Abstract
Over the past two decades, the U.S. federal government has sought to increase its capacity to find, apprehend, and deport noncitizens residing in the United States who have violated federal immigration laws. One way the federal government has done this is by partnering with state and local law enforcement agencies on immigration enforcement efforts. The present study analyzes the records of all 1,964,756 interior removals between fiscal years 2003 and 2015 to examine how, if at all, the types of criminal convictions leading to removal from the U.S. interior have changed during this period of heightened coordination between law enforcement agencies and whether there are differences by gender and region of origin in the types of convictions leading to removal. Findings show that as coordination between law enforcement agencies intensified, the proportion of individuals removed from the U.S. interior with either no criminal convictions or with a driving-related conviction as their most serious conviction increased. Findings also show that the proportion of individuals removed with no criminal convictions was greater for women than for men and that the share of individuals removed with a driving-related conviction as their most serious conviction was greater for Latin Americans than for individuals from all other regions. Given renewed investment in these types of law enforcement partnerships under the Trump administration, the patterns presented in this article may foreshadow trends to come.
Introduction
Over the past two decades, the U.S. federal government has sought to extend its enforcement reach in the interior of the country and, in the process, expand its capacity to find, apprehend, detain, and deport noncitizens residing in the United States who have in some way violated federal immigration laws. One way the federal government has done this is by engaging state and local criminal justice agencies in the everyday business of immigration enforcement. Through a number of initiatives, Immigration and Customs Enforcement (ICE), the agency tasked with enforcing federal immigration laws within the United States’ borders, has increased its coordination with state and local law enforcement departments throughout the country.
While the extent of this coordination is largely hidden from public view, a growing body of scholarship has examined how policing in several locales has changed in the wake of these partnerships (e.g., Armenta, 2017; Coleman, 2012; Coleman & Kocher, 2011; Gardner & Kohli, 2009). Two common findings emerge from these studies. The first is that despite the federal government’s claim that these partnerships are critical to public safety because they make it easier for ICE to apprehend and deport individuals with serious criminal convictions (U.S. Department of Justice, 2017; ICE, 2018), a large share of the individuals arrested and ultimately deported through these initiatives have no history of criminal convictions or have only been convicted of low-level offenses, such as a misdemeanor driving offense. For example, in his study of one of these initiatives in North Carolina, Coleman (2012) finds that in an 18-month period of time, close to 50% of ICE arrests resulting from the initiative began with a traffic stop. The second is that these partnerships open the door to more racialized policing and, in turn, can lead to increases in Latino arrests. For example, in Gardner and Kohli’s (2009) study of the Criminal Alien Program (CAP) in Irving, Texas, the authors find that after the program was established and local police officers gained round-the-clock access to an ICE officer, police began arresting Hispanics for misdemenor offenses in significantly higher numbers than Whites and African Americans.
While these studies provide a critical on-the-ground examination of how immigration enforcement operates in this era of heightened coordination across law enforcement agencies, data limitations have precluded a more macro-level analysis of changing trends in interior enforcement practices. As such, it is unclear if the findings from studies of enforcement practices in specific locales bear out at the national level. Even more generally, the extent to which interior removals increased as coordination across law enforcement agencies intensified and how, if at all, the types of criminal convictions leading to removal may have changed during this period remain unclear. This study begins to fill these gaps in extant knowledge by drawing on administrative records of all ICE removals between fiscal years 2003 and 2015, obtained through Freedom of Information Act (FOIA) requests, to (1) estimate annual counts of interior removals, (2) examine how the types of criminal convictions leading to removal from the U.S. interior have changed over time, and (3) assess differences in conviction type by gender and region.
Results indicate a substantial increase in interior removals between fiscal years 2003 and 2009, followed by a fairly steady decline. Results also indicate a shift during this 13-year period toward the deportation of individuals with driving-related convictions, and especially driving under the influence (DUI) convictions. While the number of interior removals of individuals with a non–driving-related conviction as their most serious criminal conviction increased a mere 8% in this period, the number of removals of individuals with a driving-related conviction as their most serious criminal conviction increased by more than 280%. And as suggested by other scholarship, this increase has been greatest for individuals from Latin America. Among individuals from Latin America with a criminal conviction at the time of their first removal who had initially entered the United States without authorization,1,2 and thus could not lawfully drive in most U.S. states, 22%, or more than 1 in 5, had a driving offense as their most serious criminal conviction. In comparison, among similarly situated individuals from all other regions, roughly 1 in 33 had a driving offense as their most serious conviction. Although reliable information on the method of apprehension is limited to fiscal years 2013 through 2015, analysis of these years indicates that more than 71% of removals of individuals with driving-related convictions started with an apprehension through either 287g or local CAP—two prominent initiatives that enhance coordination between local law enforcement and ICE.
