Citizenship and Sentencing: Assessing Intersectionality in National Origin and Legal Migration Status on Federal Sentencing Outcomes

Abstract

Objectives:

This study seeks to understand how national origin and legal migration status of noncitizen defendants in federal criminal courts shape incarceration and sentence length decisions.

Method:

The authors use annual United States Sentencing Commission (USSC) Monitoring of Federal Criminal Sentences (MFCS) datasets (2011–2016) to examine the impact of defendant’s (1) national origin and (2) legal versus illegal migration status on incarceration and sentence length decisions in federal criminal courts. In addition, in order to account for effects of immigration cases, supplemental analyses are conducted for (1) non-immigration offenses and (2) immigration-only offenses.

Results:

For the incarceration decision, noncitizen defendants have higher odds of incarceration than U.S. citizens, net of other factors. These effects are less consistent in the sentence length decision. These relationships systemically differ across national origin and legal migration status.

Conclusions:

Punishment disadvantages based on one’s citizenship are particularly pronounced for defendants from Mexico, Latin America, and Africa and especially for those with “illegal” migration status. As noncitizen populations continue to grow in federal courts and in the U.S. more broadly, understanding and addressing these citizenship disparities in punishment will be increasingly important.

Keywords

courts sentencing nationality immigration citizenship race ethnicity

Criminologists have had a longstanding interest in identifying the ways in which demographic characteristics shape punishment throughout the criminal justice system. Within the sentencing literature, researchers have focused on how a defendant’s extralegal characteristics, such as race/ethnicity, gender, and age, influence various sentencing outcomes. This body of research has consistently found that racial/ethnic minorities and male defendants receive harsher sentencing outcomes (i.e., incarceration, sentence length, departures) than similarly-situated Whites and females, respectively (Demuth and Steffensmeier 2004; Spohn 2000; Steffensmeier, Kramer, and Ulmer 1995; Ulmer and Kramer 1996). Furthermore, research suggests that these extralegal factors intersect in ways that aggravate, or enhance, the impact of race/ethnicity and gender (i.e., for Black males and Hispanic males, Doerner and Demuth 2010; Spohn 2000). Thus, despite the explicit goals of federal and state sentencing reforms to eliminate unwarranted disparities in sentencing (Kramer and Ulmer 1996; United States Sentencing Commission (USSC) 2004), research clearly indicates that disparities based on extralegal factors persist.

What is less clear is whether (and how) citizenship status influences punishment within the federal criminal court system, and relatedly, the degree to which these effects are shaped by defendant national origin and “legal” migration status. This is a pressing issue given that noncitizen populations have rapidly grown to include more than 44 million residents and account for nearly one out of every seven people in the United States (Lopez 2019). Likewise, the share of noncitizens in the federal criminal court system has grown over time, accounting for over 40 percent of the federal caseload in 2016 (USSC 2016).

Notably, a growing body of a literature has begun to examine how defendant citizenship status influences sentencing outcomes (Everett and Wojtkeiwicz 2002; Hartley and Armendariz 2011; Kautt and Spohn 2002; Light, Massoglia, and King 2014; Wu and Deleon 2012). These studies typically rely on either focal concerns (Steffensmeier, Ulmer, and Kramer 1998) or group threat theories (Blalock, 1967; Blumer, 1958) to explain why citizenship status may work as an aggravating factor at sentencing. Specifically, these perspectives suggest that non-U.S. citizenship may serve as a signal of defendant dangerousness, culpability, or practical concerns (e.g., flight risk or deportation proceedings) or may trigger perceptions of group threat among courtroom actors in much the same way that racial/ethnic minority status does (Wolfe, Pyrooz, and Spohn 2011).

Despite the growing attention to citizenship effects on sentencing, several key gaps and questions remain in this line of research. Notably, empirical research on the topic to date has produced mixed findings. On the one hand, some studies find that noncitizen defendants receive harsher sentences (especially for the incarceration decision) than similarly-situated U.S. citizens (Demuth 2002; Hartley and Armendariz 2011; Light et al. 2014). On the other hand, other studies find that noncitizen defendants (even those with illegal status) may actually receive less severe sentences than U.S. citizens, net of other factors (Wolfe et al. 2011; Wu and DeLone 2012).

Along with these mixed findings, research on citizenship and sentencing has often been limited to general comparisons of U.S. citizens versus noncitizens (Light 2014, Wolfe et al. 2011), with less attention given to whether citizenship effects differ by (1) national origin and (2) legal versus illegal migration status (and combinations of these categories). This oversight is particularly noteworthy given that the “noncitizen” category includes a wide variety of nationality groups (e.g., Mexican, South American, Asian, Western European, etc.), which have unique cultures and histories of migration. In addition, these groups have received dramatically different responses from the U.S. public and the criminal justice system. For example, public sentiment and crime-related fears of Mexican or Latin American migrants have been more negative than perceptions of Asian or Western European migrants (Logue 2009; Orrick and Piquero 2015). Likewise, the U.S. public has often perceived “illegal” or undocumented migrants as more crime-prone and threatening than “legal” migrants (Coutin 2005). Thus, as we discuss in the following sections, there are reasons to suspect that noncitizen effects on sentencing may depend greatly on defendant nationality, legal migration status, and the intersection between these categories (e.g., with particularly harsh sentences given to defendants who are from certain regions and have “illegal” status). However, research has yet to fully explore how these distinctions shape the effects of citizenship on sentencing outcomes.

To be clear, there have been several noteworthy exceptions that have begun this work (e.g., Light et al. 2014; Valadez and Wang 2017; Wolfe et al. 2011). However, these studies have been limited in a few ways. For example, they have typically examined legal migration status but given less attention to national origin (alone or in combination with “legal” migration status). In addition, prior analyses have often been limited to earlier times points (1990s to early 2000s; e.g. Demuth 2002; Logue 2009; Valadez and Wang 2017; Wolfe et al. 2011), a single offense type (e.g., drugs, non-immigration offenses; e.g. Logue 2009; Valadez and Wang 2017), or to a select set of locales (e.g., Hartley and Armendariz 2011; Wu and DeLone 2012).

In light of these issues, the goal of this study is to examine the intersectionality of citizenship, national origin, and legal migration status on incarceration and sentence length outcomes in federal criminal courts. To achieve this goal, we use the Monitoring of Federal Criminal Sentences (MFCS) annual datasets ranging from 2011 to 2016, which include detailed defendant demographic and case information.

Prior Research on Effects of Defendant Characteristics in Sentencing

Sentencing research has consistently shown that extralegal factors, such as defendant race/ethnicity, influence sentencing outcomes within both state and federal criminal court systems. Focusing specifically on studies of race/ethnicity, minority defendants are more likely to be sentenced to prison (Demuth and Steffensmeier 2004; Doerner and Demuth 2010; Mitchell 2005; Steffensmeier et al. 1998), receive longer average sentences (Doerner and Demuth 2010; Feldmeyer and Ulmer 2011; Steffensmeier et al. 1998), and receive less favorable departure decisions (Holmes, Feldmeyer, and Kulig 2020; Johnson, Ulmer, and Kramer 2008; Spohn and Brennan 2011) than similarly-situated White defendants. In addition, researchers have given increased attention to how one’s race/ethnicity, gender, and age intersect to influence sentencing. There is now a sizable body of research illustrating the intersectionality of extralegal factors on sentencing, which shows that Black (and sometimes Hispanic) defendants receive particularly harsh sentencing outcomes when they are young and male (Mitchell 2005; Steffensmeier, Painter-Davis, and Ulmer 2017). Furthermore, these racial/ethnic effects differ depending on the types of offenses committed (e.g., drug or violent; Mitchell 2005; Spohn 2000) and local community context in which sentencing occurs (e.g., racial/ethnic population composition, characteristics of courtroom actors; Johnson et al. 2008; Ulmer 2018; Ulmer and Parker 2020).

In order to explain these persistent racial/ethnic effects in sentencing, researchers have often relied on focal concerns theory (Steffensmeier et al. 1995, 1998, 2017; Ulmer 2012). Focal concerns theory posits that courtroom actors make decisions about sentencing and punishment based on three main considerations: (1) blameworthiness of the defendant, (2) potential danger to the community, and (3) practical constraints and consequences. According to this model, defendants that are perceived as more blameworthy and who seem to pose a greater potential threat to the community upon release are targeted with harsher sentences. In addition, courtroom actors may consider practical constraints and implications when making sentencing decisions, such as the defendant’s ability to “do time,” current jail or prison capacity, and other organizational issues related to sentencing processes (Steffensmeier et al. 1998).

Although judges and courtroom actors are expected to mete out punishment based exclusively on legal characteristics (see USSC 2012), focal concerns theory illustrates how attributions and perceptions tied to a defendant’s race/ethnicity (or other demographic characteristics) can indirectly influence sentencing decisions (Steffensmeier et al. 1998). Scholars argue that, due to incomplete information, time, and resources, judges and prosecutors rely on a defendant’s race/ethnicity as a “perceptual shorthand” for their potential blameworthiness, dangerousness, and practical considerations (Albonetti 1991; Steffensmeier et al. 2017). Thus, even if courtroom actors seek to administer punishment in a race-neutral manner, focal concerns theory explains how attributions and stereotypes about defendant race/ethnicity can unintentionally permeate sentencing and shape punishment outcomes.

