration Enforcement
DHS also relies on artificial intelligence (AI) tools to scan and analyze patterns in the data that it maintains in its own databases and that it can access through other agencies and private entities. The data to which DHS has access include individuals’ biographical and biometric data, criminal and immigration histories, social media accounts, vehicle registration and insurance, tax records, professional licenses, health care records, and credit records (Nalbandian 2022). Lucia Nalbandian (2022) identifies seven companies that are pivotal to the use of AI in migration management in the United States: Northrop Grumman, Palantir Technologies, Giant Oak, NEC Corporation, Gemalto, Thomson Reuters, and Amazon.
Several empirical studies trace the effects of these mass dataveillance systems on immigrants. These studies are uniformly qualitative in nature, relying on publicly available or leaked governmental documents concerning surveillance technologies, secondary sources like media accounts, and interviews with individuals targeted by these surveillance systems or navigating their effects as advocates and activists (Guberek et al. 2018; Muñiz 2022; Villa-Nicholas 2023). One investigation found that study respondents—all Latino immigrants—are less vigilant in avoiding online surveillance than they are about avoiding surveillance in the physical world (Guberek et al. 2018). Another, which involved interviews with Latinx residents on the US side of the border near the San Ysidro port of entry, concluded that these dataveillance technologies commodify the data of Latinx people of varying citizenship status (Villa-Nicholas 2023). The most wide-ranging study relied on ethnography, interviews, and analysis of government documents to explore information-sharing partnerships between local police, state and federal law enforcement and foreign partners (Muñiz 2022). Among other things, Muñiz’s study shows how racialized designations of dangerousness (including gang labeling) and criminal labeling in various law enforcement databases are shared with immigration enforcement and immigration benefits agencies to spur deportations and block immigration relief, including eligibility for the Deferred Action for Childhood Arrival program (DACA) (Muñiz 2022). She finds that these designations disproportionately attach to, and operate to the disadvantage of, Black and Latino residents of all immigration statuses, with severe immigration consequences for noncitizens.
These studies suggest that Latino immigrants are disproportionately subject to the mass dataveillance efforts associated with immigration enforcement. They also suggest that Black and Latino immigrants may be uniquely harmed when immigration enforcement agents rely on artificial intelligence tools and data in interoperable databases tainted by racial biases inherent in policing, economic, and health care systems to identify appropriate targets for immigration enforcement. Additional empirical inquiries into these questions could usefully supplement the limited existing data.
Technological Surveillance Tools
Many of the same private companies that contract with the US government to operate immigration detention facilities also create and maintain technological systems, such as electronic ankle bracelets, designed to track and monitor non-detained immigrants subject to official monitoring (Chacón 2022b). Individuals subject to electronic monitoring (EM) are also visited by the enforcement agents who monitor those systems.
A growing body of legal scholarship describes EM and assesses the constitutional and human rights questions that these systems raise (Marouf 2017), but there are fewer empirical studies of the operation and effects of EM. One study, based on twenty-one months of ethnographic observations and semi-structured interviews, follow-up interviews, and informal conversations with thirty immigrants subject to EM found that it was experienced as stigmatizing and burdensome across racial groups (Martinez-Aranda 2022). Although EM devices are for the purposes of civil immigration proceedings, not criminal proceedings, the visual stigma of the devices reinforced immigrants’ subordinated status and distanced them from community-support networks (Martinez-Aranda 2022). ICE visits to the wearers of these devices also “triggered anxiety” among other members of the community whenever a monitor came to check up on the individual (Martinez-Aranda 2022). Consequently, individuals subject to EM also “became isolated from supportive networks because of EM’s capabilities to attract ICE” (Martinez-Aranda 2022). EM also has an adverse effect on the children of immigrants subjected to it because EM increases parental stress and contracts their children’s social networks (Martinez-Aranda 2023). Cecilia Menjívar and Andrea Gómez Cervantes (2025) reach similar conclusions in their ethnographic study of the criminalization of Maya Guatemalan asylum seekers in rural Kansas, showing how EMs keep asylum seekers indebted and embedded in the system in the same way as other system-impacted groups in contact with the criminal legal system.
We did not identify any other empirical studies concerning the effects of immigration-enforcement-related EM on individuals’ health or economic well-being, or on the health and well-being of family members. Nor did we identify any published studies of racial disparities in the use of electronic monitoring in immigration enforcement.19
Risk-Assessment Technologies
In managing the US immigration detention system, which is the world’s largest immigration detention system, ICE uses the Risk Classification Assessment (RCA) system. The RCA is used to determine whether it is appropriate to release a person from immigration detention during the pendency of their immigration proceedings. There are several empirical evaluations of the RCA.
One study analyzes 485 risk assessment results from ICE databases from Baltimore in spring 2013 and concludes that the RCA lacked transparency, was manipulable, did not reduce overall detention rates, and may have increased them (Noferi and Koulish 2014). Later studies analyze documents obtained from ICE through litigation in 2018 that explained the RCA’s algorithms for flight risk, vulnerability, and several other detention-related factors. The authors find that the RCA fails to standardize immigration detention decisions, fails to align detention decisions with risk, and has been manipulated by ICE officers and supervisors to favor detention and to amplify the biases of its human operators (Evans and Koulish 2020; Koulish and Calvo 2021). These earlier studies do not discuss either racial or national origin discrimination, though Kate Evans and Robert Koulish (2020) suggest that the manipulability of the RCA underscored the need for studies of the racial impacts of RCA manipulation. In a more recent study, Koulish and Evans (2021, 67–68) find that the RCA replicated systemic racism through its combination of punitive bias in the algorithm and racial bias in policing, a combination that “imposed double punishment on migrants of color” in the form of unnecessary and constitutionally suspect preventative detention.
The policy implications of these findings are contested. Evans and Koulish (2020) conclude that the most appropriate response to these technological failings is to eliminate immigration detention as a method of immigration control. In contrast, Richard Frankel (2023) acknowledges the problems with the RCA identified in earlier studies, but he nevertheless advocates for the use of a “well-designed” and “transparent” risk assessment system. He posits that RCA could serve an important purpose given that immigration judges are also prone to racial bias, make nontransparent bond decisions, and do not necessarily focus on the appropriate factors in their own bond determinations. Robert Koulish and Ernesto Calvo (2021), however, demonstrate that as a practical matter, human operators to date have manipulated the existing risk assessment tools in ways that magnify, rather than mitigate, racial biases.
Workplace Enforcement: E-Verify
The E-Verify program is a federal program under which employers check the identity documents of potential employees against a federal database to determine whether those employees are “authorized workers.” Although all employers must comply with IRCA’s worker verification process, employer use of the E-Verify system is optional (with some limited exceptions). Over the past fifteen years, however, state-level E-Verify mandates have been introduced in several states, sometimes as part of a larger bundle of state law measures aimed at restricting immigration. Because some states have enacted E-Verify mandates and others have not,20 researchers have used the variations in the enforcement landscape to examine the effects of the program, including its effects on undocumented immigration, work opportunities and wages, health, and crime rates.
Effects on Undocumented Immigration
The stated legislative goal of IRCA’s employer sanctions provisions, and the related E-Verify program, is to reduce unauthorized migration. Since IRCA’s passage, however, the number of undocumented immigrants in the US has increased substantially, leading some critics to declare employer sanctions a failure (Wishnie 2007). On the other hand, one recent study that focuses on the annual percentage change in undocumented immigrants by state and year, drawing on multiple statewide data sources, has found that E-Verify mandates may be more effective than other forms of state-level immigration restrictions at deterring undocumented immigration into enacting states (Gonzalez-Gorman 2022; see also Ellis et al. 2014). The passage of the Legal Arizona Workers Act in 2007, which mandated the use of E-Verify in Arizona, also appears to have reduced the share of Hispanic noncitizens (who comprised a substantial majority of undocumented immigrant residents of Arizona at that time) residing and working in the state, at least in the short term (Chalfin and Deza 2020; Lofstrum et al. 2011).
Effects on Work Opportunities and Wages
When the US Congress debated the enactment of employer sanctions provisions in the 1980s, some civil rights and Latino advocacy organizations raised concerns employers would be likely to respond to any sanctions program by discriminating against Latino applicants (Wishnie 2007). As a result, Congress required uniform document-check practices and prohibited citizenship, immigration status, and national-origin discrimination in hiring, firing, and recruiting (Perotti 1992).21 However, a national survey of 4.6 million employers in 1990 by the Government Accountability Office “determined that a startling 19 percent [of surveyed employers] had engaged in discriminatory behavior” against non-US citizens or those with “‘foreign’ appearance or accents” (Wishnie 2007).
The more recent, selective rollout of E-Verify programs across different states presented scholars with an opportunity to assess whether the 1986 federal employer sanctions provisions of IRCA—coupled with state-mandated use of an automated, federally maintained verification system (accompanied by state law penalties for employers)—might also lead to discrimination in employment. Perhaps unsurprisingly, early study of Mexican workers’ responses to E-Verify mandates suggests that, at least in the short term, these mandates raise deportation fears, reduce interstate mobility, and curb deportees’ desire to return to the United States (Amuedo-Dorantes et al. 2013). Ethnographic research also suggests that the expanded use of E-Verify (along with other immigration enforcement-oriented programs) produces an “extra vulnerable immigrant labor force” in the United States, with less bargaining power and greater susceptibility to exploitation, and that labor leaders expend significant energy attempting to mitigate these effects (Gomberg-Munoz and Nussbaum-Barberena 2011).