Since taking office, President Trump has sought to dramatically increase immigration enforcement within the U.S. interior. His administration has redefined enforcement priorities so broadly as to ostensibly place all undocumented individuals in the country at risk of deportation regardless of personal circumstance (American Immigration Council, 2017a) and has renewed efforts to involve state and local law enforcement agencies in immigration enforcement. Given these endeavors, the patterns presented in this article, especially during the mid-2000s when interior enforcement reached its height, may foreshadow trends to come.
Background
As noted by a former Commissioner of the U.S. Immigration and Naturalization Service, the United States has constructed a “formidable enforcement machinery” (Meissner, Kerwin, Chishti, & Bergeron, 2013). This machinery extends beyond the nation’s border while also reaching deep into the nation’s interior—into local jails, courtrooms, workplaces, and homes across the country. Interior immigration enforcement is by no means a new practice, nor is the involvement of local law enforcement in policing immigration (Harwood, 1986). Yet the degree of cooperation between law enforcement agencies has grown considerably over the past three decades and this cooperation has become codified into law. Below, I provide a short overview of these changes and discuss recent scholarship on interior immigration enforcement.
The Expansion and Codification of Interagency Cooperation in Interior Immigration Enforcement
The seeds for formalizing interagency cooperation between local and state law enforcement agencies and immigration agents were laid in the 1980s. In a series of bills passed in the latter half of the decade, Congress began expanding the criminal grounds for deportation and improving coordination between law enforcement agencies to ensure that a noncitizen convicted of certain criminal offenses would be processed for deportation following completion of her or his criminal sentence. For example, in the 1986 Anti-Drug Abuse Act, Congress expanded the types of drug offenses that can lead to deportation and authorized the issuance of immigration detainers for individuals arrested for a controlled substance offense and suspected of being in the country without authorization (Hernandez, 2014). 3 These detainers enable the transfer of noncitizens from state prisons and local jails to immigration detention and have become an increasingly important enforcement tool in the federal government’s arsenal (Lasch, 2008). Other initiatives, such as the Alien Criminal Apprehension Program and the Institutional Hearing Program, sought to further coordination between the federal government and local and state law enforcement in order to expedite the removal of noncitizens with criminal convictions prior to the individual completing his or her sentence (Lasch, 2008). 4 Together, these policies have expanded the pool of deportable individuals while also strengthening coordination between law enforcement agencies to ensure that potentially deportable individuals are transferred into immigration custody from criminal custody rather than released back into the community.
In 1996, with the passage of the Illegal Immigration Reform and Immigrant Responsibility Act and the Antiterrorism and Effective Death Penalty Act, Congress went even further by directly extending the authority to enforce federal immigration policies to state and local criminal justice actors. The full scale of these reforms, however, was not realized until after September 11, 2001. Within a few months of the attacks, the Justice Department encouraged jurisdictions to establish hitherto unused 287(g) programs, leading in 2002 to the signing of the first of 80 agreements over the next decade (Watson, 2013). This program, which was authorized under section 287(g) of the Illegal Immigration Reform and Immigrant Responsibility Act, allows for the deputation of local and state law enforcement officers as immigration officers. ICE simultaneously invested in other enforcement approaches that engage local and state law enforcement, such as expanding CAP and establishing Secure Communities. Key to these programs is that they put police officers in regular contact with ICE and automatically link law enforcement databases, enabling ICE to more easily identify noncitizens in criminal custody.
After scaling up the Secure Communities program to cover all jurisdictions in the United States, the Obama administration began to roll back its interior enforcement efforts. Most notably, the director of ICE, John Morton, issued two memorandums in 2011 that redefined enforcement priorities within the U.S. interior to more actively target for removal noncitizens with serious criminal convictions and noncitizens who recently arrived in the United States. These memos altered the guidelines for prosecutorial discretion, leading ICE officers to stop referring cases for removal proceedings that fell outside of the guidelines and ICE prosecutors to administratively close nonpriority cases on their docket. In the ensuing years, there was a sizeable decrease in annual counts of interior removals (U.S. Department of Homeland Security [DHS], 2015).
Studying Interior Immigration Enforcement
While the shifts in federal immigration policy discussed above have been well documented, as have the overall trends in annual counts of removals, limitations in publicly available data impede scholars’ ability to track macro-level trends in how enforcement within the U.S. interior, specifically, has changed over time. In-depth studies of what the geographer Matthew Coleman (2012) calls the “local migration state” provide critical insights into how immigration enforcement works on the ground in different jurisdictions. Although it is clear from this body of work that the implementation of programs like 287(g) can vary substantially across locales, several of these studies point to similar trends with respect to the types of arrests resulting from these programs. In particular, a number of studies find that despite the stated aim of these programs being that they are critical to public safety because they lead to the apprehension of individuals with serious criminal convictions, a large share of the individuals apprehended through these programs have only been convicted of misdemeanor offenses or have never been convicted of a crime (e.g., Capps, Rosenblum, Chishti, & Rodriguez, 2011; Coleman, 2012; Gardner & Kohli, 2009; Nguyen & Gill, 2010).