In addition to focal concerns theory, researchers have also relied on racial threat (and related group threat) theory to account for racial/ethnic effects on sentencing and inform their analyses. Racial threat theory argues that as minority groups grow in size and gain more economic and political influence, they are perceived as greater threats to Whites (or majority groups) in positions of power (Blalock 1967). In turn, the racial threat perspective suggests that harsher social controls (including criminal justice sanctions and punishments) are imposed on minority groups as a countermeasure to protect the status quo and limit the growing influence of these groups (Blumer 1958). Although there has been mixed evidence on racial threat in sentencing (particularly for Hispanic defendants), research provides some support for this position and indicates that Black defendants receive harsher sentencing outcomes in response to growing minority presence and heightened perceptions of group threat (Feldmeyer and Ulmer 2011; see Feldmeyer and Cochran 2018).

Taken together, focal concerns and racial threat theories both provide plausible explanations for why race/ethnicity (and other demographic characteristics) could impact sentencing outcomes. As noted above, empirical findings have often supported these positions, indicating that defendant race/ethnicity (especially when combined with other case-level and defendant characteristics) is a consistent and meaningful predictor of sentence severity, net of legally-relevant factors. However, it is less clear whether the same types of disparities are found based on citizenship. Compared to the vast body of scholarship examining racial/ethnic effects on sentencing, research has given far less attention to the role that citizenship status plays in sentencing, with particularly limited attention given to the way that citizenship effects intersect with defendant nationality and legal migration status (which is the focus here). This is problematic considering that noncitizens account for nearly half (41 percent) of all federal criminal defendants (USSC 2016), making them one of the largest minority groups in the federal criminal caseload. In addition, as we review below, both focal concerns and racial threat theories suggest that noncitizen defendants may be disadvantaged in sentencing outcomes in much the same way as racial/ethnic minority defendants.

Theoretical Explanations for Noncitizen Effects

Although focal concerns and group threat theories have often been applied to the study of race and ethnicity, the same theoretical arguments provide a conceptual framework for explaining why noncitizen defendants may receive harsher punishments than their U.S. citizen counterparts. Notably, the current study does not provide explicit tests of focal concerns or group threat theories, but rather uses the two as conceptual guides to inform the analyses. As Lynch (2019) notes, although this approach is fairly common in extant sentencing research, there is also a need to include more direct measures and tests of these theories—a point we return to in our discussion section.

Drawing on focal concerns theory, judges may perceive noncitizen defendants as particularly blameworthy because they committed a crime while receiving benefits from the host nation, thus violating the host country’s hospitality (Wolfe et al. 2011). In addition, these perceptions of culpability may be magnified when noncitizens hold “illegal” migration status, indicating that they violated immigration restrictions in addition to committing a criminal offense.¹ In terms of dangerousness, noncitizens have been portrayed as more crime-prone than U.S. citizens in media and public discourse (Trager and Kubrin 2014), which may lead judges to view them as a greater potential threat and more deserving of harsh penalties (Wolfe et al. 2011). Last, in terms of practical constraints, judges may adjust sentences for noncitizen defendants based on considerations of potential flight risk and deportation proceedings (Wolfe et al. 2011). However, research suggests that these practical considerations could work in opposing ways depending on the sentencing outcome examined. Specifically, judges may be more likely to give incarceration sentences to noncitizens to reduce flight risk but give relatively short sentences knowing that deportation proceedings are likely to follow (Valadez and Wang 2017; Wolfe et al. 2011).

Group threat theory (Blalock 1967) also suggests plausible reasons why noncitizen defendants may receive harsher punishments than their U.S. citizen counterparts. Immigrant groups have been among the fastest growing populations in the United States in recent decades. The foreign-born population in the United States has risen from approximately 10 million people in 1970 (about 4 percent of the population) to approximately 45 million (14 percent of the U.S. population) by 2018 (Migration Policy Institute 2019). As a result, this rapid growth in immigrant and noncitizen populations may have contributed to growing perceptions of group threat toward these populations (Light et al. 2014). In response, courtroom actors may impose harsher punishments on noncitizen defendants to counter rising levels of perceived threat.

These perspectives provide plausible reasons why noncitizens may receive harsher sentences than similarly situated U.S. citizens. However, there are also reasons to believe that noncitizen disadvantages may apply to some groups more than others, which is the focus of the current study. That is, some nationalities and migrant groups may be seen as a greater crime risk, more blameworthy, and more threatening than others. As described above, research on race/ethnicity and sentencing has long recognized the importance of intersectionality in punishment and shown that race/ethnic effects in sentencing are magnified when combined with other legal and extralegal factors (e.g., gender, age, offenses type) (Doerner and Demuth, 2010; Steffensmeier et al. 2017). A similar intersectional relationship could be at play here. That is, noncitizens may be treated more harshly in some sentencing outcomes, but that disadvantage may be especially pronounced for defendants from certain countries and for those who are undocumented.

In terms of national origin, there has been particularly sharp public concern and political rhetoric surrounding migration from Mexico and Latin American, which was voiced in the early 2000s but became louder during the 2016 presidential election. Donald Trump’s 2016 presidential campaign relied on a narrative portraying migrants from Central and South America as “bad hombres” and as criminal and economic threats (Trump 2016). Trump referred to immigration on the Mexican border as an “invasion” and claimed that Mexico was sending drug dealers and rapists across the border (Lamont, Park, and Ayala-Hurtado 2017). To be sure, this harsh rhetoric against noncitizens focused largely on Mexico but also included a variety of other Latin American and African nations, which Trump referred to as “shithole countries” (Watkins and Phillip 2018). In light of these narratives, there is reason to expect that noncitizen defendants from Mexico and Latin America (and perhaps African nations) may be perceived as particularly dangerous, threatening, culpable, and deserving of harsher court sanctions (Carey and Marak 2018; Logue 2009). In contrast, migrants from Asian and European countries have typically been viewed more favorably in both public and political spheres, especially concerning their potential crime risks. Thus, they may not elicit the same degree of concern about criminal threat and may see greater leniency in sentencing than the previously mentioned groups.

Similarly, undocumented immigrants may also elicit greater perceptions of blameworthiness, crime risk, and threat. Public fears and harsh political rhetoric toward noncitizens have often focused on “illegal” immigrants (but with broad ramifications for all noncitizens). In addition, noncitizen defendants with “illegal” status may elicit different sets of practical considerations as judges weigh their potential flight risk and future deportation proceedings when calculating punishment decisions. Notably, prior research has shown that “illegal” migration status generally elicits more negative sentencing outcomes (Demuth 2002; Light et al. 2014). However, research has given less attention to the intersections between legal migration status and nationality to identify how these combined identities shape noncitizen effects at sentencing (a point we return to below).

In sum, the arguments outlined above suggest that “noncitizen” may not be a one-size-fits-all category and could fail to adequately capture the unique ways in which national origin and legal migration status intersect to shape sentencing. Although these extralegal factors may not be the primary determinants of sentencing decisions (Doerner and Demuth 2010; Ulmer and Parker 2020), focal concerns and group threat theories suggest that it would be difficult for courtroom actors to be completely immune to narratives and stereotypes about noncitizens when coming to sentencing decisions. Instead, it is likely that some of these portrayals of noncitizens, especially for certain nationalities and those with “illegal” status, may creep into the courtroom decision-making process, even if these influences are unintentional.

Empirical Findings on the Relationship between Citizenship and Sentencing

Although citizenship has received less empirical attention than race/ethnicity (or gender or age), there is a growing body of research examining sentencing disparities between U.S. citizens and noncitizens. However, this line of research has produced somewhat mixed findings. Several studies find that noncitizen defendants receive more punitive sentencing outcomes than similarly-situated U.S. citizen defendants, especially in the incarceration decision (Demuth 2002; Hartley and Armendariz 2011; Light et al. 2014). To illustrate, research by Demuth (2002) indicates that noncitizens are more likely to be sentenced to prison, compared to U.S. citizens with the same offense charges. Similarly, Light et al. (2014) report that noncitizen defendants are four times more likely to receive incarceration than their U.S. citizen counterparts, and undocumented or “illegal” noncitizen defendants have seven times greater odds of incarceration.

However, other studies indicate that noncitizen defendants are treated similarly or even more favorably than U.S. citizens, particularly in sentence length decisions. For example, Valadez and Wang (2017) find that noncitizens are disadvantaged at incarceration but are sentenced to similar prison lengths as U.S. citizens. Similarly, Kautt and DeLone (2006) find no difference in sentence lengths between U.S. citizens and noncitizens, net of other factors. In contrast, several other studies actually find noncitizenship advantages in sentence length. Both Wolfe et al. (2011) and Wu and DeLone (2012) find that noncitizens are disadvantaged in the incarceration decision but receive shorter average sentences compared to U.S. citizens, even when they hold “illegal” status. Taken together, these studies provide a mixed picture of citizenship effects in sentencing, which differ depending on the dependent variable examined. That is, noncitizens appear to be disadvantaged in incarceration decisions, but the relationship between sentence length and citizenship is less clear.