Relying on ten years of survey data, a study of naturalized Mexicans (who are authorized workers by virtue of their citizenship) finds that they themselves perceived less employment discrimination after E-Verify mandates were imposed, but they were also more likely to perceive that discrimination was impeding the success of Latino residents than were their counterparts in states that did not enact E-Verify mandates (Amuedo-Dorantes and Pozo 2019). One study used a synthetic control method to select a group of states without E-Verify mandates against which the labor market trends of Arizona could be compared (Bohn, Lofstrum, and Raphael 2015). The authors conclude that not only did Arizona’s E-Verify mandate fail to improve labor market conditions for authorized workers who compete with unauthorized workers but actually increased unemployment among authorized low-skilled workers in Arizona (Bohn, Lofstrum, and Raphael 2015).
The long-term effects of E-Verify mandates on wages remain uncertain. One recent study finds that E-Verify mandates have a negative effect on the wages of likely undocumented immigrant men, but positive wage effects for likely undocumented immigrant women (Orrenius and Zavodny 2021). Any negative wage effects of E-Verify mandates on undocumented immigrants do not obviously redound to the benefit of authorized workers. Recent studies find that E-Verify mandates produce no notable effects on farm wages (Luo and Kostadini 2022) or on the wages of US-born, non-Hispanic Whites (Orrenius and Zavodny 2021). The impact of E-Verify on Latino workers may also depend, to some extent, on where they live within a state with a mandate. Irene Browne and colleagues (2022) find that the mandates have no impact on wage-employed workers. However, self-employed Latinx residents in this study experienced significant earnings losses when living outside predominantly Latinx counties but significant earnings gains when living within such counties.
Undocumented immigrants appear to suffer a wage penalty as compared to authorized workers, one which increases over the lifetime (Borjas and Cassidy 2019). And immigration policies affect the size of that penalty, which increases when immigration restrictions tighten (as through mandatory E-Verify laws) and decreases when immigration restrictions loosen (as through DACA) (Borjas and Cassidy 2019). But additional study could help further disaggregate the effects of race and immigration status, and of employer sanctions laws and other forms of restrictive immigration policies.
Effects on Health
Two studies evaluate the effects of employer sanctions and E-Verify mandates on health and health care access. One study shows that E-Verify mandates are associated with undocumented immigrants’ decreased access to health insurance in the period immediately after implementation—the logical result of employers declining to hire wage workers who are not authorized to work (Churchill 2021). Another study finds that E-Verify mandates are associated with a decline in birth weight and gestational age for infants born to immigrant mothers with demographic profiles matching the undocumented population, as well as for infants of native-born mothers (Strully et al. 2020). These authors suggest that the economic and political climate that gives rise to E-Verify mandates poses risks to infant health among those groups associated with undocumented status (usually, but not always Latinos), regardless of actual immigration status.
Effects on Crime
Proponents of greater immigration restrictions, including E-Verify mandates, often justify their support for these initiatives on the grounds that greater immigration enforcement increases public safety. Efforts to understand the relationship between governmental employer sanctions and crime predate E-Verify. One study focusing on Bexar County, Texas, found an increase in felony drug charges in the heavily Latino Bexar County, Texas, after the enactment of IRCA, and theorized that this was a result of reduced labor market opportunities for unauthorized residents (Freedman et al. 2018). An earlier study suggested that changes in police behavior in Hispanic neighborhoods following the enactment of IRCA could confound estimates of the effects of IRCA on criminal activity (Bohn, Freedman, and Owens 2015). Freedman and colleagues controlled for this effect in their 2018 study. Another study uses data from the National Incident Based Reporting System to conclude that the enactment of E-Verify mandates is associated with a 7 percent reduction in property crime incidents involving Hispanic arrestees but no reduction in “violent criminal incidents” among Hispanic arrestees (Churchill et al. 2021). This study, however, does not fully explore the possible confounding effect of changes in police behavior.
Policy Implications and Future Research
There is a dearth of empirical studies on how DHS maintains enforcement databases, employs dataveillance, and uses artificial intelligence, perhaps because the opacity of these systems and processes impedes direct research. However, the existing studies of these systems consistently suggest that within the United States, Latinos are the primary targets of DHS’s immigration enforcement-related dataveillance (Cox and Miles 2013; Heyman 2022; Muñiz 2022). Studies documenting the effects of proliferating immigration enforcement should be understood as operating in a context of racialized implementation, even when the studies do not directly focus on racial effects of the program.
Studies also document racially disproportionate enforcement practices in database labeling and in the uses of electronic monitoring technologies, with Black and Latino residents more likely to experience their negative effects (Muñiz 2022). Qualitative studies exploring the impact of mass surveillance and electronic monitoring highlight the social and emotional harms of such monitoring. Additional empirical work could further document and specify the degree to which these technologies engender racial discrepancies in immigration enforcement. More research is also needed to add specificity to our understanding of how particular enforcement tools and policies contribute to racialization and affect different racialized populations, including subpopulations within large pan-ethnic groups.
In the area of worksite enforcement, which has a more robust set of related empirical studies, significant gaps and ambiguities exist. Studies suggest that E-Verify mandates tend to be adopted in states with higher levels of anti-immigrant sentiment, and such sentiment tends to be enacted (both in policy and in daily practice) in ways that target individuals according to race, with race perhaps standing in for perceived immigration status. But studies on the effects of employer sanctions and E-Verify mandates on immigrant health and crime rates are still relatively scarce. There is no scholarly consensus regarding the effects of the increased use of E-Verify on the wages of authorized and unauthorized workers, let alone the racial impacts of E-Verify use. Finally, to our knowledge, no study has focused specifically on the employers who are the purported target of both IRCA’s employer sanctions provisions and state-level E-Verify mandates. Given that the relevant criminal sanctions are aimed at employers, it might be worth exploring whether patterns of sanctions or criminal prosecutions for IRCA violations themselves reflect racial bias.
On the whole, existing studies of DHS enforcement databases, dataveillance, and uses of artificial intelligence in enforcement suggest a troubling lack of transparency in systems that shape immigration enforcement and have significant consequences in the lives of residents of the US, regardless of citizenship status. These studies also suggest that DHS’s enforcement technologies overwhelmingly target Latinos, and consequently, that any errors, misinterpretations of data, problematic exercises of discretion, and even technological glitches, in implementing the system on the ground disproportionately affect Latinos. Finally, the available data suggest that these systems not only reproduce racial disparities of criminal enforcement systems automatically but also may exacerbate them by amplifying the effects of law enforcement biases through algorithmic manipulation. Overall, these studies point to the urgent need not only for more study but also for greater transparency in immigration enforcement technologies, greater legislative oversight of the uses of these technologies in enforcement, and more robust legal protections against discriminatory law enforcement practices.
From Immigration Enforcement to Court Proceedings
In this section, we consider scholarly empirical studies that have examined what happens to immigrants who enter the US court system after being charged with immigration-related violations. Before proceeding with our review, it is important to make clear which bodies of research are excluded from our review. Studies that examine whether and to what extent criminal defendants’ citizenship status affects criminal court outcomes for all types of criminal charges are beyond the scope of our review (see, for example, Koo et al. 2022; Light et al. 2014). Research on immigration adjudication by service agencies within the Department of Homeland Security, such as the US Citizenship and Immigration Services (USCIS), is also outside the scope of our review (Ryo and Humphrey 2022; Schoenholtz et al. 2014). Finally, we do not examine studies of immigration cases decided by US Courts of Appeals and the US Supreme Court (see, for example, Law 2010; Vogler 2019).
Our review focuses on the following two sets of research. First, we focus on research on criminal prosecution of immigration crimes in criminal courts. By immigration crimes, we mean violations of immigration law that are sanctionable through criminal punishment. Immigration crimes generally include improper entry and reentry, smuggling, transporting, and harboring aliens, and immigration-related fraud (Santamaria 2023). Second, we focus on research on civil prosecution of immigration law violations in immigration courts. Specifically, we review the growing body of research on removal proceedings.22 Removal proceedings constitute the vast majority of proceedings conducted by immigration judges (Executive Office for Immigration Review 2022, 103). We begin our discussion of each of these bodies of research with a brief overview of the relevant legal background that contextualizes our subsequent discussions.
Research on Immigration-Crime Cases in Criminal Courts
Certain features of federal cases involving immigration crimes set them apart from other federal criminal cases. The most common types of immigration crimes are smuggling offenses, illegal-entry offenses, and fraudulent document offenses (US District Courts 2020a, table D-2). Most of these cases are brought in southwestern border districts (US District Courts 2020b, table D-3). The defendants in these cases are predominantly male, Hispanic and non-US citizens (Schmitt and Jeralds 2022, 18). Between fiscal years 2018 and 2020, prosecutions of immigration crimes constituted the largest share of all federal criminal prosecutions (Schmitt and Jeralds 2022, 5). To lighten these cases’ burden on the court system, some districts place these cases in fast-track programs, which allow prosecutors to offer reduced sentences to those who waive certain procedural rights, enter a prompt guilty plea, and agree to be sentenced immediately (Hartley and Tillyer 2012, 79–80).