Additionally, a number of these studies suggest that there is uneven policing of Latinos and non-Latinos, especially when it comes to traffic stops. For example, analysis of policing data in Wake County, North Carolina, where the county jail had a 287(g) and Secure Communities contract, found that Latinos are disproportionately pulled over for driving offenses (Coleman & Kocher, 2011). And in Irving, Texas, researchers found that after the implementation of CAP in the local jail, discretionary arrests of Latinos for petty offenses—especially minor driving offenses—rose dramatically (Gardner & Kohli, 2009). Last, Armenta (2017) provides an in-depth study of policing under the 287(g) program in Nashville, a city where only 8% of the 200,000-plus traffic stops made in 2007 (the first years of the surrounding county’s 287(g) program) resulted in arrest, while stops made of Latino drivers—roughly 5% of the total number of stops—led to arrests 29% of the time (Howard, 2008).
Taking a national and longitudinal view, this article adds to this body of work by (1) analyzing change in the annual number of interior removals between fiscal years 2003 and 2015, (2) examining how the types of crimes leading to removal changed during this period, and (3) investigating differences in conviction type by gender and region.
Data and Method
ICE Removal Records
Information on ICE removals largely come from two data sources—the Yearbook of Immigration Statistics and specific reports on ICE activity, both released annually by the DHS. While these sources provide important information on the annual numbers of individuals deported and voluntarily returned and the countries to which these individuals are removed, it is not possible with these data to ascertain the full picture of changes over time in interior removals or identify those most affected by these changes. For one, there are inconsistencies across years in how detailed these reports are, making it challenging to track certain trends over time. Prior to 2008, for example, the federal government does not separately report how many ICE removals began with an apprehension in the U.S. interior. Additionally, because there is not detailed information about the types of offenses of which removed individuals have been convicted, it is not possible with these data to track changes over time in removals for specific offenses or to examine differences in conviction type by nationality and gender.
The administrative records of ICE removals, obtained through an FOIA request initiated by the Transactional Records Access Clearinghouse, overcome the above data limitations. 5 These data capture the 3,925,758 removals conducted by ICE between fiscal years 2003 and 2015 and, for each removal, include information on the gender, nationality, and age of the removed individual; the date the individual most recently entered the United States; 6 the individual’s legal status at time of entry; 7 the agency involved in the apprehension; 8 and whether the individual has a criminal conviction and, if so, the individual’s most serious criminal conviction. The data also include information on the immigration-related charge leading to removal, an indicator for whether the individual was previously deported, an indicator for whether a detainer was issued, 9 the type of removal (i.e., expedited removal, regular removal, voluntary departure, voluntary return, or reinstatement of prior removal order), 10 the state and port from which the individual was removed, and the date of removal. Because these data are organized at the removal rather than the individual level, it is not possible to track whether the same individual is removed more than once during this period.
Federal Bureau of Investigation Uniform Crime Reporting Program Data
To assess whether the trends I identify in removals by conviction type reflect broader crime trends in the United States, I also draw on data from the Federal Bureau of Investigation (FBI) Uniform Crime Reporting (UCR) Program. Similar to the structure of the ICE removal data, the UCR data capture the number of arrests, as reported by law enforcement agencies, not the number of individuals arrested. Speficially, the UCR data provide information on the total number of arrests each year and the number of arrests for 28 separate offenses. Of the six most common offenses in the ICE removal data, five have comparable offenses listed in the UCR data. I use the UCR data to calculate the share of all arrests in the United States attributable to each of these five offenses in a given year. The one offense prominent in the ICE removal data and not available in the UCR data is driving offenses other than DUI.