There are several potential explanations for the seemingly mixed findings reported in prior research. First, several of these studies have focused on different sets of districts and on specific or single offense types (e.g., drug offenses, non-immigration offenses) (Demuth 2002; Logue 2009; Valadez and Wang 2017), which may have produced different findings. Second, these studies have relied on different time periods (typically early 2000s) to examine noncitizen effects on sentencing. To illustrate, Demuth (2002) uses data from 1996 to 1999 and Wolfe and colleagues (2011) use data from 2005 to 2006, while Light et al.’s (2014) data covers 1992 to 2008. Thus, the mixed findings across studies could be a result of the different time periods examined. Light et al. (2014) echoes this point, noting that noncitizen effects in federal sentencing appear to have widened over time. Third, research also suggests that the mixed findings, especially for sentence length, may reflect practical considerations of sentencing for noncitizens. Specifically, judges may give incarceration to noncitizens at a higher rate due to concerns about flight risk, but they also may give short sentences to noncitizens in anticipation of swift deportation (Valadez and Wang 2017; Wolfe et al. 2011). Last, prior research may have produced differing citizenship effects because some studies separate noncitizens by national origin (or legal migration status) (e.g., Holland 2017; Light et al. 2014; Valadez and Wang 2017), while others do not (e.g., Kautt and DeLone 2006; Wu and DeLone 2012). As a result, citizenship effects on sentencing may have been obscured depending on how noncitizen defendants were defined and categorized.

The mixed evidence in prior studies (especially as it relates to sentence length) suggests that further research is needed to obtain a more complete picture of citizenship and sentencing. Based on the issues highlighted above, there is a pressing need to examine citizenship effects in more recent time periods (post-2010), for multiple offense types (rather than one or two), and covering all U.S. districts (rather than a select set of districts or locations). Perhaps most importantly, there is a need to examine the intersectionality of national origin and legal migration status in citizenship effects. As noted in the previous section, focal concerns and group threat theories both suggest that noncitizen effects in sentencing may be more severe for defendants from certain countries (e.g., Mexico, Central and South America, and Africa) and for those without legal migration status. However, research to date has given limited attention to this potential intersectionality in citizenship effects, despite longstanding evidence of the intersectionality between race/ethnicity, gender, and age at sentencing. There are several noteworthy exceptions that have begun this work, focusing largely on the ways in which legal migration status shapes citizenship effects on sentencing (Light 2014; Wolfe et al. 2011). In contrast, analyses that differentiate citizenship by national origin (especially when combined with legal migration status) are less common.² As such, examining effects of national origin (and its intersection with legal migration status) is important for providing a more complete understanding of the precise ways in which citizenship influences federal sentencing.

In light of the gaps in research identified above, the current study seeks to extend sentencing research by examining the effects of defendant (1) national origin and (2) legal migration status on incarceration and sentence length decisions in federal criminal courts, using data from 2011 to 2016. Based on the theoretical arguments outlined above and prior literature on racial/ethnic and citizenship effects on sentencing outcomes, we predict the following:

Hypothesis 1: Compared to U.S. citizens, noncitizen defendants will be more likely to receive incarceration decisions, net of other factors.

Hypothesis 1a: Noncitizen effects will be stronger for groups that have been viewed more negatively in public and political rhetoric (e.g., Mexican, Latin American, African).

Hypothesis 1b: Noncitizen effects will be stronger for noncitizens with “illegal” status compared to those with “legal” status.

Hypothesis 1c: National origin and legal migration status effects will intersect to create greater incarceration risks for defendants who are from countries with more negative immigration narratives and who have “illegal” migration status (e.g. Mexican “illegal” and Latin American “illegal” defendants).

Hypothesis 2: Compared to U.S. citizens, noncitizen defendants will receive longer prison sentence length decisions, net of other factors.

Hypothesis 2a: Noncitizen effects will be stronger for groups that have been viewed more negatively in public and political rhetoric (e.g., Mexican, Latin American, African).

Hypothesis 2b: Noncitizen effects will be stronger for noncitizens with “illegal” status compared to those with “legal” status.

Hypothesis 2c: National origin and legal migration status effects will intersect to create longer sentences for defendants who are from countries with more negative immigration narratives and who have “illegal” migration status (e.g. Mexican “illegal” and Latin American “illegal” defendants).

Data and Method

Data for this study are drawn from the annual United States Sentencing Commission (USSC) Monitoring of Federal Criminal Sentences (MFCS) datasets from 2011 to 2016. These data include all felony and serious misdemeanor criminal cases in federal criminal courts with sentencing dates between October 1, 2011 and September 30, 2016.³ After combining five years of data, the total number of cases in the dataset is 378,789. The MFCS dataset is particularly well-suited for our analysis because it provides detailed information on defendant citizenship status, national origin, and legal migration status, with enough cases within each grouping for meaningful statistical analyses.⁴

Dependent Variables

Criminal sentencing has been conceptualized as a two-step process: (1) the decision to incarcerate and (2) the decision of how long to incarcerate (i.e., prison sentence length) (Doerner and Demuth 2010; Steffensmeier et al. 1998). To reflect this, the current study uses two dependent variables.

First, incarceration is measured as a dichotomous variable (1 = Prison; ref. = No Prison). Second, sentence length is measured continuously as months of incarceration ordered. Aligning with USSC recommendations (USSC 2004, 2012), sentence length is capped at 470 months—the USSC representation of a life sentence. Due to heavy right skew of the sentence length measure, we use the natural log of sentence length in regression models to impose a more normal distribution (for similar practices, see Feldmeyer and Ulmer 2011; Ulmer and Parker 2020).

Independent Variables

The main independent variable in this study is defendant’s national origin-by-legal status. National origin indicates a defendant’s country of citizenship. The USSC measure of country of citizenship includes more than 150 different nationalities. Due to the small number of cases from many countries, we constructed regional groupings for noncitizen defendants to create substantively meaningful categories with enough cases to allow for statistical analyses (Shadish, Cook, and Campbell 2002). Specifically, we combine nationalities into seven categories based on the United Nations Regional Groups (UNRG) system: African, Asia-Pacific, Eastern European, Latin American/Caribbean, Mexican, Western European, and U.S. Citizens (reference group) (for similar practices, see Holland 2017).⁵ In addition, the USSC data provide the legal status of noncitizens, indicating whether a defendant’s migration status is “legal” or “illegal.” With this information, we combined the national origin and legal migration status measures to create national origin-by-legal status categories. This resulted in 12 dummy variables indicating the defendant’s legal migration status for each of the non-U.S. national origins (i.e., African legal, African illegal, Latin American/Carribean legal, Latin American/Carribean illegal, etc.), in addition to the reference group of U.S. citizens.⁶

Control Variables

Based on the practices of prior literature, we include several relevant extralegal and legal control variables (Demuth 2002; Feldmeyer and Ulmer 2011; Light et al. 2014; Steffensmeier et al. 1998). We use six extralegal controls: race/ethnicity, gender, educational attainment, dependents, age, and age-squared. We use dummy variables to measure defendant race/ethnicity (1 = Black, Hispanic, Other; ref. = White), gender (1 = Female, ref. = Male), educational attainment (1 = High School Graduate, Some College, College Graduate; ref. = Not High School Graduate), and dependents (1 = 1+ Dependents; ref. = No Dependents). Defendant’s age is measured continuously in years. We also include an age-squared variable to account for potential curvilinear effects of age (Doerner and Demuth 2010; Steffensmeier et al. 1995).