Although legal scholars have written extensively about the nature and operation of federal prosecution of immigration crimes (see, for example, Chacón 2021b; Eagly 2010; Lee 2013), including the racist origins of laws that gave rise to these prosecutions in criminal courts (Fish 2022), empirical research on the prosecution of these immigration crimes remains relatively scarce. The small number of existing empirical studies relies exclusively on quantitative methods, and the outcome of interest in these studies is limited to sentencing decisions. To understand the determinants of these sentencing decisions, empirical studies have examined the following set of predictors: defendant characteristics (such as their race and gender), characteristics of the communities in which the courts are located, and offense types (such as trafficking, illegal entry, and document-related).
Hartley and Tillyer (2012) use the US Sentencing Commission (USSC) data from fiscal year 2008 to investigate the effect of legal, extralegal, and case-processing variables on sentence lengths of immigration-crime cases. This study finds that racial disparities in sentencing outcomes in these cases vary by type of offense and district in which the cases are adjudicated. Specifically, their results show that Hispanic defendants, compared to White defendants, received shorter sentences in illegal-entry cases, though not in trafficking nor documented-related cases (Hartley and Tillyer 2012, 98). However, Hispanic defendants received significantly longer sentences than White defendants in two of the five southwestern border districts examined (Hartley and Tillyer 2012). In addition, Hartley and Tillyer (2012, 99) find that in southwest border districts, Black defendants received significantly longer sentences compared to White defendants.
In a subsequent study, Tillyer and Hartley (2016) analyzed the same USSC data but narrowed their research focus to sentencing outcomes in illegal-entry cases in fast-track eligible districts. Their results indicate that the use of fast-track programs, the amount of sentence reduction applied in fast-track cases, and the overall sentence length are differentially impacted by various legal, extralegal, case-processing, and district-level variables. Of most relevance to our review, this study finds that male and Hispanic offenders received relatively longer sentences, leading Tillyer and Hartley (2016, 1638) to conclude that judges may be using sentence length to attempt to deter Hispanic immigrants, who constitute the majority of undocumented immigrants, from entering the United States.
Finally, Melanie Holland (2020) analyzed USSC data on sentences imposed on criminal defendants convicted of immigration offenses between 2006 to 2017. One of this study’s main foci was on the significance of gender on the likelihood that a judge will engage in a downward departure from sentencing guidelines under the fast-track program. She finds that although the odds of receiving a downward departure under the fast-track program did not differ significantly between male and female defendants in the study sample, gender disparities did exist across racial groups and countries of origin. For example, Holland (2020, 142) finds that Black and Hispanic women were less likely to receive downward sentence departures than White women. However, controlling for Hispanicity, Mexican women were more likely to receive this departure than women from other countries. In terms of the magnitude of the sentence length reduction among defendants who received a downward departure, women received larger sentence reductions than their male counterparts, but Hispanic women received significantly shorter reductions than White women.
A qualitative study of federal drug crime prosecutions along the southern border by Mona Lynch (2017) complicates these findings by suggesting that immigration sentencing data need to be understood in relation to drug prosecutions, at least along the southwestern border. Lynch’s study focuses on the criminal prosecution of immigrants—most commonly undocumented—who were apprehended by US Border Patrol while transporting drugs in backpacks as part of their payment to a coyote to make it across the increasingly difficult-to-cross US-Mexico border. Drawing on her observations of criminal proceedings, court documents, and interviews with attorneys and judges, Lynch (2017, 113) concludes that in sharp contrast to the drug prosecutions that she studied in other regions of the country, drug cases along the southwestern border are “immigration cases, and the prevailing adjudicatory logic is concerned with defendants’ status as unauthorized outsiders” This means that “the imperative driving their criminal adjudication is swift and efficient resolution to get them out of the system and out of the country” (Lynch 2017, 121). Thus, these immigrant defendants typically were mass processed whereby in exchange for a guilty plea, felony illegal reentry charges were dropped to misdemeanor illegal entry charges, and felony drug-trafficking charges were reduced to misdemeanor drug possession charges. These guilty pleas result in more lenient sentences compared to those in non-border districts, but this system also generates what Lynch characterizes as a unique form of over-criminalization of immigrants.
Taken together, the findings of these empirical studies show that the effect of defendants’ race on sentencing outcomes in cases involving immigration crimes are complex and conditional. These studies suggest that the effects of race on sentencing outcomes can vary by gender and national origin. They also suggest that a full understanding of the effects of race on sentencing outcomes requires considering geography and offense types, particularly courts’ proximity to the US-Mexico border, and the distinction between illegal-entry offenses and other types of immigration and criminal offenses.
Research on Immigration Court Process and Outcomes
Not all immigrants who are apprehended by US immigration authorities are placed in formal removal proceedings decided by immigration judges. Many immigrants who are encountered at the border are summarily expelled by frontline immigration officials without an opportunity to appear in immigration court (Leong 2022).23 However, immigrants placed in a removal proceeding must appear before an immigration judge. The two main types of hearings in removal proceedings are master calendar hearings and merits hearings.24 Because some immigrants in removal proceedings are detained during the pendency of their proceedings, immigration courts also hold bond hearings—also known as custody determination or redetermination hearings (Executive Office for Immigration Review 2022, 124). Roughly six hundred immigration judges preside over these types of hearings across approximately seventy locations throughout the United States (Executive Office for Immigration Review 2023). Immigration courts are located within the US Department of Justice—the department of the executive branch tasked with law enforcement—rather than the judicial branch.
A growing body of research examines immigration court process and adjudication. These studies use diverse methodologies (both qualitative and quantitative) to investigate wide-ranging topics. The following sampling of topics from recent studies of immigration court is illustrative: the use of interpreters and technology in immigration courtrooms (Barak 2021; Eagly 2015), the justifications that immigration judges provide for their decisions (Asad 2019), the role of legal representatives and their advocacy strategies (Levesque et al. 2023; Miller et al. 2015; Ryo 2018), how immigrant respondents view immigration judges and the court system (Barak 2023; Ryo 2017), and the importance of continuances granted by immigration judges (Hausman and Srikantiah 2016). The influence of legal and extralegal factors on immigration judges’ decisions has been of enduring interest to empirical researchers working in this area. In brief, studies have found that personal characteristics of immigration judges matter (Keith et al. 2013; Ramji-Nogales et al. 2009; Ryo and Peacock 2021). So do state and local contexts in which immigration courts are located (see, for example, Blue et al. 2021; Chand et al. 2017; Miller et al. 2014), and outside influences and pressures on immigration judges (Hausman et al. 2022; Heyes and Saberian 2019; Kim and Semet 2020a, 2020b).
Inequalities in the immigration court process and outcomes have been of core concern to scholars. However, empirical research that systematically examines racial bias or disparities in immigration court process or outcomes is relatively scarce. Instead, the dominant scholarly focus has been on disparities by national origin. For example, in a national study of access to counsel among immigrants in removal proceedings decided between 2007 and 2012, Ingrid Eagly and Steven Shafer (2015) investigated disparities in representation rates by immigrants’ nationality. They found that Mexicans, who constitute the largest nationality group in removal proceedings, were the least likely to be represented by counsel, with a representation rate of 21 percent (Eagly and Shafer 2015, 92). The next three nationality groups with low representation rates were also from Latin America: Honduras, Guatemala, and Nicaragua. In sharp contrast, Eagly and Shafer found that most Chinese nationals in removal proceedings—92 percent—were legally represented.
Research on immigration bond hearings by Ryo (2019) is another study that investigates disparities by national origin. This study examines immigration judges’ determination that certain immigrant detainees pose a danger to the community by analyzing audio recordings of 350 immigration bond hearings decided between 2013 and 2014 in the Central District of California. Controlling for a variety of detainee background characteristics and criminal-conviction-related measures, the study shows that immigration judges are more likely to find Central Americans as a danger to the public than non-Central Americans (Ryo 2019, 245). This study offers two possible explanations for this finding that warrant future empirical investigation: One possible explanation is that the sudden surge in Central American removal cases during the study period triggered in immigration judges a heightened sense of criminal threat from this group. The second possible explanation is that in a decisional environment characterized by time pressure and incomplete information, social stereotypes of Central Americans as criminals and gang members might have become highly salient.
Finally, two recent studies show that events that threaten national security or public health can stigmatize certain nationality groups, which can lead to judicial bias against those groups in immigration court. Abel Brodeur and Taylor Wright (2019) examined the effect of the 9-11 attacks on asylum decisions made by immigration judges. The study’s difference-in-differences analysis showed that the attacks significantly reduced the likelihood that applicants from Muslim-majority countries were granted asylum. Moreover, this estimated effect was even more pronounced among immigrants who shared the same nationality (Egypt, Lebanon, Saudi Arabia, and United Arab Emirates) as the 9-11 attackers (Brodeur and Wright 2019, 130). In another study, Peacock and Ryo (2022) investigated the effect of the COVID-19 pandemic on immigration judges’ removal decisions. They found that Chinese nationals, the group that was widely blamed for the pandemic, became significantly more likely than other groups to be ordered removed during the early pandemic period compared to the pre-pandemic period. Moreover, East and Southeast Asian nationals, the groups perceived to be most closely associated with, or belonging to the same ethnoracial category, as Chinese nationals, also experienced a significantly higher removal rate during the early pandemic period (Peacock and Ryo 2022, 581–83).