Identifying Interior Removals
Key for this analysis is distinguishing removals of individuals apprehended while trying to enter the country from removals of individuals apprehended after settling in the country. Because Border Patrol officers at times apprehend individuals within 100 miles of the border who have already settled in the United States (e.g., American Civil Liberties Union, 2013; Garrova, 2018; Stillman, 2018) and because ICE at times apprehends individuals trying to enter the country, 11 it is not possible to simply use the agency involved in the apprehension to distinguish between interior and border apprehensions. Instead, I take a multipronged approach to identify removals of individuals who were apprehended after settling in the United States (i.e., interior removals). First, among the 2,041,542 removals where Border Patrol was not involved in the apprehension and the apprehension did not occur at a port of entry, I exclude all expedited removals (n = 34,213). These removals only occur in cases where the individual is apprehended within 14 days and 100 miles of the border (American Immigration Council, 2017a). Next, among the remaining cases where Border Patrol was not involved in the apprehension and the apprehension did not occur at a port of entry I compare the entry date and removal date and exclude an additional 113,449 removals where the individual was removed within 90 days of entering the country. 12 Last, to identify individuals likely to have been apprehended by Border Patrol after already residing in the country, I mark as an interior apprehension any removal that occurred more than 5 years after the recorded entry date and where Border Patrol was responsible for the apprehension (n = 70,876 cases). 13 In total, I estimate that there were 1,964,756 removals from the U.S. interior between fiscal years 2003 and 2015. 14
Results
Trends in Interior Removals
Of the 1,964,756 interior removals ICE conducted between 2003 and 2015, 15 approximately 94% of the individuals are male; 86% hail from Mexico or the Northern Triangle countries of El Salvador, Guatemala, and Honduras; 75% are between the ages of 20 and 40; and 79% initially entered the United States without authorization. With respect to the type of removal, 65% of individuals were ordered deported by an immigration judge, 24% had a prior removal order reinstated, and the remainder returned voluntarily. Descriptive statistics of all interior removals by year are presented in Appendix B.
Annual counts of interior removals changed considerably during this 13-year period, and these changes map onto shifts at the federal level in enforcement priorities. As illustrated in Figure 1, counts of interior removals increased annually between 2003 and 2009, peaking at 239,083 removals in 2009. The most dramatic increase in removals occurred between 2006 and 2008, with interior removals increasing roughly 120%. During this period, the federal government substantially increased its federal funding for interior immigration enforcement efforts (American Immigration Council, 2017b) and the number of formal partnerships between ICE and local law enforcement agencies expanded (ICE, 2013; Watson, 2013). Annual counts of ICE removals from the interior started to decline substantially following 2011—the year the Obama administration narrowed its criteria for who in the United States should be prioritized for deportation—and continued to decline in each year thereafter. By 2015, the number of interior removals fell below 2003 levels.
Annual counts of interior removals by gender, fiscal years 2003-2015.
Shifts in Types of Criminal Convictions Among Removed Individuals
As annual counts of interior removals increased in the first half of this period, the share of removed individuals with either no criminal conviction or only a misdemeanor conviction also increased. ICE classifies criminal convictions into three seriousness levels: the most serious (Level 1) covers the crimes listed in federal immigration legislation as “aggravated felonies.” Level 2 offenses cover other felonies, while Level 3 offenses are misdemeanors, including petty and other minor violations of the law. 16 Between 2003 and 2009, the share of individuals removed for the first time with either no conviction or where the most serious criminal conviction was classified as Level 3 increased from 59% to 77% of all first-time interior removals. 17 Once the Obama administration issued new guidelines prioritizing the removal of individuals convicted of Level 1 or Level 2 crimes (Morton, 2011a, 2011b), the share of individuals removed for the first time with no criminal history or with only a misdemeanor conviction began to decrease. By 2015, this group accounted for 44% of all first-time interior removals, with over 75% of these individuals having been convicted of a misdemeanor offense. Figure 2 illustrates these trends and shows that a much higher share of removed women throughout this period had no criminal history. Across all years, approximately 70% of women removed from the interior for the first time had no criminal convictions as compared to roughly 43% of men removed for the first time.
Interior removals by conviction level for men and women, FY 2003-2015.
Furthermore, as the annual number of interior removals increased, an increasingly large share of the individuals removed for the first time with a criminal conviction had a driving offense as their most serious conviction. Figure 3 presents the over-time trends for the six leading crime categories represented in the data. In 2003, of the individuals removed for the first time and who had a criminal conviction, roughly 1 in 3 (or 33%) had a drug-related conviction. 18 By 2015, drug-related convictions accounted for approximately 18% of first-time removals of individuals with criminval convictions. While the annual number of removals where the most serious criminal conviction recorded was drug-related (i.e., “drug-related removals”) increased during much of this period, the share of drug-related removals decreased considerably because of the increase in removals of individuals with DUI and other driving-related convictions, both of which are classified by ICE as Level 3 offenses. Between 2003 and 2011, the annual number of removed individuals with the most serious criminal conviction being a DUI increased more than sevenfold. In this same period, the annual number of removed individuals with a non-DUI driving offense as their most serious criminal conviction increased nearly 17-fold. By fiscal year 2015, the number of removals where the most serious criminal conviction recorded was either a DUI or some other kind of driving offense (i.e., “driving-related removals”) both dropped, although DUI remained the most common criminal conviction among first-time–removed individuals. With respect to the other common charges, the share of removed individuals convicted of a Level 1 assault offense remained relatively stable throughout this period as did the share of removed individuals convicted of a Level 1 sexual offense. 19
Leading crime categories for individuals removed from the interior, as share of total criminal removals each year, FY 2003-2015.