We also include a variety of legal variables that have commonly been used in prior federal sentencing studies (Feldmeyer and Ulmer 2011; Ulmer et al. 2010; Ulmer and Parker 2020). Guideline minimum range for imprisonment is measured continuously as determined by the court.⁷ This variable is capped at 470 months (USSC 2004) and due to its heavy right skew, we use the natural log of the variable in regression models. Research indicates that criminal history exerts an influence on sentencing outcomes, above and beyond its contribution to the guideline prescribed minimum (Holmes and Feldmeyer 2019; Ulmer 2000). As a result, we control for criminal history category (ranging one to six) using a series of six dummy variables (with Category 1 as the reference group). Furthermore, we include a series of dummy variables representing departures (1 = Upward, Downward, Government-Assisted; ref. = Within Range),⁸ focal offense type (1 = Property, Drug, Fraud, Firearm, Immigration, Other; ref. = Violent), presentence detention (1 = In Custody, Other; ref. = Bail/Released On Own Recognizance), acceptance of responsibility (1 = Adjustment Applied; ref. = No Adjustment Applied), and mode of disposition (1 = Trial; ref. = Plea).⁹ In the sentence length model only, we control for whether the defendant received supervised release using a dichotomous indicator (1 = Yes; ref. = No).¹⁰

Analytic Approach

The analysis takes place in three steps. First, we provide descriptive statistics for all dependent, independent, and control variables. Second, we use logistic regression to isolate the effect of noncitizen defendant’s national origin-by-legal status on the odds of incarceration, net of other factors. Third, we use ordinary least squares (OLS) regression to isolate the effect of noncitizen defendant’s national origin-by-legal status on the length of sentence ordered, net of other factors. Throughout the analyses, we conduct Wald’s tests for equality of coefficients to compare the magnitude of the legal versus illegal effects within the same nationality groupings (for similar practice in sentencing research, see Holleran and Spohn 2004; Spohn et al. 2014; Steffensmeier et al. 2017). Due to the nesting of cases within federal districts, standard errors are clustered by sentencing district to account for intra-district correlation (Mastrobuoni and Pinotti 2015; Orrick and Piquero 2015).¹¹

Results

Descriptive Statistics

Table 1 provides the descriptive statistics for all dependent, independent, and control variables used in the incarceration analysis. For dichotomous variables, percentages (as opposed to proportions) are reported for ease of interpretation. Similar to prior federal sentencing studies (Doerner and Demuth 2010), the vast majority of federal defendants (90.92 percent) receive a prison sentence. Among those sentenced to incarceration, the average prison sentence is 53.26 months (S.D. = 65.39). With respect to our national origin-by-legal status measures, Table 1 shows that close to 40 percent of the defendants in the sample (38.46 percent) are noncitizens. Among noncitizens, Mexican defendants with illegal status account for the largest share of defendants (27.44 percent), followed by Latin American/Caribbean defendants with illegal status (5.98 percent) and Mexican defendants with legal status (1.89 percent).

Table 1.

Descriptive Statistics of Federal Criminal Cases Sentenced, FY 2012–2016.

Type of Variables	Variables		Mean(%)	SD
Dependent Variables	Incarceration Sentence (ref. = No)		90.92%	0.29
Dependent Variables	Sentence Length		53.26	65.39
Non-U.S. citizen: Nationality-by-legal status	African	Legal	0.20%	0.04
	African	Illegal	0.19%	0.04
	Asian	Legal	0.51%	0.07
	Asian	Illegal	0.24%	0.05
	Eastern European	Legal	0.17%	0.04
	Eastern European	Illegal	0.15%	0.04
	Latin American/Caribbean	Legal	1.28%	0.11
	Latin American/Caribbean	Illegal	5.98%	0.24
	Western European	Legal	0.20%	0.04
	Western European	Illegal	0.20%	0.04
	Mexican	Legal	1.89%	0.14
	Mexican	Illegal	27.44%	0.45
	U.S. (ref.)	U.S. Citizen	61.54%	0.49
Extralegal Controls	Race	White (ref.)	24.97%	0.43
		Black	21.41%	0.41
		Hispanic	49.55%	0.50
		Other Race	4.06%	0.20
	Females (ref. = Male)		13.49%	0.34
	Education	Less than High School Graduate (ref.)	47.08%	0.50
		High School Graduate	30.30%	0.46
		Some College	16.63%	0.37
		College Graduate	5.99%	0.24
	Has Dependents (ref. = No)		64.58%	0.48
	Age		36.64	11.12
	Age Squared		1466.29	922.55
Legal Controls	Guideline Minimum		58.15	72.15
	Criminal history category	1 (ref.)	44.31%	0.50
		2	13.92%	0.35
		3	17.10%	0.38
		4	9.67%	0.30
		5	5.55%	0.23
		6	9.44%	0.29
	Departure	Within Range (ref.)	46.87%	0.50
		Above	2.19%	0.15
		Government Assistance	29.27%	0.46
		Below	21.67%	0.41
	Offense type	Violent (ref.)	3.18%	0.18
		Property	1.73%	0.13
		Drug	32.04%	0.47
		Fraud	15.46%	0.36
		Firearm	11.12%	0.31
		Immigration	28.61%	0.45
		Other	7.86%	0.27
	Pre-sentence status	Custody	73.97%	0.44
		Release on Bail (ref.)	25.44%	0.44
		Other Detention	0.58%	0.08
	Adjustment for Acceptance of Responsibility (ref. = No)		95.21%	0.21
	Trial (ref. = Plea)		3.23%	0.18
	Supervised Release (ref. = No Supervised Release)		77.87%	0.42

Note: Descriptive statistics presented for the incarceration sample (n = 318,817). Percent presented for dichotomous variables—interpreted as the percent of cases coded as “1.” SD = Standard Deviation.

Among extralegal controls, the majority of the sample is Hispanic (49.55 %), followed by White (24.97 percent), Black (21.41 percent), and Other (4.06 percent). In addition, the sample is largely male (86.51 percent), tends to hold less than a high school diploma (47.08 percent), has one or more dependents (64.58 percent), and has an average age of approximately 37 years old (S.D. = 11.12).

Among legal controls, the average case has a guideline minimum score of about 58.15 months (S.D. = 72.15) and falls into Criminal History Category 1 (44.31 percent). Cases are typically sentenced within-guideline range (46.87 percent), and only 2.19 percent of cases receive an upward sentencing departure. Among focal offense types, the majority of cases have a focal drug (32.04 percent) or immigration (28.61 percent) offense. The majority of defendants are in custody pre-sentencing (73.97 percent), receive a sentencing mitigation for acceptance of responsibility (95.21 percent), and have their case disposed of via plea (96.77 percent). For those sentenced to incarceration (and thus in the sentence length models), 77.87 percent receive some form of supervised release.

Multivariate Analysis of Incarceration

Table 2 presents the estimates produced by our multivariate logistic regression analysis, isolating the effects of national origin-by-legal status on the odds of incarceration, net of legal and extralegal factors. Focusing first on our variables of interest, several noteworthy effects emerge. With respect to nationality, Mexican and Latin American/Caribbean defendants have among the highest odds of incarceration compared to U.S citizens. Specifically, Mexican and Latin American/Caribbean defendants with illegal status have 10.82 times and 9.88 times greater odds of incarceration (respectively) compared to U.S. citizens, net of other factors. Similarly, noncitizens from Africa also have significantly higher odds of incarceration (8.83 times greater for those with illegal status; 1.95 times greater for those with legal status) than U.S. citizens. European (Eastern and Western) and Asian defendants with illegal status also have higher odds of incarceration than U.S. citizens. However, these effects are not as severe as those observed for the groups noted above (incarceration odds falling between 3-to-5.5 times greater than U.S. citizens). Taken together, these findings suggest that although almost all noncitizen groupings have higher odds of incarceration than U.S. citizens, the odds of incarceration for Mexican, Latin American/Caribbean, and African noncitizens (who are subject to more negative and crime-related stereotypes) are particularly elevated.

Table 2.

Logistic Regression Examining the Odds of Incarceration.

	b	S.E.	Odds Ratios
Nationality-by-Legal Status (ref. = U.S. Citizen)
African Legal Non-Citizen	0.67**^†	0.24	1.95
African Illegal Non-Citizen	2.18***^†	0.38	8.83
Asian-Pacific Legal Non-Citizen	0.05^†	0.09	—
Asian-Pacific Illegal Non-Citizen	1.10***^†	0.21	3.00
Eastern European Legal Non-Citizen	0.50**^†	0.17	1.65
Eastern European Illegal Non-Citizen	1.70***^†	0.31	5.50
Latin American/Caribbean Legal Non-Citizen	0.51***^†	0.12	1.66
Latin American/Caribbean Illegal Non-Citizen	2.29***^†	0.37	9.88
Western European Legal Non-Citizen	−0.09^†	0.14	—
Western European Illegal Non-Citizen	1.69***^†	0.36	5.41
Mexican Legal Non-Citizen	0.64***^†	0.17	1.89
Mexican Illegal Non-Citizen	2.38***^†	0.67	10.82
Race (ref. = White)
Black	0.16**	0.05	1.18
Hispanic	0.17	0.09	—
Other	0.18*	0.07	1.19
Gender (1 = female)	−0.48***	0.03	0.62
Education (ref. = Less than HS Graduate)
High school graduate	−0.09**	0.04	0.91
Some college degree	−0.06	0.04	—
College graduate	−0.04	0.05	—
Dependents (1 = yes/have)	−0.08**	0.03	0.92
Age	0.02**	0.01	1.02
Age squared	0.00***	0.00	1.00
Natural-logged guideline minimum range	0.96***	0.02	2.61
Criminal History (ref. = Cat 1)
Cat 2	0.17*	0.07	1.19
Cat 3	0.40***	0.09	1.49
Cat 4	0.56***	0.12	1.74
Cat 5	0.62***	0.10	1.85
Cat 6	0.70***	0.12	2.01
Departure (ref. = Within Range)
Above	2.99***	0.26	19.83
Government assistant	−2.60***	0.13	0.07
Below	−2.39***	0.15	0.09
Offense type (ref. = Violent)
Property	−0.68***	0.09	0.51
Drug	0.36***	0.09	1.43
Fraud	−0.02	0.09	—
Firearm	−0.15	0.10	—
Immigration	0.02	0.16	—
Other	−0.34***	0.09	0.71
Pre-sentence detention (ref. = Bail/ROR)
Custody	2.32***	0.09	10.13
Other	−0.06	0.10	—
Acceptance of responsibility (1 = Yes)	−0.86***	0.09	0.42
Trial (1 = Yes)	0.08	0.11	—
Constant	0.74*	0.25	2.10
Observations	318,817
Pseudo R2	0.53

Note: *p < .05, **p < 0.01, ***p < .001. ^†Significant Difference Between Legal and Illegal Estimates. Clustered standard errors based on sentencing district.