As the foregoing discussion indicates, much of the focus in this body of research has been on national origin, but two recent exceptions are worth noting. In a study of crime-based removal proceedings, Ryo and colleagues (2025) find that Hispanic immigrants with drug-related charges and Black immigrants with domestic violence or firearms charges obtain worse case outcomes than their counterparts. They also find that non-White immigrants fare better when their presiding judges are of the same, rather than different, race. In another recent study, Ryo and colleagues (forthcoming) find that race and criminal record are key predictors of whether an individual in removal proceedings obtains legal representation and the effect of that representation on case outcomes. Immigrants are much less likely to be represented if they have criminal charges, and Hispanic immigrants (whether or not they have criminal charges) have substantially lower rates of representation than any other immigrant group. Moreover, the positive effect of representation on case outcomes varies by criminal record and race. The representation effect is substantially greater for immigrants with criminal charges than for those without criminal charges. In addition, the representation effect is the greatest for White immigrants with criminal charges and the smallest for Black immigrants without criminal charges. Finally, this study finds that White immigrants with White lawyers obtain more favorable outcomes than White immigrants with non-White lawyers, but non-White immigrants (particularly Black immigrants) face less favorable outcomes when they are represented by same-race lawyers.
In sum, there is a growing body of empirical studies on immigration court adjudication that have focused on disparities in case outcomes by national origin. More research, however, is needed to identify the mechanisms that generate such disparities. In addition, there is a continuing need for scholarship that systematically examines the complex nature and sources of racial disparities in immigration court process and outcome, as well as the interaction between race and criminal record that might be shaping case outcomes. Finally, recent research highlights the importance of examining not only the racial identities of immigrants but also those of lawyers representing them and of judges presiding over their cases.
Policy Implications and Directions for Future Research
Whether and to what extent racial minorities experience harsher treatment or face worse outcomes are questions of fundamental importance for any court system. Empirical research suggests that race and nationality may be important axes of inequality in the adjudication of immigration crimes and removal proceedings, respectively. The policy implications of these study findings are potentially wide-ranging. For example, insofar as the source of such disparities is the activation of pernicious stereotypes, there is a need to develop effective antibias training for judges and provide them with expanded resources to create decisional environments (such as those characterized by less time pressure and more complete case information) that reduce the risk of reliance on mental shortcuts and categorical thinking. On the other hand, to the extent that the source of disparities is structural—for example, unequal access to legal resources, language assistance, or information networks—policy solutions must include community-based and court-based programs that address those specific structural barriers.
The foregoing discussion suggests that an important direction for future research in this area is identifying the specific mechanisms that generate disparities along racial lines or nationality. In addition, with respect to adjudication of immigration crimes, empirical research is needed to understand the effects of race on court outcomes beyond sentencing. For example, future research on immigration crimes should seek to address basic questions such as whether and to what extent racial disparities exist in decisions to prosecute and decisions to offer plea bargains. With respect to removal proceedings, expanding the research focus beyond national-origin disparities is an urgent task. As we noted earlier, conspicuously missing in the existing empirical literature are systematic examinations of racial disparities in immigration court process and outcomes. Filling this gap in empirical knowledge is critical to obtaining a more holistic view of how racial disparities across multiple legal institutions (from criminal justice to immigration enforcement systems) might have mutually reinforcing effects that might compound over time to generate cumulative disadvantages for immigrants of color.
Impact on Families
Scholarship over the past fifteen years has significantly expanded our understanding of the effects of immigration enforcement on families.25 Despite the US government’s putative support for keeping families together, US immigration law “contains many provisions that keep families apart and uncertain of when or if family reunification will take place” (Enchautegui and Menjívar 2015, 32). Using data from the American Community Survey and interviews with immigrants in Phoenix, Arizona, María Enchautegui and Cecilia Menjívar (2015) identify four key mechanisms through which immigration law engenders family separation: setting restrictions on who gains legal admission to the United States, mandating that immigrants in the United States maintain a certain level of income before sponsoring a family member, prohibiting legal entry to those who already crossed the US border without documentation, and regulating the lives of immigrants without legal status.26 While family separation through removal remains one of the most severe outcomes of immigration enforcement, the looming threat of separation also plays a critical role in shaping the contours and composition of families, as well as the meanings, relations, and practices that constitute family life.
This section focuses on empirical scholarship that explores the ways that immigration enforcement shapes life outcomes for, and relations between, parents, intimate partners, children, and other family members. Many of these family-impact studies show that immigration enforcement has a long arm that extends beyond the purported target of enforcement (that is, unauthorized immigrants) to reach even those who have legal status or are US citizens.27 For example, scholars find that US citizen partners of unauthorized immigrants live the social experience of noncitizenship (López 2015) and that immigrant families experience “collective liminality” that threatens their financial security and emotional well-being when a family member is detained (Martinez-Aranda 2020). Nina Rabin (2018, 1) emphasizes the importance of studying “secondary immigration enforcement” to capture how “young people, regardless of their own immigration status, are harmed by immigration enforcement aimed at their parents.” In addition, as we will detail at length, family-impact studies suggest that the effects of immigration enforcement often reverberate throughout entire communities in which immigrant families are embedded.28
Nearly all family-impact studies of immigrant life in the United States examine the experiences of Latinos. This research strategy is unsurprising; as we explained earlier, Latinos—and specifically Mexicans and Central Americans—remain the primary targets of immigration enforcement in the United States. As a result, this literature does not directly address disparities in treatment or outcomes across different racial groups. In the conclusion of this section, however, we revisit this issue by highlighting the urgent need for studies that systematically analyze the role of race in shaping how immigrant families are treated and impacted by immigration enforcement.
Approaches to Family-Impact Research
We begin by discussing the two main approaches that researchers have adopted to analyze the effects of immigration enforcement on immigrant families. The first approach focuses on the effects of specific enforcement policies or practices versus the overall enforcement climate.29 The second approach focuses on the “direct” versus “indirect” effects of enforcement policies and practices.
Specific Laws and Policies Versus the Overall Enforcement Climate
Some studies seek to evaluate the effects of specific laws and policies, while others aim to unveil the effects of the overall enforcement climate. Many of the studies that evaluate the effects of specific laws and policies have focused on those adopted by state and local governments (for an overview, see the section on immigration enforcement federalism in this article). For example, Scott Rhodes and colleagues (2015) analyze the relationship between 287(g) agreements and health care among Hispanic/Latino and non-Hispanic/Latino groups in North Carolina. The authors find that delaying and receiving poor-quality prenatal care was relatively common among Hispanic/Latina women, although the adoption of a 287(g) agreement did not significantly change care use or quality (Rhodes et al. 2015). This and other family-impact studies thus highlight immigration policy as a “de facto family policy,” if not a de facto family separation policy (Rangel and Peck 2022, 189).
A second group of studies examines how the more diffuse enforcement environment shapes immigrant families. Scholars measure the overall enforcement climate in various ways. To assess how the enforcement environment generates parent-child separations among Central American immigrants, Catalina Amuedo-Dorantes and colleagues (2015) consider the joint effect of the 287(g) program, Secure Communities, and state omnibus bills. Their study results show that these initiatives collectively increase the rate of separations and that “deportees with children left in the United States are more likely to report the intention to remigrate compared with deportees with no children and deportees who did not leave young children back in the United States” (Amuedo-Dorantes et al. 2015, 1845). Other studies examine immigration enforcement climate at a broader level than formal laws and policies. For instance, Ashley Walsdorf and colleagues (2022) explore how Latinx immigrant parents in the greater Washington, DC, area view the overall exclusionary environment—defined as threats to programs such as temporary protection, the tightening of border enforcement, and nativist sentiment in the media. In examining how this environment affects the health and well-being of their adolescent children, the study finds that parents “holding a more precarious status … describe their adolescent children as being more adversely impacted by immigration changes than do parents who have greater security by virtue of being permanent residents or citizens” (Walsdorf et al. 2022, 59).
Finally, some family-impact studies track changes in the enforcement climate over time or analyze specific waves of enforcement action that have punctuated different time periods. Catalina Amuedo-Dorantes and Esther Arenas-Arroyo (2018) interrogate how an evolving sociopolitical context affects Hispanic immigrant families with an index that accounts for the “intensity” of overall enforcement policies between 2001 and 2015. This study finds that the higher the index score, the greater the proportion of Hispanic youth in foster care (Amuedo-Dorantes and Arenas-Arroyo 2018, 372). In a rare study that compares the experiences of Latino and non-Latino families, Sandra Candel and Shahla Fayazpour (2019) explore how Mexicans and Iranians navigated the post 9-11 and Trump-era “zero-tolerance” policies regulating immigration. This study shows that whereas Mexican immigrant families were most strongly affected by the fear of or actual deportation, their Iranian counterparts were more preoccupied with US travel bans from Muslim countries and their children’s uncertain socioeconomic prospects.
“Direct” Versus “Indirect” Effects
Individual immigrants and their family members experience enforcement action with varying degrees of contact with the enforcement system. Thus, the second key approach in family-impact studies attends to the “direct” versus “indirect” effects of enforcement on (mostly Latino) families. There is no uniform definition of “direct” and “indirect” effects that scholars have adopted, but one useful conceptualization is Joanna Dreby’s (2012) classification system. Drawing on a four-year-long ethnography of Mexican immigrant households and the effects of enforcement on Mexican children, Dreby (2012) proposes what she refers to as a “deportation pyramid.” The deportation of a parent is situated at the apex of the pyramid, representing the most acute, permanent, and arguably “direct” form of immigration enforcement. The bottom of the pyramid reflects less conspicuous experiences of enforcement shared across a larger population of immigrant families that can be viewed as an “indirect” effect (for example, the threat of deportation, and fears about family stability). Under this conceptualization, actual arrest, detention, or deportation can be considered a “direct” effect, whereas the fear or threat of enforcement action can be considered an “indirect” effect.