The decrease in the share of drug-related removals and increase in the share of driving-related removals do not simply reflect broader trends in law enforcement activity during this period of time. Analyses of data from the FBI’s UCR program on the number of arrests annually for drug possession, drug sales, DUI, and rape and other sex offenses suggest that while there are minor fluctuations year-to-year in the number of arrests for each of these offenses, overall, the share of arrests for each offense did not change considerably between 2003 and 2015 (see Appendix C). 20 As such, the increase in the share of removed individuals whose most serious criminal conviction is a DUI is a trend specific to the immigration enforcement arena. This trend maps onto the period in which local law enforcement agencies became more involved in immigration enforcement efforts through programs like 287g and Secure Communities. Whereas Congress had passed legislation as early as 1986 that authorized the issuance of immigration detainers for individuals arrested for a controlled substance offense and suspected of being unauthorized (Hernandez, 2014), it is not until the mid-2000s that there is an efficient and formalized process in place to enable a traffic stop to turn into an ICE arrest and, ultimately, a deportation. 21 And while there was a sharp decrease in the number of removals of individuals with non-DUI driving convictions following the Obama administration’s change in enforcement priorities in 2011, it is clear from these findings that individuals with a DUI conviction continued to be viewed by the federal government as priorities for deportation. 22
Differences by Region in Driving-Related Removals
Studies of how immigration enforcement unfolds in specific locales suggest uneven policing of Latinos and non-Latinos, especially when it comes to traffic stops (e.g., Armenta, 2017; Coleman, 2012; Gardner & Kohli, 2009; Schmalzbauer, 2014). To descriptively explore if this observation may bear out at the national level, I examine whether a the share of removed individuals with driving-related convictions was greater for individuals from Latin American countries than from all other regions of the world (i.e., “others”). Because the data do not include information on removed individuals’ ethnicity, I use nationality to approximate the Latino population. Additionally, I restrict the analysis to individuals with a criminal conviction at time of first removal and who entered the country without authorization, and therefore could not lawfully drive in most U.S. state. 23
Among Latin Americans, 22%, or more than 1 in 5, had a driving offense as their most serious criminal conviction. In comparison, among all others, 3%, or 1 in 33, had a driving offense as their most serious criminal conviction. Figure 4 breaks out these trends over time. While the share of removals of individuals with driving-related convictions is persistently higher for Latin Americans than for others, this disparity grows over much of the observed time period. In 2003, the share of removals of individuals with a DUI conviction was roughly 7% for Latin Americans versus 1.2% for others; by 2015 the share was 16% for Latin Americans versus 1.3% for others. For removals of individuals with non-DUI driving convictions, the share for Latin Americans versus others in 2003 was 2.1% and .8%, respectively, and peaked in 2011 at 12% and 2.1%, respectively.
DUI and non-DUI driving-related removals for Latin Americans and others, as share of total criminal removals each year, FY 2003-2015.
Method of Apprehension
As suggested by several of the site-specific studies of enforcement programs, it is also clear from these data that a large share of the individuals removed with either no criminal history or with a misdemeanor offense as their most serious conviction were initially brought to ICE’s attention because of local and state law enforcement agencies’ cooperation with ICE. Due to data quality issues, it is not possible to comprehensively identify the method of apprehension for all interior removals. However, between 2013 and 2015, there is reliable information about whether the CAP or 287(g) initiatives were involved in the initial apprehension. In all three years, the CAP initiative in local jails accounts for the greatest share of apprehensions leading to removal (see Appendix B). Among individuals removed after being identified through a local CAP initiative, 44% had a misdemeanor-level offense as their most serious conviction and 14% had no criminal convictions. With respect to 287g, 43% of all individuals removed after being apprehended through this program between 2013 and 2015 had a misdemeanor-level offense as their most serious conviction and 8% had no criminal convictions. 24 In terms of removals of individuals with a driving-related conviction, specifically, more than 71% of these removals between 2013 and 2015 started with an apprehension through either 287g or local CAP. Another 21% of these removals began with an apprehension as a result of some “other” method. It is not possible with these data to unpack the “other” category. However, Schmalzbauer’s (2014) ethnography of enforcement in Montana provides insight into how in places with strong informal ties between local police and ICE, programs like 287(g) and CAP are not required for a traffic stop to result in an ICE apprehension.
Discussion
Taken together, the findings from this study advance our understanding of U.S. interior immigration enforcement in several ways. First, this study provides rigorous estimates of interior removals between fiscal years 2003 and 2015—a period that began with a substantial ramping up of the federal government’s interior enforcement efforts and ended with a set of executive memos in place that more clearly circumscribed who from the interior should be prioritized for deportation. The most dramatic increase in removals occurred between 2006 and 2008, with interior removals increasing roughly 120%. Following the issuance of the Obama administration’s prosecutorial discretion memos in 2011, annual counts of interior removals started to decline substantially and continued to decline in each year thereafter. By 2015, the number of interior removals fell below 2003 levels.