In addition, the results in Table 2 also show that across nearly all nationality categories, noncitizen defendants with “illegal” status have significantly greater odds of incarceration than their “legal” counterparts, especially for defendants from Mexico, Latin America, and Africa. For example, Mexican defendants with illegal status have 10.82 times greater odds of receiving a prison sentence than U.S. citizen defendants, while Mexican defendants with legal status have only 1.89 times greater odds of incarceration than similarly-situated U.S. citizens (with a statistically significant difference between “illegal” and “legal” coefficients, p < 0.05). Similarly, noncitizens with illegal status from Africa and from Latin America/Caribbean have 8.83 and 9.88 times greater odds of receiving a prison sentence than U.S. citizens (respectively), while their legal counterparts have only 1.95 and 1.66 times greater odds of incarceration than U.S. citizens (with a statistically significant difference between “illegal” and “legal” coefficients, p < 0.05 for African group, p < 0.001 for Latin America and Caribbean group). Again, this same pattern holds across all nationality groupings, showing that noncitizen defendants with “legal” status receive much greater leniency in incarceration decisions than their counterparts with “illegal” status, net of other controls.

It is worth noting that although incarceration is common for all defendants in federal courts (about 90 percent), our findings indicate that it is even more likely for noncitizens and especially for those from certain regions and with illegal migration status. To illustrate these differences, Online Appendix Table 2 provides descriptive statistics for incarceration (and sentence length) variables for each of the nationality and legal status categories examined here. Online Appendix Table 2 shows that approximately 85.66 percent of U.S. citizen defendants are incarcerated. In contrast, noncitizens with “illegal” status from Africa, Eastern Europe, Latin America, and Mexico are incarcerated in more than 95 percent of cases. Thus, even though most defendants receive incarceration, there are meaningful and substantively important differences in incarceration across these groupings.

The effects of the control variables on incarceration largely mirror those shown in prior research. With respect to extralegal characteristics, Black defendants are more likely to receive prison sentences than White defendants (1.18 times higher odds). However, the odds of incarceration for Hispanic defendants do not differ significantly from those of White defendants. Female defendants have lower odds (OR = 0.62) of incarceration than males, and defendants who graduated from high school are less likely to receive incarceration than those who did not. Defendants with dependents in their care are less likely to receive a prison sentence (OR = 0.92) than those without dependents. In addition, older defendants are, in general, more likely to receive a prison sentence. However, this effect is curvilinear based on the quadratic term, indicating lower odds of incarceration for the youngest and oldest defendants.

Turning to our legal controls, higher guideline minimums and higher criminal history categories are associated with increased odds of incarceration. Defendants who receive an upward departure have greater odds of incarceration compared to those sentenced within the guideline recommended range. Likewise, defendants who receive downward departures (both government-initiated and judge-initiated) are less likely to receive a prison sentence than those sentenced within range. In terms of offense types, defendants with drug offenses have higher odds (OR = 1.43) of incarceration, and those with property offenses have lower odds of incarceration (OR = 0.51), compared to those with violent offenses. Notably, those convicted of immigration offenses have similar odds of incarceration as those convicted of violent offenses. Defendants in custody pre-sentencing have 10.13 times greater odds of incarceration than those released pre-sentencing. Last, defendants who receive a sentence mitigation for acceptance of responsibility are less likely to receive a prison sentence, but defendants who go to trial are equally likely to receive a prison sentence than those who enter a plea.

Multivariate Analysis of Prison Sentence Length

Table 3 presents the estimates produced by our multivariate OLS regression analysis, isolating effects of national origin-by-legal status on sentence length. In contrast to the incarceration models, nearly all the national origin-by-legal status categories examined do not produce significant effects. We find only one exception to this pattern: defendants from Western Europe tend to receive shorter average sentences than U.S. citizens. Specifically, Western European defendants with legal status receive approximately 13 percent shorter prison sentences and those with illegal status receive about 10 percent shorter prison sentences compared to U.S. citizen defendants. Notably, there is not a significant difference in “legal” versus “illegal” effects observed for the Western European group.

Table 3.

OLS Regression Examining the Length of Sentence Ordered.

	b	S.E.	% Change
Nationality-by-Legal Status (ref. = U.S. Citizen)
African Legal Non-Citizen	−0.10^†	0.06	—
African Illegal Non-Citizen	0.11^†	0.06	—
Asian-Pacific Legal Non-Citizen	−0.01	0.05	—
Asian-Pacific Illegal Non-Citizen	0.04	0.06	—
Eastern European Legal Non-Citizen	−0.04	0.06	—
Eastern European Illegal Non-Citizen	0.05	0.05	—
Latin American/Caribbean Legal Non-Citizen	−0.04	0.06	—
Latin American/Caribbean Illegal Non-Citizen	0.01	0.03	—
Western European Legal Non-Citizen	−0.14*	0.06	−13.21
Western European Illegal Non-Citizen	−0.11*	0.05	−10.08
Mexican Legal Non-Citizen	−0.07	0.04	—
Mexican Illegal Non-Citizen	−0.07	0.04	—
Race (ref. = White)
Black	0.03	0.02	—
Hispanic	−0.02	0.03	—
Other	−0.11***	0.03	−10.02
Gender (1 = female)	−0.19***	0.02	−17.68
Education (ref. = Less than HS Graduate)
High school graduate	0.01*	0.00	1.11
Some college degree	0.00	0.01	—
College graduate	−0.01	0.02	—
Dependents (1 = yes/have)	−0.01	0.01	—
Age	0.02***	0.00	1.70
Age squared	−0.00***	0.00	−0.02
Natural-logged guideline minimum range	0.64***	0.02	0.64
Criminal History (ref = Cat 1)
Cat 2	−0.08	0.05	—
Cat 3	−0.06	0.03	—
Cat 4	−0.01	0.02	—
Cat 5	0.04	0.02	—
Cat 6	0.14***	0.02	14.73
Departure (ref. = Within Range)
Above	0.61***	0.04	84.42
Government assistant	−0.69***	0.05	−49.93
Below	−0.57***	0.03	−43.65
Offense type (ref. = Violent)
Property	−0.20***	0.04	−18.50
Drug	0.03	0.02	—
Fraud	−0.15***	0.03	−13.73
Firearm	0.06*	0.02	5.83
Immigration	−0.21***	0.05	−18.68
Other	−0.03	0.02	—
Pre-sentence detention (ref. = Bail/ROR)
Custody	0.56***	0.04	75.88
Other	0.04	0.07	—
Acceptance of responsibility (1 = Yes)	−0.18***	0.02	−16.12
Trial (1 = Yes)	0.20***	0.03	22.35
Supervised release (1 = Yes)	0.10*	0.04	10.69
Constant	0.86***	0.06
Observations	284,673
R2	0.70

Notes: *p < .05, **p < 0.01, ***p < .001. ^†Significant Difference Between Legal and Illegal Estimates. Clustered standard errors based on sentencing district.

Turning our focus to the extralegal control variables, we note several findings. Effects of defendant race/ethnicity on sentence length are non-significant, with the exception of “Other” race defendants receiving shorter sentences than Whites.¹² Consistent with prior literature (Etienne 2010), females receive shorter average sentences than male defendants. In addition, defendants who graduate high school receive slightly longer sentences than those without a high school diploma. Age has curvilinear effects on sentence length, indicating that the youngest and oldest defendants receive shorter sentences on average than middle-aged defendants.

With respect to legal factors, results generally mirror those found in prior research. Defendants with greater guideline minimums and in the highest criminal history categories (i.e., level 6) receive significantly longer sentences, compared to defendants with lower guideline minimums and criminal history scores. Defendants who receive upward departures are sentenced to longer sentences (84.42 percent longer) than those sentenced within the guideline range. Individuals who receive government-initiated or judge-initiated downward departures receive shorter average sentences than those who are sentenced within the guideline range (49.93 percent and 43.65 percent shorter, respectively). With respect to offense type, defendants convicted of property, fraud, and immigration offenses tend to receive shorter sentences (18.50 percent, 13.73 percent, 18.68 percent shorter, respectively), while those convicted of firearms offenses tend to receive longer (5.83 percent) sentences than defendants convicted of violent offenses. Defendants who are in custody pre-sentencing receive longer sentences than those who are released pre-sentencing. Defendants with a sentencing mitigation for acceptance of responsibility receive significantly shorter (16.12 percent) sentences on average. Last, defendants who go to trial or receive supervised release tend to receive longer sentences than those who accept a plea or do not receive supervised release (22.35 percent and 10.69 percent, respectively).