While distinguishing between “direct” and “indirect” effects can be a helpful heuristic device, this distinction comes with limitations. Though most family-impact studies tend to address direct or indirect effects independently, these effects are closely intertwined and can interact in ways that compound experiences of insecurity and precarity among immigrant families. Indeed, different forms of enforcement “often blur and build on each other, either because they remain unresolved or because they prompt a second or third episode” (Dreby and Macias 2023, 119). As Mirian Martinez-Aranda (2020, 757) writes, “Damage to families emerges when the looming, ever-present threat of deportation becomes more intense and tangible at the moment a loved one is detained … and their relatives increasingly fear deportation.” Moreover, direct and indirect effects are not static categories defined by a uniform set of consequences. For example, the same type of enforcement action may influence a child differently depending on its intensity, duration, and when and where it took place in terms of the child’s development. One study of immigrant families finds that, for younger children, a lack of comprehension and awareness may mitigate some of the harmful repercussions of parental deportation even as the effects of parental separation linger in the long term (Dreby et al. 2022). Another study finds that, within Mexican mixed-status families, caretaking responsibilities often fall on older siblings after parental detention or deportation (Londoño et al. 2022).
Effects on Four Main Domains of Family Life
In this section, we discuss studies that evaluate the direct and indirect impact of specific policies and programs or the overall immigration climate. These studies examine how immigration enforcement shapes family and household dynamics, meanings, roles, and practices across four key domains (see Menjívar et al. 2016): familial relationships, economic and material conditions, health and health care, and education. These categories are overlapping, and some studies assess the effects of enforcement across multiple domains. For example, in their study of Mexican families with a father deported from the United States, Victoria Ojeda and colleagues (2020) note that deportation generates insecurities that compromise mental and physical health, academic performance, financial stability, employment, social networks, and family life. Moreover, as we will highlight, effects in one domain of life can initiate a cascading effect that influences life in other arenas. For example, an analysis of Latino adolescents in Southern California by Kristina Lovato (2019) shows that the deportation of a parent erodes children’s psychological health, which subsequently impedes their academic performance.
Family Dynamics and Relationships
Family separation—both potential and actual—remains a prominent theme in family-impact studies. Some research helps illuminate the consequences of family separation from the perspective of family members who are arrested, detained, or deported. For example, Abigail Andrews and Fátima Khayar-Cámara (2022) study evolving fathering practices following men’s deportation to Tijuana, Mexico. They find that “men deported directly tended to remain undeterred, while those who faced complex deportations began to acquiesce to the barriers to parenting, and those deported from jail or prison resigned themselves to the institutional obstacles to fatherhood” (Andrews and Khayar-Cámera 2022, 706). More common, however, are studies that focus on the experiences of left-behind family members. Much of this work traces how family separation damages family relations and stability. For instance, Kristina Lovato and Laura Abrams (2021, 39) find in their study of Latinx students in Southern California that after a parent is deported, families “(a) modified family structures and relied upon extended familial support, (b) experienced familial tensions, (c) experienced financial difficulties and housing instability, and (d) retreated from social life, including drastically shrinking social networks.” Because deportees tend to skew male, family separations often create new or deepened preexisting strains between “suddenly single mothers” and their children. For example, one fifteen-year-old interviewee in Lovato and Abrams’ study was left to shoulder the responsibility of caring for her younger brother after her father’s deportation forced her mother to find paid employment.
While actual family separation destabilizes family life and relationships, many scholars contend that private and public discourse in Latino communities are so saturated with discussions about immigration enforcement, legal status, and the like, that the risk of family separation is also a significant source of pervasive insecurity and anxiety for many Latino families. Studies show that deportability can play a powerful role in shaping parenting practices and household dynamics. To take one example, Fernanda Lima Cross and colleagues’ (2022, 569; see also Lykes et al. 2013) Michigan-based study finds that the fear of deportation has the potential to recalibrate how unauthorized Latinx mothers engage with their children; whereas some women chose to withhold information about, or abstained from discussing, the topic of deportation with their children, others felt obliged to “prepare their children for possible family separation and the necessary steps to reunification.” In Shannon Gleeson’s (2015) study of workplace exploitation among Latinos in California, she notes that the pressure to provide for one’s family and retain paid employment suppresses the extent to which Latino workers advocate for their labor rights, as well as how openly they discuss mistreatment in the workplace with family members. Deisy Del Real (2019) shows how the environment of fear and uncertainty create conditions for the abuse and exploitation of undocumented immigrants by their relatives, intimate partners, and friends who have legal status.
Finally, an important strand of family-impact studies highlights how families can serve as sites of resilience, strategic decision-making, security, and affection. Blanca Ramirez (2024) discusses how Latinx mixed-status families engage in “anchoring work,” a family maintenance strategy that aims not only to prevent the deportation of a relative but also to mitigate the repercussions associated with their potential removal. For example, families employed “strategic secrecy,” withholding details about the criminal record of a parent at risk of deportation from their children to shield them from additional anguish (Ramirez 2024). In line with these studies that emphasize the protective role of families, Alan Dettlaff and colleagues (2009, 781) show that Latino “immigrant families are significantly less likely to experience high family stress, actively use drugs, or exhibit poor parenting skills than their native-born peers.” However, even supportive family relationships can be fraught with complications for immigrants. When children engage in “legal brokering” to help older, undocumented family members make sense of immigration law and advocate for their rights (Delgado 2020), “this burden can be disruptive to family dynamics and challenge normative power relations between parents and children” (Gleeson 2015, 32). Likewise, Jennifer Kam and colleagues (2024) find that while undocumented college students’ communications with their families about their legal rights as immigrants are associated with higher levels of felt safety, some types of knowledge gained through these communications are also indirectly associated with increased levels of anxiety.
Economic and Material Conditions
A major strand of the family-impact literature examines immigration enforcement’s material consequences, which are often severe and long term in nature. Families face numerous economic disadvantages after a relative is detained or deported. As William Lopez and colleagues (2022, 63) find, “the most immediate repercussion of worksite raids is the loss of employment” previously held by the affected family member. Ajay Chaudry and colleagues’ (2010) study of communities that experienced raids in six sites across Arkansas, California, Florida, Iowa, Massachusetts, and Nebraska reports that the incomes of (mostly Latino) immigrant families contracted by half or more on average in the months following a relative’s deportation and that some families still had trouble purchasing food twelve months later. Many families also incur thousands of dollars in legal fees and other costs related to attempts at family reunification, which swiftly deplete their savings (Boyce and Launius 2020; Chaudry et al. 2010). Further disruptions to household finances occur when deportation removes the former breadwinner, leading their (usually female) spouse to “[work] more hours or [take] additional jobs, limiting time with their children… to compensate for lost income” (Lopez et al. 2022, 63).
Studies find that immigration enforcement has detrimental financial and material repercussions for Latino families even when family members are not detained or deported. Subfederal immigration enforcement programs are notable in this regard. Stephanie Potochnick and colleagues’ (2017, 1047) study finds that “Mexican non-citizen households residing in a metro-area that adopted 287(g) were 10 percentage points more likely to experience food insecurity than their peers in metro-areas without 287(g)”—an association that did not extend to the Hispanic US citizen population. Likewise, Jacob Rugh and Matthew Hall (2016, 1068) show that counties with 287(g) agreements were more likely to see substantial housing foreclosures among Latinos compared to White and Black residents, especially in places “where more immigrants are identified for deportation and where a higher share of the undocumented population resides in owner-occupied homes.” Cecilia Ayón’s (2018) study of Latino immigrants in Arizona highlights the barriers created by the Legal Arizona Workers Act, which obliged employers to verify a job candidate’s immigration status before hiring them. This requirement not only exacerbated immigrants’ ability to find employment, especially when undocumented, but also effectively relegated workers to lower-wage and more exploitative jobs.
A smaller body of research illuminates how financial insecurity resulting from immigration enforcement affects family relations. Lopez and colleagues (2022) explain how deportations can disrupt the relationship between immigrants in the United States who send remittances to family members in their countries of origin and receivers of those remittances: to pay for an attorney for a relative apprehended in a workplace raid, some respondents in their study requested that family members in their home country return some of the money they had previously received in the form of remittances, while other respondents had to stop sending remittances altogether. In addition, the inability to locate gainful employment in the United States because of immigration enforcement policies can itself spur family separation, compelling immigrant parents in the United States to return to their country of origin to work while their children remain in the United States (Rabin 2018).
Contributing to this body of research, Ajay Chaudry and colleagues (2010) point to how religious, nonprofit, and other community groups stepped in following a workplace raid to collectively initiate “disaster-relief operations.” In Postville, Iowa, these groups raised approximately $1 million to assist affected Latino families (Chaudry et al. 2010). Still, the authors note that such grassroots efforts were inherently limited—not only because of the challenges in addressing a series of disparate needs but also because of difficulties tied to organizing and sustaining the humanitarian response for as long as these needs persisted (Chaudry et al. 2010). Ironically, Chaudry and colleagues’ (2010) account also suggests that community assistance is contingent on the scale and publicity of workplace raids themselves; as a result, families living in places where fewer immigrants were impacted by a raid received less support despite facing similar challenges.