Second, this study unpacks the government’s category of “criminal removal” to examine the specific offenses removed individuals have been convicted of and explore how the types of convictions have changed over time. Results show that as the number of interior removals increased in the first half of the observed time period, the share of removed individuals with no criminal convictions or with a misdemeanor conviction as their most serious criminal conviction increased. Following the implementation of the Obama administration’s prosecutorial discretion guidelines in 2011, a greater share of removed individuals had a felony conviction at the time of first removal. With respect to specific offenses, results show that the share of drug-related removals decreased over time while the share of driving-related removals increased. The fact that the share of drug-related and DUI arrests in the country did not fluctuate substantially from 2003 to 2015 suggests that the trends identified in this article are specific to the immigration enforcement arena.
Third, in terms of differences in conviction type by gender and region, this article shows that a greater share of removed women did not have criminal convictions as compared to men and that the increase in the share of driving-related removals nationally was greatest for Latin Americans. Among individuals from Latin America with a criminal conviction at the time of their first removal who had initially entered the United States without authorization, 22%, or more than 1 in 5, had a driving offense as their most serious criminal conviction. In comparison, among similarly situated individuals removed to all other regions, roughly 1 in 33 had a driving offense as their most serious conviction. Last, results illustrate the role of law enforcement partnerships in generating the arrests that lead to the removal of individuals with driving-related convictions. Among all individuals removed for the first time between 2013 and 2015 with a driving-related conviction as their most serious criminal conviction, more than 71% were first apprehended through either 287g or local CAP. Taken together, these findings largely reinforce those found in site-specific studies of immigration enforcement dynamics in areas with formal partnerships between ICE and local law enforcement.
Since taking office, President Trump has sought to dramatically increase immigration enforcement within the U.S. interior. As part of this effort, his administration has defined enforcement priorities so broadly as to ostensibly place all unauthorized migrants in the country at risk of deportation regardless of personal circumstance (American Immigration Council, 2017a) and renewed efforts to involve state and local law enforcement agencies in immigration enforcement, more than doubling the number of active 287g agreements within their first year and a half in office (ICE, 2018). In addition to redefining enforcement priorities and supporting further coordination between local and state law enforcement agencies and ICE, the Trump administration is also challenging efforts by states and localities to limit their cooperation with ICE. Most notably, they have begun suing states over sanctuary policies (Zapotosky, 2018). Jurisdictions with these policies tend to be less willing to honor ICE detainers and have been shown to have lower rates of immigration detention (Moinester, 2018).
While levels of interior apprehensions and removals under the Trump administration have yet to reach those seen during the height of interior enforcement in 2008 and 2009, the data-sharing infrastructure is clearly in place and the interagency relationships forged to enable a further ramping up of enforcement. As evidenced by removal trends in earlier years, if cooperation between local police and ICE goes unchecked, unauthorized immigrants, and especially unauthorized Latino immigrants, may likely be most affected. A large share of these individuals may have no criminal history or, at most, a misdemeanor conviction.
Increasing apprehensions and removals from the U.S. interior will have serious consequences not only for those removed but also for the families and communities they leave behind. Nearly two thirds of the more than 10 million unauthorized adults residing in the United States have lived in the country for at least 10 years and close to half are parents of minor children (Taylor, Lopez, Passel, & Motel, 2011). As such, these individuals are deeply entrenched within community, and research shows that these communities are negatively affected by heightened enforcement. Detentions and deportations from the U.S. interior create profound economic and emotional strain for affected families (Baum, Jones, & Berry, 2010; Capps, Castañeda, Chaudry, & Santos, 2007; Rugh & Hall, 2016); lead immigrants to withdraw from key institutions in American life, including the medical and education systems (Hagan, Castro, & Rodriguez, 2009; Satinsky, Hu, Heller, & Farhang, 2013); and erode trust between immigrants and police, undermining public safety (Menjivar & Bejarano, 2004).
Given that deportations from the U.S. interior are once again increasing and that there are clear social and economic consequences of these enforcement actions for individuals, families, and communities, there is a need for future research along several dimensions. For one, it is important to extend this analysis to include removals through the present. Doing so will help to illuminate the types of criminal convictions leading to current removals and the enforcement programs involved in these removals. Additionally, there is a need for more in-depth research studying local police department’s coordination with ICE across different locales and collaborative initiatives. As shown in this research, a large share of interior removals began with an apprehension through CAP. While there are a few studies of this program (Cantor, Noferi, & Martínez, 2015; Gardner & Kohli, 2009), more research is needed to better understand the program’s reach and the ways in which its presence in local jails, specifically, may affect policing. It is also clear from other research that high levels of enforcement have been achieved in locales with few, if any, formal partnerships on the books between ICE and local law enforcement (Moinester, 2018; Schmalzbauer, 2014). Future research should also explore how, if at all, local police support immigration enforcement efforts in these places. Last, this article presents important descriptive trends suggesting that as cooperation between ICE and local and state law enforcement agencies increased, the share of driving-related removals increased, especially for individuals from Latin American countries. Future research should seek out new data to test whether there is a causal impact of these interagency partnerships on types of arrests and on the number of Latinos arrested.