Supplemental Analyses by Offense Type

The analyses presented thus far include immigration offenses alongside other offense types (violent, property, drug, etc.). However, it is possible that the citizenship effects observed above may differ for (or be driven largely by) immigration offenses. Immigration offenses are one of the most common offense types in the federal criminal caseload and account for almost 30 percent of annual federal criminal cases (Hartley and Tillyer 2012; Demleitner and Sands 2002). More importantly for the current analysis, noncitizens are particularly likely to be convicted on immigration charges (more than half of cases involving noncitizens). Thus, the citizenship effects observed in the original analyses may simply reflect findings for immigration cases that do not extend to other offenses. This is also likely if (as the intersectionality arguments would suggest) noncitizen disadvantages are amplified for offense types (i.e., immigration cases) that are more commonly associated with noncitizen groups. As in our original hypotheses, we again would expect noncitizen disadvantages to be particularly strong for those groups that have been at the center of public concerns about immigration-related crime risks (e.g., noncitizens from Mexico, Central America, and with illegal status). To address these concerns and more fully exhaust the data, we replicated all analyses for (1) non-immigration offenses (our original models minus immigration cases) and then separately for (2) immigration offenses only, shown in Table 4 (incarceration) and Table 5 (sentence length).

Starting with incarceration, findings from our supplemental models generally mirror those from our main models, suggesting that immigration offenses are not driving our findings. Specifically, Table 4 shows that nearly all noncitizen defendant groups are more likely to receive incarceration than U.S. citizens for both immigration and non-immigration cases. Similar to our main analysis, for nearly all groups these effects also tend to be stronger for defendants with “illegal” status (with one exception of Eastern Europeans with an immigration offense).¹³ Notably, we find a few small differences in these effects and a few coefficients do not reach statistical significance in the immigration-only model of incarceration (perhaps due to reduced sample sizes for some of the nationality categories after disaggregating by offense type). However, the substantive findings from these models are generally consistent with our main analysis, showing that noncitizens are more likely to receive incarceration than U.S. citizens, net of other factors. In addition, these disadvantages are again particularly strong for noncitizens with “illegal” status from Mexico, Latin American/Caribbean, and Africa.

Table 4.

Logistic Regression Examining Odds of Incarceration by Offense Type.

	Non-Immigration Cases			Immigration Cases
	b	S.E.	Odds Ratios	b	S.E.	Odds Ratios
Nationality-by-Legal Status (ref. = U.S. Citizen)
African
Legal	0.63*^†	0.21	1.88	1.26^†	0.92	—
Illegal	1.85***^†	0.36	6.34	2.00**^†	0.72	7.42
Asian-Pacific
Legal	0.04^†	0.11	—	0.08^†	0.39	—
Illegal	0.99*^†	0.30	2.69	0.89^†	0.48	—
Eastern European
Legal	0.27^†	0.17	—	2.17**^†	0.68	8.78
Illegal	1.56***^†	0.40	4.74	1.37*^†	0.61	3.93
Latin American/Caribbean
Legal	0.49*^†	0.14	1.63	0.42^†	0.32	1.51
Illegal	1.49***^†	0.23	4.44	1.89***^†	0.45	6.64
Western European
Legal	−0.07^†	0.15	—	−0.29^†	0.68	—
Illegal	1.22*^†	0.36	3.39	2.26**^†	0.76	9.61
Mexican
Legal	0.73***^†	0.21	2.07	0.45^†	0.26	—
Illegal	2.02***^†	0.25	7.54	1.80***^†	0.37	6.05
Observations	227,588			91,229
Pseudo R2	0.53			0.43

Note: *p < .05, **p < 0.01, ***p < .001. ^†Significant Difference Between Legal and Illegal Estimates. Clustered standard errors based on sentencing district. Immigration and Non-Immigration models include the full set of control variables from our main models (effects not shown, available upon request).

Turning to the supplemental sentence length models, we do find some noteworthy differences from our main model. As the reader may recall, our main model showed almost no significant differences in sentence length between U.S. citizens and noncitizens across nationality-by-legal status groupings. In contrast, our supplemental models show some significant disadvantages in sentence length for noncitizens, but mainly for those with “illegal” status. Among non-immigration cases, Table 5 shows that African, Eastern European, and Latin American/Caribbean defendants with illegal status receive longer sentences than similarly situated defendants with U.S. citizenship. The same groups are more disadvantaged when focusing exclusively on immigration offenses. Here, we also find that Mexican defendants with illegal status (for immigration offenses) receive significantly longer prison sentences compared to U.S. citizen defendants. Notably, these findings are masked in our main model which combines immigration and non-immigration cases (a point we return to in our discussion of areas for future research). Regardless, the supplemental models offer some evidence that noncitizens may, in fact, be disadvantaged in sentence length decisions. However, this effect appears to mainly apply to those with “illegal” status and is largely limited to defendants from Africa, Eastern Europe, and Latin America/Caribbean countries.¹⁴

Table 5.

OLS Regression Examining Length of Sentence by Offense Type.

	Non-Immigration Cases			Immigration Cases
	b	S.E.	% Change	b	S.E.	% Change
Nationality-by-Legal Status (ref. = U.S. Citizen)
African
Legal	−0.04†	0.06	−3.85	0.23†	0.12	25.86
Illegal	0.12*†	0.05	12.69	0.32*†	0.14	37.95
Asian-Pacific
Legal	0.01	0.04	—	−0.10	0.21	—
Illegal	0.08	0.07	—	0.16	0.11	—
Eastern European
Legal	0.01	0.05	—	−0.31	0.37	—
Illegal	0.12*	0.05	12.27	0.24**	0.09	27.68
Latin American/Caribbean
Legal	−0.01	0.06	—	0.08	0.06	—
Illegal	0.05*	0.02	4.98	0.20**	0.06	22.65
Western European
Legal	−0.10	0.06	—	−0.12	0.25	—
Illegal	−0.08	0.04	—	0.15	0.14	—
Mexican
Legal	0.01	0.03	—	0.09	0.05	—
Illegal	−0.03	0.04	—	0.12*	0.06	13.25
Observations	196,950			87,723
R2		0.60			0.79

Discussion

The purpose of this study was to extend research on citizenship effects in sentencing to examine how (1) national origin and (2) “legal” versus “illegal” migration status of noncitizen defendants are intertwined with incarceration and prison sentence length decisions in federal criminal cases. In doing so, the current study makes several contributions to sentencing research. For one, we extend prior literature on the relationship between citizenship and sentencing by using more recent data, including all offense types (along with supplemental models disaggregating results by immigration offenses), and providing national coverage of U.S. federal districts (rather than focusing on a select set of districts or locales). More importantly, the current study draws on a longstanding tradition of assessing intersectionality in sentencing effects to show how national origin and legal migration status may combine to create particularly strong disadvantages in sentencing for some noncitizens, especially in the incarceration decision.

With this in mind, the results of this study can be briefly summarized. First, noncitizen defendants, are more likely to be sentenced to incarceration than U.S. citizens. However, once they receive a prison sentence, the length of that sentence does not significantly differ between U.S. citizens and noncitizens (except for some evidence in the supplemental analysis of immigration offenses).

Second, citizenship effects differ across nationality groupings. Although all noncitizen groups are disadvantaged in incarceration decisions, certain nationalities are particularly disadvantaged. As our hypotheses suggested, Mexican, Latin American/Caribbean, and African noncitizen defendants have the greatest odds of receiving a prison sentence, followed by European and Asian noncitizen defendants. This aligns with focal concerns (and group threat) arguments that noncitizens from countries which have been the target of more negative crime-related stereotypes and political rhetoric may be recipients of harsher punishment.

Third, the noncitizen effect is magnified when the defendant holds “illegal” status, which again aligns with our predictions and with prior research on intersectionality in sentencing. Across all national origin categories, defendants with “illegal” migration status receive harsher punishments than U.S. citizens and their “legal” noncitizen counterparts, especially for incarceration. Although this effect is less consistent in sentence length, when effects are observed (i.e., in the supplemental models) they show that defendants with “illegal” status tend to receive the longest sentences of any groups examined. Moreover, the findings show that certain nationalities who tend to be subject to crime-related stereotypes (e.g., Mexican, Latin American/Caribbean, African) are particularly disadvantaged when defendants hold “illegal” status—again highlighting the intersectional nature of these effects.

Fourth, effects of citizenship by nationality and legal status are more consistent for incarceration than for sentence length. Notably, the null effects of citizenship in our sentence length models are consistent with several other studies that find limited influence of citizenship on sentence length (Kautt and DeLone 2006; Valadez and Wang 2017). As suggested earlier, this may reflect deportation considerations. If a noncitizen defendant is scheduled to be deported, judges may be more inclined to give them incarceration (perhaps due to concerns of flight risk) but not feel the need to impose a longer prison sentence (Wu and DeLeon 2012; see Orrick, Compofelice, and Piquero 2016). However, the supplemental analysis does reveal some noncitizen disadvantages in sentence length, but only for defendants with “illegal” status and only when immigration versus non-immigration offenses are examined separately.