Health and Health Care
While immigration enforcement has relatively immediate effects on financial security and the material conditions of immigrant households, health effects generally unfold more gradually over time. Applying an “embodiment of racism” approach to “examine how racial and ethnic social hierarchies shape people’s biology, or ‘gets under their skin,’” Airín Martínez and colleagues (2018, 189) show that for Mexican mixed-status family members living in Arizona after the passage of SB 1070,30 fear of deportation produces oral inflammation—“an indicator and predictor of oral health, immune function, and chronic disease.” Among Latino parents, even simply knowing someone who has been deported increases the odds that “their child has been referred or diagnosed with a developmental disorder,” regardless of that parent’s immigration status (Vargas and Benitez 2019, 1164).
The effects of immigration enforcement—both actual and potential—on mental health is a particularly robust line of inquiry in the family-impact literature. Family separation can spark a range of mental health issues—both for the deportee and for their family members who continue to reside in the United States (for a review, see Naseh et al. 2024). Victoria Ojeda et and colleagues (2020, 10) observe this dual impact in their investigation of Mexican immigrant fathers’ deportations, which not only generated lasting “sadness and depression … chronic crying, anger and resentment, and even a suicide attempt” among their children, but also feelings of inadequacy among deportees who could no longer perform their parenting role. Another important insight from this literature is that the mental health of immigrants often suffers because of anxieties about the safety, security, and well-being of other family members—efor example, children concerned about parental deportation (Gulbas et al. 2016; Lieberman et al. 2020) or parents distressed about the discrimination that their children face at school owing to their race or Spanish-language use (Barajas-Gonzalez et al. 2022b). For example, Kathleen M. Roche and colleagues (2018, 527) surveyed Latino parents living in a large city in the mid-Atlantic and found that concerns about family separation and the overall anti-immigrant climate, as well as interactions with immigration authorities, significantly increased the likelihood that respondents developed “high psychological distress”—measured as “depression, anxiety, and somatization.”
Family-impact studies show that parental health can shape not only the cognitive and behavioral development but also the psychological and physical health of their children (see, for example, Barajas-Gonzalez et al. 2022a; Koball et al. 2015). For instance, Florencia Torche and Catherine Sirois (2019, 30) indicate that the mere prospect of a tougher immigration reform can harm children in utero. Employing data on live births, the authors find that exposure to the adoption of Arizona’s 2010 immigration enforcement bill, SB 1070, during pregnancy (measured by references to the bill in newspapers and Google searches in the state) “had a negative effect on birth weight among Latina immigrant women” whereas their Black, White, or non-immigrant Latina counterparts remained unaffected. Nicole Novak and colleagues (2017) find in their study of a major immigration raids in Postville, Iowa, similar effects for infants born to not only Latina immigrant mothers but also US-born Latina mothers. Not surprisingly, health effects on children vary by their age and stage in the life course (Chaudry et al. 2010; Lieberman et al. 2020; Lovato 2019; Roche et al. 2018). For example, Joanna Dreby and colleagues (2022) draw on interview data with young adults raised in both Latinx and non-Latinx immigrant families in New York to find that children ages six to fourteen experienced the most severe mental health effects following an enforcement interaction. Dreby and colleagues (2022, 726) explain that whereas younger children lacked awareness of enforcement and older children could more easily manage these interactions, those in the middle age range “were old enough to remember them and for parents to request [the children’s] help, but children felt bewildered and confused and unsure of what to do.”
Finally, research suggests that legal precarity obstructs health care access. In particular, fear of immigration enforcement leads immigrants to avoid interactions with medical professionals. For example, Edward Vargas (2015, 88) finds that, all other variables held constant, Medicaid use is higher among Mexican mixed-status families compared to US-born White families; however, once deportation risk is incorporated into the model, Medicaid use among Mexican mixed-status families declines, suggesting a “chilling effect” for undocumented immigrants. In some cases, Latino immigrants, such as Indigenous Guatemalans in rural Kansas, avoid contact with medical professionals to treat the health conditions that fear of immigration enforcement often triggers (Gómez Cervantes and Menjívar 2020). As with other enforcement effects, scholars also locate spillover health effects that extend beyond the unauthorized population. Investigating how the wider enforcement environment influences the psychological health of Latino immigrants with adolescent children, Kathleen Roche and colleagues (2018, 529) find that, “[although] parental worries and behavior modifications tied to immigration actions and news were least prevalent among US citizens, pernicious immigration-related consequences were by no means limited to the undocumented”; even non-US citizen parents with legal status evaded health care and expressed concern for their families, fearing for their more legally vulnerable members.
Education
Studies suggest that immigration enforcement has detrimental effects on immigrant children’s academic performance and attainment, which have potentially long-term implications given the association between educational achievement and health, as well as employability. Carolyn Heinrich and colleagues (2023) found higher rates of school absences among Hispanic students in the aftermath of a 2018 workplace raid in Morrison, Tennessee, as these students grew unwilling to attend classes owing to the fear of family separation. For those who did return to school, interviews with school administrators revealed that “meeting children’s educational needs took a backseat to addressing their emotional and basic care (food, safety) needs” in the wake of the raid (Heinrich et al. 2023, 366). Likewise, J. Jacob Kirksey and colleagues (2020) examined the effect of deportations on achievement among Latino, Asian, Black, and White elementary and secondary school students in all public school districts in the United States. They found that the higher the number of deportations within a twenty-five-mile radius of the school, the larger the White-Latino gaps in math achievement and number of school absences.
Studies also suggest that even in the absence of actual immigration enforcement, both specific policies and the overall immigration climate cultivate the fear of deportation or deportability and intensify anti-immigrant sentiment, which in turn can undermine children’s schooling. For example, during and after the 2016 election of Donald Trump, R. Gabriela Barajas-Gonzalez and colleagues (2022b, 8) found that Latinx children of undocumented immigrants in New York City “felt unsafe and had difficulties focusing in school [and] experienced a range of separation anxiety symptoms including worry about being separated from loved ones and fear of being away from home.” Latina immigrant mothers also reported that children suffered from heightened anti-immigrant (and especially anti-Mexican) discrimination and bullying at school (Barajas-Gonzalez et al. 2022b). In a rare study comparing how the adoption of the Secure Communities program affected children’s educational development across different racial groups, Laura Bellows (2019) find that Hispanic students struggled in English Language Arts, whereas non-Hispanic Black students struggled in English Language Arts and math. However, the study finds no change in White students’ performance.
Though most research on children’s education in immigrant families captures the effects of immigration enforcement on young children, some scholars spotlight the distinct challenges that high schoolers or college applicants face. For instance, Cassaundra Rodriguez argues that Latino citizen college applicants in Los Angeles County experienced “legal violence” (Menjívar and Abrego 2012) and “multigenerational punishment” (Enriquez 2015) when filing for financial aid; Rodriguez (2019, 714) explains that during the application process, parental undocumented status becomes particularly salient, as “illegality enters the lives of citizen young adults in mixed-status families.” Providing insight into the (limited) potential of inclusionary immigration policies, Daniel Corral (2021) compared Hispanic undocumented students with their US-born peers to measure the impact of sanctuary policies on educational attainment across the two groups. Though as of 2017, 128 US counties had adopted sanctuary policies, Corral (2021) did not find that these policies significantly affect the likelihood of graduating high school or pursuing a college degree. Corral (2021, 10) concluded that “although sanctuary policies counteract immigration enforcement, for the most part, they have minimal associations with educational attainment.”
In a related vein, an emerging body of research suggests that schools can serve as a critical locus of community support for immigrant families against the backdrop of a hostile immigration environment. During a workplace raid in one community in Tennessee, for example, Heinrich and colleagues (2023) describe how schools worked to support affected students and their families in a variety of ways. At the same time, this study also emphasizes the degree to which schools and their staff remained ill-equipped to address the large-scale repercussions of workplace raids. Interviews revealed not only that staff lacked the training and capacity to adequately assist all students and families but that they were unable to counter “the chilling effects of immigration enforcement activities [that] deterred children and families from attending healthcare appointments and seeking other services and supports” (Heinrich et al. 2023, 366).
Policy Implications and Directions for Future Research
Family-impact studies make several significant advances in the study of immigration enforcement and extend the existing body of research on collateral consequences of contacts with the criminal justice system (see, for example, Kirk and Wakefield 2018). This literature makes strides in illuminating how immigration enforcement shapes the lives and opportunities of immigrant families in the United States and how these effects reach far beyond the individuals directly targeted by immigration enforcement. As we noted, family-impact studies are predominantly focused on Latino communities. Some of these studies challenge the understanding of Latino populations as a homogenous group, raising questions about the “potential differences that may exist between immigrant and native-born families,” as well as immigrants with and without legal status (Dettlaff et al. 2009, 776). Further, recent studies highlight heightened vulnerabilities among Indigenous vis-à-vis mestizo Mexican and Central American immigrants (Gómez Cervantes and Menjívar 2020).