Footnotes
Appendix A
My counts of interior removals do not exactly match the numbers ICE began reporting in fiscal year 2008. As shown in Table A1 below, my estimates of interior removals differ from ICE in all years. This difference is greatest in fiscal years 2014 and 2015.
Differences likely occur for several reasons. First, the data used in this article are based on actual deportation data ICE recorded, unlike with ICE’s annual statistics where there can be a lag in reporting. 25 Second, current ICE records do not always replicate counts published in ICE’s historical statements, indicating data quality issues. Last, my strategy for identifying interior removals likely differs some from ICE’s, although it is challenging to assess the extent of these differences given that ICE does not detail their identification strategy. For example, in ICE’s annual reports, they define an interior removal as “individuals apprehended by ICE officers and agents in the interior” and border removals as “an individual removed by ICE who is apprehended by a CBP officer or agent while attempting to illicitly enter the United States at or between the ports of entry” (DHS, 2015). Given these definitions, it is unclear whether an individual who is apprehended by Customs and Border Patrol while residing in the United States and then removed is classified by ICE as a border or an interior removal, or whether such an individual is even recorded in ICE’s annual statistics. 26 To further complicate matters, in other reporting, ICE uses a different definition of border removals and provides estimates of border removals that do not match those given in their annual reports. For example, in a Congressional hearing on border security, then executive associate director of Enforcement and Removal Operations, Thomas Homan, defined border removals as “cases initiated by CBP, expedited removals, or removals of individuals within three years of entry into the United States” and said that ICE conducted 240,363 border removals in fiscal year 2012 (Homan, 2013, pg 21-22). This number is larger than that presented in ICE’s annual reports.
Appendix B
Descriptive Statistics of ICE Interior Removals, FY 2003-2015.
| FY 2003 (%) | FY 2004 (%) | FY 2005 (%) | FY 2006 (%) | FY 2007 (%) | FY 2008 (%) | FY 20009 (%) | FY 2010 (%) | FY 2011 (%) | FY 2012 (%) | FY 2013 (%) | FY 2014 (%) | FY 2015 (%) | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Deport group | |||||||||||||
| Reinstatement | 23.89 | 26.98 | 15.26 | 16.29 | 19.36 | 16.83 | 18.64 | 22.30 | 22.76 | 27.07 | 34.14 | 41.09 | 43.62 |
| Removal | 75.95 | 72.82 | 84.48 | 83.42 | 68.14 | 61.58 | 60.94 | 60.83 | 62.76 | 61.68 | 55.57 | 54.25 | 54.51 |
| Voluntary departure/return | 0.16 | 0.21 | 0.25 | 0.3 | 12.49 | 21.59 | 20.42 | 16.87 | 14.48 | 11.26 | 10.29 | 4.66 | 1.87 |
| MSCC group | |||||||||||||
| No conviction | 36.21 | 35.30 | 35.35 | 40.63 | 52.67 | 57.28 | 51.42 | 37.38 | 32.26 | 25.16 | 17.38 | 13.54 | 7.53 |
| Level 1 crime | 37.95 | 37.55 | 37.74 | 33.85 | 24.68 | 19.88 | 21.62 | 26.81 | 27.26 | 30.06 | 35.68 | 38.48 | 45.27 |