Taken together, the results of this study raise several important implications for sentencing research and theory. First, as mentioned above, the size of the noncitizen population in the both the federal criminal caseload and U.S. population more broadly is growing. As a result, disparities in sentencing based on citizenship present an increasingly pressing problem for the federal court system. The federal sentencing guidelines were designed and implemented for the expressed purpose of creating fairness, uniformity, and consistency in sentencing and with the goal of eliminating unwarranted disparities based on defendant characteristics (USSC 2004, 2012). The findings here raise serious concerns about the degree to which this uniformity and fairness has been achieved and suggests that defendant citizenship, nationality, and legal migration status may result in significant disparities in punishment, even after controlling for legally-relevant factors. Thus, in addition to addressing race/ethnicity and gender disparities in punishment, it is equally important for the federal system to target unwarranted disparities in punishment based on citizenship, national origin, and legal migration status.

Second, in terms of theory, focal concerns and group threat perspectives have largely been used to explain racial/ethnic effects in sentencing. However, the current results suggest that these frameworks may also be relevant for citizenship effects in sentencing. The effects of citizenship observed here are consistent with these theoretical positions. In addition, the greater disadvantages for (1) defendants with “illegal” status and (2) defendants from nations that have been the focus of political and public scrutiny (Mexico, Latin America/Caribbean, African) further align with these theoretical arguments. Like many sentencing studies, we are unable to directly measure courtroom actors’ focal concerns or perceptions of “threat” based on defendant’s citizenship status, national origin, or legal migration status. In addition, we do not provide direct tests of racial/ethnic threat, examining how changing citizenship contexts shape effects of defendant citizenship (i.e., using cross-level interactions between percent noncitizen in a federal district and defendant citizenship). Future research is needed that explores such relationships in detail to more directly test these theoretical arguments (for a more detailed discussion on this topic, see Lynch 2019). With that being said, our findings provide an initial indication that focal concerns theory and group threat perspective may offer valuable explanations for the types of citizenship effects observed here.

Although this study makes several important contributions to research on federal sentencing and citizenship, it is not without limitations. First, in order to provide sufficient cases for our national origin categories, this study combined countries into six groups (seven including U.S. citizens). However, there is likely meaningful within-group variation in sentencing effects within these broader categories. For example, in this study, there are 47 countries in our Asia-Pacific group and 45 countries in the African group. Sentencing outcomes may differ across individual national origins with the broader regional groupings used here. Thus, we encourage future research to explore these potential differences in sentencing effects for individual countries of origin. Second, there is a need for additional analyses exploring citizenship and nationality effects across separate offense categories. Our supplemental analysis suggests that there may be important differences in punishment across nationality-by-legal citizenship categories when immigration offenses are examined separately from other offense categories. However, more work is needed to fully explore how citizenship, nationality, and legal migration status shape sentencing for immigration offenses and other specific offense types. Related to this, research is also needed that more thoroughly examines how deportation proceedings and considerations shape these decisions, especially for immigration versus non-immigration cases. Third, there is a need to explore the way that race/ethnicity may intersect with the nationality and “legal” status effects observed here. This is particularly noteworthy given that the most disadvantaged noncitizen groups in our analysis (Mexican, Latin American/Caribbean, African) overlap with racial/ethnic groups (Black and Latino) that have traditionally been disadvantaged in sentencing decisions. Our analyses control for a defendant’s race/ethnicity to isolate the effect of nationality, citizenship, and legal/illegal status. However, given the substantial overlap in these categories, future research is needed that explores the intersectionality of race/ethnicity and citizenship by nationality more closely.

Last, research is needed that extends these analyses to examine whether they differ across federal sentencing districts and depend on court and district characteristics. Research on racial threat has shown some indications that minority group disadvantages are greater in places with large or growing minority populations (see Feldmeyer and Ulmer 2011; Feldmeyer, Harris, and Scroggins 2015). Similarly, Light et al. (2014) find that citizenship effects are more pronounced in areas with growing immigrant populations. However, additional research is needed to provide a more direct test of the group threat perspectives discussed here and to assess how such contextual effects may differ depending on a defendant’s nationality and legal migration status.

Clearly, much more research is needed to develop a complete understanding of the ways in which citizenship, national origin, and legal migration status shape sentencing decisions. With that being said, the current study offers an important step toward this goal and indicates that noncitizens are highly disadvantaged in incarceration decisions (but less so in sentence length decisions) in federal criminal courts compared to the similarly-situated U.S. citizens. Moreover, this disadvantage is particularly pronounced for defendants from Mexico, Latin America, and Africa and for those with “illegal” migration status. Although the federal court system has long pursued equality and fairness in sentencing and sought to reduce unwarranted disparities in punishment across race/ethnicity and other demographic characteristics, the current findings suggest that disparities based on citizenship and nationality persist and, for some groups, are quite severe. Thus, as noncitizen populations continue grow within the U.S. population and the federal criminal caseload, understanding and addressing these citizenship disparities in punishment will be an increasingly important goal for researchers and for the federal criminal justice system.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Doyun Koo

Bryan Holmes

Notes

References

Albonetti

C. A.

1991. “An Integration of Theories to Explain Judicial Discretion.” Social Problems 38:247–66.

Blalock

H. M.

Jr . 1967. Toward a Theory of Minority Group Relations. New York: Capricorn Books.

Blumer

1958. “Race Prejudice as a Sense of Group Position.” Pacific Sociological Review 23:3–7.

Carey

Marak

. 2018. “A Historical Perspective of Immigration and Drug Control Policy.” Pp. 19–32 in Routledge Handbook on Immigration and Crime, edited by Peguero

Miller

H. V.

. Abingdon: Taylor & Francis.

Coutin

S. B.

2005. “Contesting Criminality: Illegal Immigration and the Spatialization of Legality.” Theoretical Criminology 9:5–33.

Demleitner

N. V.

Sands

J. M.

. 2002. “Non-Citizen Offenders and Immigration Crimes: New Challenges in the Federal System.” Federal Sentencing Reporter 14:247–54.

Demuth

2002. “The Effect of Citizenship Status on Sentencing Outcomes in Drug Cases.” Federal Sentencing Reporter 14:271–5.

Demuth

Steffensmeier

. 2004. “Ethnicity Effects on Sentence Outcomes in Large Urban Courts: Comparisons among White, Black, and Hispanic Defendants.” Social Science Quarterly 85:994–1011.

Doerner

J. K.

Demuth

. 2010. “The Independent and Joint Effects of Race/Ethnicity, Gender, and Age on Sentencing Outcomes in US Federal Courts.” Justice Quarterly 27:1–27.

10.

Etienne

2010. “Sentencing Women: Reassessing the claims of Disparity.” Journal of Gender, Race, and Justice 14:73–84.

11.

Everett

R. S.

Wojtkiewicz

R. A.

. 2002. “Difference, Disparity, and Race/Ethnic Bias in Federal Sentencing.” Journal of Quantitative Criminology 18:189–211.

12.

Feldmeyer

Cochran

J. C.

. 2018. “Racial Threat and Social Control: A Review and Conceptual Framework for Advancing Racial Threat Theory.” Pp. 283–316 in Building a Black Criminology, Vol 24, edited by Unnever

J. D.

Gabbidon

S. L.

Chouhy

. New York: Routledge.

13.

Feldmeyer

Ulmer

J. T.

. 2011. “Racial/Ethnic Threat and Federal Sentencing.” Journal of Research in Crime and Delinquency 48:238–70.

14.

Feldmeyer

Harris

C. T.

Scroggins

. 2015. “Enclaves of Opportunity or “Ghettos of Last Resort?” Assessing the Effects of Immigrant Segregation on Violent Crime Rates.” Social Science Research 52:1–17.

15.

Feldmeyer

Warren

P. Y

Siennick

S. E.

Neptune

. 2015. “Racial, Ethnic, and Immigrant Threat: Is There a New Criminal Threat on State Sentencing?” Journal of Research in Crime and Delinquency 52:62–92.

16.

Fischman

J. B.

Schanzenbach

M. M.

. 2011. “Do Standards of Review Matter? The Case of Federal Criminal Sentencing.” The Journal of Empirical Legal Studies 40:405–37.

17.

Fox

2015. Applied Regression Analysis and Generalized Linear Models. 3rd ed. Thousand Oaks, CA: Sage.

18.

Hartley

R. D.

Armendariz

L. F.

. 2011. “Border Justice? Sentencing Federal Narcotics Offenders in Southwest Border Districts: A Focus on Citizenship Status.” Journal of Contemporary Criminal Justice 27:43–62.

19.

Hartley

R. D.

Tillyer

. 2012. “Defending the Homeland: Judicial Sentencing Practices for Federal Immigration Offenses.” Justice Quarterly 29(1):76–104.

20.

Holland

M. M.

2017. “Equal Justice: Examining the Effects of Citizenship, Documentation Status, and Country of Origin on Carceral Punishment across Federal Districts.” Criminal Justice Review 43:419–39.

21.