Yet, much work remains to be done in this area of research, especially as it relates to questions of racial disparities and bias. More research is needed to disentangle the effects of immigration enforcement within the Latino population, as Latinos who are racialized as Indigenous or Afro-descendent may experience the same system quite differently than those who are racialized as White (Adler 2006; Gómez Cervantes, forthcoming). In addition, scholars must broaden their target of inquiry beyond Latinos. For example, while some studies analyze Latino groups versus non-Latino groups (see, for example, Rhodes et al. 2015), analyses of the experiences of Black, Asian, or Middle Eastern immigrant families with US immigration enforcement (for example, Candel and Fayazpour 2019; Twersky 2019) might reveal that racialization processes or racial bias against those groups manifest or operate differently than against Latinos.
Similarly, the effects of immigration enforcement on families might vary across different racial groups even when they are subjected to the same facially race-neutral enforcement policies and practices. This might be the case for several reasons. First, perceptions of racial bias are likely to vary across racial groups given their varying history of migration and context of reception. These differences in perceptions of racial bias, in turn, can shape the coping strategies of individual immigrants and their family members (see, for example, Ayón 2018; Cardoso et al. 2021; Cavanagh and Cauffman 2015). Second, immigrant communities’ resources and mobilization capacities, as well as origin-country responses to their treatment in the United States, vary across different racial groups (see Benavides et al. 2021; Ramirez 2024); these differences mean that the same set of facially race-neutral enforcement policies and practices can have varying effects on families from different racial groups.
Another important line of inquiry for future research concerns whether and to what extent race might intersect with other demographic characteristics to shape how immigration enforcement impacts family life and family relationships. For example, Rachel Adler (2006) finds in her study of a neighborhood in New Jersey a “Latino ethnic hierarchy” based on national origin, race, social class, and legal status, pointing to the intra-Latino racialization of Indigenous immigrants (Gómez Cervantes, forthcoming; Menjívar 2021). An emerging body of scholarship also suggests that gender and age are important axes that generate variations in the effects of immigration enforcement across different family members within Latino communities (see, for example, Dreby 2015; Dreby et al. 2022; Menjívar and Gómez Cervantes 2025; Rodriguez 2019). Thus, a fruitful avenue of research in this area is to explore whether these and other demographic characteristics might operate in the same or different ways for family members of various racial groups. Geography and context matter as well (Roubenoff et al. 2023), as research has found variation in how the same facially neutral enforcement policies have dissimilar gender effects on Latino immigrants (Simmons et al. 2021) across states. When researchers examine these types of complex interactions between race and other demographic characteristics, leveraging such research methods as experimental lab or field study designs (which are less commonly used methods in this area of research) may prove to be especially valuable. These methods can generate not only new insights about the nature and extent of racial disparities but also the mechanisms that produce such disparities.
Finally, while many studies have examined how immigration enforcement negatively affects the lives of immigrant families, few studies address the extent to which progressive reforms to the immigration enforcement system may help to neutralize or mitigate these consequences (see Corral 2021 for an exception). In addition, no study to our knowledge considers how such initiatives may potentially improve the conditions for different groups of immigrant families other than Latinos. Similarly, while several studies examine how nonprofit and nongovernment organizations and community groups provide a partial social safety net for Latino communities affected by immigration enforcement (Dettlaff et al. 2009; Hagan et al. 2010; Koball et al. 2015), we lack a systematic understanding of the limitations and constraints that these support networks face, and whether and to what extent such networks might exist for non-Latino families. Investigating these issues remains an urgent task that will inform efforts to develop reform measures and support networks that can advance family unity and well-being across different immigrant communities.
CONCLUSION
Our review of the existing literature on the criminalization of immigration reveals three key themes. Across these themes, research shows that race and racialization play an important role at multiple layers and in various components of the immigration enforcement system to disproportionately affect immigrants racialized as non-White in the United States. Thus, it is unsurprising that a great deal of the existing scholarship that we have reviewed focuses on Latino communities, individuals, and families. For example, the growing body of research on the collateral consequences of immigration enforcement uncovers a multitude of negative effects on various aspects of family life in Latino communities.
The first important theme that emerges from our review of the literature on criminalization of immigration is that, although immigration enforcement is generally a matter of federal civil law, immigration enforcement has become substantially intertwined with criminal law enforcement in ways that have broad-ranging and radiating effects on immigrant communities of color. State and local law enforcement agents participate in immigration enforcement efforts, both at the request of federal officers and under their own authority and subfederal initiatives. Databases that store, link, and disseminate information about individuals for law enforcement purposes (whether criminal or immigration) are interconnected or simply combined. State and local prisons and jails hold civil immigration detainees. Companies that provide alternatives to detention in the criminal context do the same for individuals in civil removal proceedings, and risk-assessment tools designed for release determinations in the criminal context have been adapted to manage similar determinations in the context of civil immigration detention. The federal government increasingly relies on federal criminal law prosecutions to enforce immigration laws, and these prosecutions contribute to making criminal subjects out of immigrants.
Second, race and racialization play an important and bidirectional role in the increasing criminalization of immigration. As we explained earlier, criminalizing immigration fosters the racialization of certain immigrant groups as the dangerous other, and such racialization in turn legitimates the policies and practices that target or disproportionately impact those immigrant groups. Studies suggest that the racial makeup of a location may affect state and local political decisions as to whether and how to cooperate with (or even amplify) federal immigration enforcement efforts. The increasing role of state and local law enforcement in immigration enforcement may lead to more racially discrepant policing against Latinos. In addition, the expanding role of state and local law enforcement may decrease the trust that local immigrant communities have in policing agencies, which in turn increases the rate of crime victimization in those communities. Moreover, because of the data sharing between law enforcement agencies and immigration enforcement agencies, racialization in ordinary criminal policing—including the gang designations or surveillance and monitoring that disproportionately target Black and Latino residents—ultimately affects (and is affected by) immigration policing. Although the evidence concerning the effects of race on sentencing outcomes are mixed, it is clear that federal criminal prosecutions are overwhelmingly of Latinos. Racial bias also appears to play a role in the use of algorithmic risk-assessment tools in immigration detention bond determinations, as well as in individualized release determination by immigration judges.
Finally, the effects of the criminalization of migration in the United States are not limited to the United States. Immigration enforcement programs have been externalized, and at the same time, data sharing and status determinations about migrants (including the affixing of gang labels or communications concerning criminal records) are both exported to, and received from, foreign governments. This externalization has racializing effects, as certain migrant populations (here, our review of the scholarship has focused on those moving through Mexico) are racialized in these processes and subjected to violence (in this context, racialized violence) during their migration journeys. This externalization of migration control follows the United States’ exportation of certain highly punitive features of its criminal legal systems (McLeod 2010). Externalization adds a distinct new layer to this phenomenon, enabling the United States to use other legal systems and enforcement actors as a primary means of effectuating its own migration control goals. The result is an intersecting regime of crime and migration control where individuals’ records and identity markers can serve as the basis for their exclusion and punishment by a wide array of transnational actors.
Significant gaps and open questions require additional research. First, as we noted earlier, the targeting of Latino immigrants and communities for immigration enforcement has produced specific harms on this population; as a result, empirical work has focused on capturing the effects of immigration enforcement or restrictionist policies on Latino communities. There is some work on other racial groups in specific contexts. However, on the basis of this review, we argue that much more scholarly work is needed to substantiate racial disparities in enforcement (or even to document “racially discrepant enforcement”) and to understand the differential effects of immigration enforcement policies—and the criminalization of immigration—across racial groups.
Second, we note that in the extant scholarship that we have reviewed, “Latinos” are mostly treated and understood as a homogenous racial group (with some, but not a lot of, attention to gender or national-origin differences within this heterogenous group). More careful attention to racial and other differences under the umbrella term Latino is needed in empirical studies of immigration enforcement policies and studies concerning the criminalization of immigration. The incipient scholarship that examines differences within the Latino population points to important differences in experiences with immigration enforcement between Indigenous and Latinos racialized as White (Gómez Cervantes 2021).
Third, further study could also enhance our understanding of why and how state and local agencies implement immigration enforcement initiatives, how those choices are shaped by racial bias, and how racial dynamics shape both these initiatives and the experience of these initiatives in communities. Although concepts like racial threat are useful in understanding how perceptions and racial biases are seen by the nonimmigrant local population, more research is needed to better theorize this concept and to better understand whether and how racial threat shapes immigration enforcement policies. Similarly, studies need to attend to refining theories concerning the benefits of diversification of law enforcement agents, particularly as it pertains to trust among, and treatment of, racialized immigrant communities. Relatedly, more work is needed to understand how and why racial minorities perceive and experience local immigration enforcement policies and practices as racially discriminatory and biased.
Our review of the scholarship on the criminalization of immigration suggests the urgent need for policy and regulatory changes in the governance of migration. When evaluating the benefits of immigration-enforcement initiatives, including cooperative policies with state and local governments, policymakers should take into account the costs of enforcement policies on broader populations, including citizen and lawfully present immigrant residents, who experience many of the negative consequences of increased enforcement efforts. Second, policymakers should attend to, and seek to interrupt, the ways that racial bias and racialization practices can shape immigration enforcement choices, whether that is through uncritical acceptance of “gang” designations made by individual officers or through reliance on the dangerousness designations made by governments abroad.