Holleran

Spohn

C. C.

. 2004. “On the Use of the Total Incarceration Variable in Sentencing Research.” Criminology 42:211–40.

22.

Holmes

Feldmeyer

. 2019. “Reassessing the Influence of Criminal History in Federal Criminal Courts.” Justice Quarterly 36:1206–28.

23.

Holmes

Feldmeyer

Kulig

T. C.

. 2020. “Sentencing Departures and Focal Concerns: The Joint Effects of Race and Gender on Departure sin United States District Courts.” Journal of Crime and Justice 1–25. doi: 10.1080/0735648X.2020.1730933.

24.

Iles

Adegun

O. V.

. 2018. “Geography and Sentencing: Does Country of Citizenship Influence Sentence Longevity?” International Journal of Criminology and Sociology 7:48–58.

25.

Johnson

B. D.

Ulmer

J. T.

Kramer

J. H.

. 2008. “The Social Context of Guideline Circumvention: The Case of Federal District Courts.” Criminology 46:711–83.

26.

Kautt

Spohn

. 2002. “Crack-ing Down on Black Drug Offenders? Testing for Interactions among Offenders’ Race, Drug Type, and Sentencing Strategy in Federal Drug Sentences.” Justice Quarterly 19:1–35.

27.

Kautt

P. M.

DeLone

M. A.

. 2006. “Sentencing Outcomes Under Competing but Coexisting Sentencing Interventions: Untying the Gordian Knot.” Criminal Justice Review 31(2):105–31.

28.

Kim

Cano

M. V.

Kim

Spohn

. 2016. “The Impact of United States v. Booker and Gall/Kimbrough v. United States on Sentence Severity: Assessing Social Context and Judicial Discretion.” Crime & Delinquency 62(8):1072–1094.

29.

Kitchens

2010. Federal Sentencing Data and Analysis Issues. Washington, DC: Office of Research and Data United States Sentencing Commission.

30.

Kitchens

2019. Commission Collection of Individual Offender Data. Washington, DC: Office of Research and Data United States Sentencing Commission.

31.

Kramer

J. H.

Ulmer

J. T.

. 1996. “Sentencing Disparity and Departures from Guidelines.” Justice Quarterly 13:81–106.

32.

Lamont

Park

B. Y.

Ayala-Hurtado

. . 2017. “Trump’s Electoral Speeches and his Appeal to the American White Working Class.” The British Journal of Sociology 68:153–80.

33.

Light

M. T.

2014. “The New Face of Legal Inequality: Noncitizens and the long-term Trends in Sentencing Disparities Across US District Courts, 1992–2009.” Law & Society Review 48:447–78.

34.

Light

M. T.

Massoglia

Dinsmore

. 2019. “How do Criminal Courts Respond in Times of Crisis? Evidence from 9/11.” American Journal of Sociology 125(2):485–533.

35.

Light

M. T.

Massoglia

King

R. D.

. 2014. “Citizenship and Punishment: The Salience of National Membership in US Criminal Courts.” American Sociological Review 79:825–47.

36.

Logue

M. A.

2009. “The Price of being Mexican”: Sentencing Disparities between Noncitizen Mexican and non-Mexican Latinos in the Federal Courts.” Hispanic Journal of Behavioral Sciences 31:423–45.

37.

Lopez

2019. “Who Are Today’s U.S. Immigrants?” Pew Research Center 19. (https://us1.campaign-archive.com/?u=434f5d1199912232d416897e4&id=a9e724f458&e=68e3355c63)

38.

Lynch

2019. “Focally Concerned about Focal Concerns: A Conceptual and Methodological Critique of sentencing Disparities Research.” Justice Quarterly 36(7):1148–75.

39.

Mastrobuoni

Pinotti

. 2015. “Legal Status and the Criminal Activity of Immigrants.” American Economic Journal: Applied Economics 7:175–206.

40.

Migration Policy Institute. 2019. Migration Data Hub: U.S. Immigration Trends. Washington, DC: Migration Policy Institute.

41.

Mitchell

2005. “A Meta-Analysis of Race and Sentencing Literature: Explaining the Inconsistencies.” Journal of Quantitative Criminology 21:439–66.

42.

Orrick

E. A.

Piquero

A. R.

. 2015. “Assessing the Impact of Mexican Nativity on Sentence Length.” Criminal Justice Policy Review 26:643–64.

43.

Orrick

E. A.

Compofelice

Piquero

A. R.

. 2016. “Assessing the Impact of Deportable Status on Sentencing Outcomes in a Sample of State Prisoners.” Journal of Crime and Justice 39:28–40.

44.

Shadish

W. R.

Cook

T. D.

Campbell

D. T.

. 2002. Experimental and Quasi-Experimental Designs for Generalized Causal Inference. Boston, MA: Cengage.

45.

Spohn

2000. “Thirty Years of Sentencing Reform: The Quest for a Racially Neutral Sentencing Process.” Criminal Justice 3:427–501.

46.

Spohn

C. C.

Kim

Belenko

Brennan

P. K.

. 2014. “The Direct and Indirect Effects of Offender Drug use on Federal Sentencing Outcomes.” Journal of Quantitative Criminology 30:549–76.

47.

Spohn

Brennan

P. K.

. 2011. “The Joint Effects of Offender Race/Ethnicity and Gender on Substantial Assistance Departures in Federal Courts.” Race and Justice 1:49–78.

48.

Steffensmeier

Kramer

Ulmer

. 1995. “Age Differences in Sentencing.” Justice Quarterly 12:583–602.

49.

Steffensmeier

Ulmer

Kramer

. 1998. “The Interaction of Race, Gender, and Age in Criminal Sentencing: The Punishment Cost of being Young, Black, and Male.” Criminology 36:763–98.

50.

Steffensmeier

Painter-Davis

Ulmer

J. T.

. 2017. “Intersectionality of Race, Ethnicity, Gender, and Age on Criminal Punishment.” Sociological Perspectives 60:810–33.

51.

Trager

Kubrin

C. E.

. 2014. “Complicating the Immigration-Crime Nexus: Theorizing the Role of Gender in the Relationship between Immigration and Crime.” Pp.527–48 in The Oxford Handbook of Gender, Sex, and Crime, edited by Gartner

McCarthy

. London, UK: Oxford University Press.

52.

Trump

2016. “Transcript of the Third Debate.” The New York Times, October 20. Retrieved from (https://www.nytimes.com/2016/10/20/us/politics/third-debate-transcript.html)

53.

Ulmer

J. T.

2000. “The Rules have Changed-so Proceed with Caution: A Comment on Engen and Gainey’s Method for Modeling Sentencing Outcomes under Guidelines.” Criminology 38:1231–43.

54.

Ulmer

J. T.

2012. “Recent Developments and New Directions in Sentencing Research.” Justice Quarterly 29:1–40.

55.

Ulmer

2018. “Race, Ethnicity, and Sentencing.” In Oxford Research Encyclopedia of Criminology and Criminal Justice. Accessed June 10, 2021. (https://oxfordre.com/criminology/view/10.1093/acrefore/9780190264079.001.0001/acrefore-9780190264079-e-262)

56.

Ulmer

J. T.

Parker

B. R.

. 2020. “Federal Sentencing of Hispanic Defendants in Changing Immigrant Destinations.” Justice Quarterly 37:541–70.

57.

Ulmer

J. T.

Kramer

J. H.

. 1996. “Court Communities Under Sentencing Guidelines: Dilemmas of Formal Rationality and Sentencing Disparity.” Criminology 34:383–408.

58.

Ulmer

J. T.

Einstein

Johnson

B. D.

. 2010. “Trial Penalties in Federal Sentencing: Extra-Guideline Factors and District Variation.” Justice Quarterly 27:560–92.

59.

United States Sentencing Commission. 2004. Guidelines Manual. Washington, DC: Office of General Counsel United States Sentencing Commission.

60.

United States Sentencing Commission. 2012. Report on the Continuing Impact of United States v. Booker on Federal Sentencing. Washington, DC: Office of General Counsel United States Sentencing Commission.

61.

United States Sentencing Commission. 2016. Monitoring of Federal Criminal Sentences, [United States], 2014-2015. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor]. Accessed November 9, 2016. (https://doi.org/10.3886/ICPSR36962.v1)

62.

Valadez

Wang

. 2017. “Citizenship, Legal Status, and Federal Sentencing Outcomes: Examining the Moderating Effects of Age, Gender, and Race/Ethnicity.” The Sociological Quarterly 58(4):670–700.

63.

Watkins

Phillip

. 2018. “Trump Decries Immigrants from ‘Shithole Countries’ coming to US.” CNN Politics, January 12, 12.

64.

Wolfe

S. E.

Pyrooz

D. C.

Spohn

C. C.

. 2011. “Unraveling the Effect of Offender Citizenship Status on Federal Sentencing Outcomes.” Social Science Research 40 349–62.

65.

DeLone

M. A.

. 2012. “Revisiting the Normal Crime and Liberation Hypotheses: Citizenship Status and Unwarranted Disparity.” Criminal Justice Review 37:214–38.