Finally, our review underscores the critical importance of improving government transparency across all levels of immigration enforcement and adjudication agencies. For example, empirical studies of immigration surveillance, dataveillance, electronic monitoring, and risk-assessment policies have been impeded by the opacity of these systems and data inaccessibility to researchers and the public. Likewise, little empirical research attends to racial disparities in immigration court proceedings and outcomes owing to data scarcity, which the government is in a unique position to address. In brief, greater governmental transparency is needed not only to allow for systematic investigations of the efficacy and effects (including racially disparate impacts) of these surveillance and adjudication systems but also to ensure democratic accountability.
FOOTNOTES
↵1. The racial category of “White” (as with all other racial categories) has always been “contingent, changeable, partial, inconstant, and ultimately social” (Haney López 2006). The US Office of Management and Budget (OMB) used to define “White” as “a person having origins in any of the original peoples of Europe, the Middle East, or North Africa” (US Census Bureau 2022). However, in March 2024, the OMB published a new definition of “White” as “individuals with origins in any of the original peoples of Europe, including, for example, English, German, Irish, Italian, Polish, and Scottish” (US Census Bureau 2024). According to the OMB, “people who identify their origin as Hispanic, Latino, or Spanish may be of any race” (US Census Bureau 2022, 2024). Reflecting the fluidity of Latino identity, the self-identification of the Hispanic or Latino population as “White” alone decreased from 47.4 in the 2010 Census to 17.6 in the 2020 Census (US Census Bureau 2023).
↵2. Although under the Immigration and Nationality Act, 8 U.S.C. § 1101(a)(15), “immigrant” is a term of art that has a specific narrow legal meaning, we use this term more broadly in this article to refer to non-US citizens.
↵3. We do, however, take up some of these studies in the section on the effects of immigration enforcement on immigrant families.
↵4. For a review of the empirical literature that considers both subfederal enforcement and sanctuary policies, see Boyke and Horner 2024.
↵5. For an analysis of the evolution of the 287(g) Program over time, see American Immigration Council 2021a.
↵6. As of April 2025, 86 law enforcement agencies in 25 states operate under the jail enforcement model, 180 law enforcement agencies in 27 states operate under the warrant service officer model, and 190 law enforcement agencies in 23 states operate under the task force model (US ICE 2025).
↵7. A study by Bianca Wirth and Eric Baumer (2024) is a notable exception. They find in their study of factors associated with 287(g) Program adoption between 2007 and 2009 that racial diversity among police forces is not a significant predictor. However, Wirth and Baumer (2024, 371) note that this result may be merely a reflection of the limited racial diversity found in the police agencies sampled. As they note, “Very few of the law enforcement agencies sampled reach a critical mass of officers from racial-ethnic minority groups.”
↵8. Qualitative research on the everyday lives of Latino border enforcement agents presents a complex picture in which even as these agents are “inclined toward a more humane approach to immigration control than their white counterparts” (Vega 2018, 2558), they also “function as both enforcers and targets of racial power” (Correa and Thomas 2015, 242; see also Heyman 2002).
↵9. Before 2008, Ecuador used to extend multi-entry visas to practically all countries, including the Global South. In 2018, all visa requirements were eliminated by a presidential decree, but this was short lived. In 2010, little by little and in selective fashion, visa requirements were reintroduced. Since then, visa requirements have been in place for 37 countries, almost all in the Global South. The reintroduction of visas occurred in the context of externalization, as a strategy of containment of migratory flows through Ecuador.
↵10. Using a wide array of government data, Martha Rojas-Wiesner (2022) provides a chronology of various periods of externalization history in migration through Mexico.
↵11. Precedent may be found in offshoring policies to intercept, detain, and deport seafaring migrants, especially Haitians and Cubans, in the post-Mariel period (Loyd and Mountz 2018; Shull 2021).
↵12. The racialization of Central American migrants in Mexico as dangerous is not new, however, as historically they were associated with Central American guerrillas and civil turmoil across the continent (Minian 2020; Shull 2021).
↵13. It is also worth noting that the US Coast Guard focuses its activities largely to stop Haitians, Cubans, Dominicans, and Chinese nationals who travel across the Caribbean to reach the United States (Ryan 2010).
↵14. All of these practices supplement, rather than replace, physical inspection by humans and canines.
↵15. The term authorized workers refers specifically to individuals who are authorized to work under the federal Immigration Reform and Control Act. This population includes citizens, lawful permanent residents, other individuals in the United States on visas that authorize their employment, and individuals legally authorized to work by the DHS secretary even though they lack legal immigration status (such as some individuals who have been paroled into the US for humanitarian purposes). Not all immigrants who are authorized to be present in the US are authorized to work in the United States, and not all immigrants who are authorized to work in the United States are “documented” immigrants. Nevertheless, because most empirical studies fail to acknowledge this distinction—perhaps because there is substantial overlap between the categories of unauthorized workers and undocumented immigrants—we will continue to refer to the population of unauthorized workers as “undocumented immigrants,” with the caveat that the term in this context is both over- and under-inclusive.
↵16. On January 20, 2025, the Trump administration ended the use of CBP One and cancelled all asylum interviews that had been scheduled using that application (Aleaziz and Villegas 2025).
↵17. A growing literature examines the subjective experiences of land border crossers, including cross-border commuters, along the US-Mexico border (Castañeda Perez 2020; Chávez 2016). These studies might fruitfully be compared with and explored alongside the literature exploring how different populations experience and respond to negative domestic street policing incidents.
↵18. Racially discriminatory treatment of travelers predates current technological screening tools (Carbado 2023, 140–43).
↵19. A report based on a review of a limited body of public records finds that Black immigrants were subjected to electronic monitoring at significantly higher rates than the members of other racial groups (Giustini et al. 2021), but empirical study could enhance our understanding of the costs, benefits, and discriminatory applications of EM.
↵20. Even in states that mandate E-Verify, fewer than 50 percent of employers in those states are actually enrolled in the program (Orrenius et al. 2020).
↵21. The Department of Justice’s Civil Rights Division, Immigrant and Employee Rights Section, enforces these provisions (Perotti 1992).
↵22. In 1996, the Illegal Immigration Reform and Immigrant Responsibility Act of 1996 (IIRIRA), replaced “deportation” proceedings and “exclusion” proceedings with “removal” proceedings (Executive Office for Immigration Review 2022, 61).
↵23. The key types of nonjudicial removals are expedited removal and reinstatement of prior removal orders. In fiscal year 2021, about 42 percent of removals from the United States were expedited removals, and 35 percent were based on the reinstatement of prior removal orders (Leong 2022).
↵24. The master calendar hearing is the first hearing that an immigration judge holds during the removal process. These hearings are held for pleadings, scheduling, and other similar matters (Executive Office for Immigration Review 2022, 69). During merits hearings, also known as individual calendar hearings, immigration judges decide challenges to removability and applications for relief from removal (Executive Office for Immigration Review 2022, 78).
↵25. Our review and analysis of the existing literature primarily focuses on the effects of immigration enforcement on families in the United States; however, there is an emerging body of research that examines impact on family members, such as US citizen children, who are deported or experience forced relocation and are residing in countries of origin at the time of their research participation (see, for example, Fuentes-Balderrama et al. 2024; Hamilton et al. 2023).
↵26. Other scholarship examines how individual characteristics such as sexuality, nationality, age, and socioeconomic status interact with these mechanisms to keep immigrant families apart (Jones 2023; Rodriguez 2019).
↵27. As Carlos Santos and Cecilia Menjívar (2013, 8) have observed, “the undocumented population is not a separate group easily classifiable and isolated from the documented (or U.S. citizens). Undocumented immigrants live, work, go to school, pray and play with documented immigrants as well as U.S. citizens, and in the context of the family and schools, they are often found under the same roof.”
↵28. For helpful reviews of studies on the broader question of how immigration shapes families, see Glick 2010; Van Hook and Glick 2020.
↵29. This is an important distinction, as Amy Johnson and colleagues (2024, 1) show in their recent study comparing the effects of “discrete dramatic societal events that independently signal a change to the country’s approach to deportation” versus “more gradual changes to the country’s everyday institutional … and social … environment of deportation threat.”
↵30. Passed in 2010 in Arizona, Senate Bill 1070 (“Support Our Law Enforcement and Safe Neighborhoods Act”) instituted a series anti-immigration policies targeting undocumented persons; while some measures were overturned following a 2012 US Supreme Court decision, the measure requiring state and local law enforcement to request identification documents and confirm legal status during stops or arrests (that is, the “show me your papers” provision) was upheld.
- © 2025 Russell Sage Foundation. Ryo, Emily, Jennifer M. Chacón, and Cecilia Menjívar. 2025. “Criminalization of Immigration.” RSF: The Russell Sage Foundation Journal of the Social Sciences 11(3): 282–343. https://doi.org/10.7758/RSF.2025.11.3.07. We are grateful to Denni Arnold, Jane Bahnson, Natasha Bluth, Danielle Flores, Ramón Garibaldo Valdéz, Paul Moorman, Isabella Palmer, and Maia Wenger for their excellent research assistance. Emily Ryo’s research was supported by the National Science Foundation (Award No. 2407056). Direct correspondence to: Emily Ryo, at ryo{at}law.duke.edu, 210 Science Drive, Durham, NC, 20083, United States.
Open Access Policy: RSF: The Russell Sage Foundation Journal of the Social Sciences is an open access journal. This article is published under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.
REFERENCES
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.