| Level 2 crime | 6.19 | 6.49 | 6.81 | 6.51 | 5.11 | 4.47 | 5.25 | 6.77 | 7.08 | 9.60 | 11.06 | 11.69 | 12.14 |
| Level 3 crime | 19.65 | 20.66 | 20.10 | 18.98 | 17.53 | 18.36 | 21.81 | 29.05 | 33.40 | 35.19 | 35.87 | 36.29 | 35.06 |
| Entry status | |||||||||||||
| Asylum | 0.16 | 0.07 | 0.05 | 0.06 | 0.04 | 0.04 | 0.10 | 0.20 | 0.21 | 0.29 | 0.31 | 0.38 | 0.35 |
| False claims | 0.21 | 0.23 | 0.21 | 0.16 | 0.14 | 0.10 | 0.10 | 0.11 | 0.10 | 0.11 | 0.12 | 0.14 | 0.12 |
| LPR | 1.28 | 1.72 | 1.85 | 1.66 | 1.42 | 1.12 | 1.46 | 1.92 | 2.35 | 2.86 | 3.68 | 3.73 | 3.28 |
| Not applicable | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.11 | 32.72 |
| Other | 46.90 | 35.79 | 30.10 | 22.56 | 14.21 | 20.76 | 7.05 | 5.87 | 4.73 | 3.60 | 2.74 | 2.22 | 1.91 |
| Present with admission | 5.20 | 6.81 | 7.36 | 6.52 | 6.27 | 5.02 | 5.24 | 6.34 | 6.39 | 6.62 | 6.58 | 6.72 | 5.23 |
| Present without admission | 46.26 | 55.38 | 60.42 | 69.03 | 77.91 | 72.95 | 86.05 | 85.56 | 86.22 | 86.52 | 86.56 | 86.71 | 56.39 |
| Gender | |||||||||||||
| Female | 8.06 | 8.15 | 8.35 | 8.94 | 7.92 | 7.32 | 5.99 | 6.02 | 6.26 | 5.46 | 4.76 | 4.55 | 3.96 |
| Male | 91.89 | 91.82 | 91.62 | 91.02 | 92.06 | 92.66 | 93.99 | 93.97 | 93.73 | 94.53 | 95.23 | 95.44 | 96.03 |
| Unknown | 0.05 | 0.03 | 0.03 | 0.04 | 0.02 | 0.01 | 0.01 | 0.02 | 0.01 | 0.01 | 0.01 | 0.01 | 0.01 |
| Apprehension method | |||||||||||||
| 287(g) | 13.29 | 12.25 | 10.55 | 9.00 | 8.81 | 7.86 | |||||||
| Border Patrol | 4.49 | 5.85 | 5.64 | 4.16 | 2.46 | 1.91 | |||||||
| CAP federal | 9.01 | 17.07 | 20.13 | ||||||||||
| CAP local | 49.72 | 43.81 | 35.60 | ||||||||||
| CAP state | 7.08 | 9.08 | 11.54 | ||||||||||
| Other | 82.23 | 81.91 | 83.81 | 20.30 | 18.77 | 21.98 | |||||||
| Place of origin | |||||||||||||
| El Salvador | 3.23 | 3.85 | 4.07 | 4.22 | 4.02 | 3.76 | 3.95 | 4.42 | 4.95 | 5.12 | 5.05 | 5.44 | 5.56 |
| Guatemala | 4.50 | 5.09 | 5.28 | 5.71 | 5.97 | 5.97 | 6.41 | 6.54 | 7.42 | 7.51 | 7.01 | 7.52 | 7.20 |
| Honduras | 3.23 | 3.31 | 3.75 | 4.05 | 4.32 | 4.52 | 5.03 | 5.90 | 6.21 | 6.11 | 6.42 | 7.01 | 6.75 |
| Mexico | 65.30 | 64.89 | 63.39 | 65.22 | 68.78 | 72.05 | 73.82 | 71.94 | 70.33 | 70.37 | 71.24 | 69.89 | 70.03 |
| Africa | 2.29 | 1.99 | 1.98 | 1.65 | 1.44 | 1.06 | 0.78 | 0.77 | 0.69 | 0.67 | 0.68 | 0.70 | 0.67 |
| Asia | 5.90 | 4.94 | 5.32 | 5.00 | 3.91 | 3.01 | 2.09 | 2.31 | 2.17 | 2.05 | 1.84 | 1.87 | 1.67 |
| Caribbean | 6.34 | 6.07 | 6.34 | 4.93 | 3.85 | 2.93 | 2.39 | 2.29 | 2.36 | 2.61 | 2.76 | 2.88 | 3.40 |
| Europe | 2.62 | 2.58 | 2.55 | 2.30 | 1.78 | 1.62 | 1.17 | 1.23 | 1.13 | 1.13 | 1.11 | 1.15 | 1.11 |
| Latin America | 5.73 | 6.47 | 6.53 | 6.26 | 5.43 | 4.71 | 4.01 | 4.22 | 4.40 | 4.07 | 3.46 | 3.08 | 3.09 |
| North America | 0.60 | 0.58 | 0.55 | 0.49 | 0.34 | 0.24 | 0.22 | 0.24 | 0.20 | 0.22 | 0.27 | 0.26 | 0.30 |
| Oceania | 0.25 | 0.24 | 0.24 | 0.17 | 0.16 | 0.14 | 0.12 | 0.13 | 0.14 | 0.14 | 0.16 | 0.20 | 0.21 |
| Deported previously | |||||||||||||
| No | 69.69 | 65.55 | 64.89 | 62.94 | 64.02 | 64.57 | 59.06 | 56.07 | 54.56 | 51.05 | 46.48 | 42.87 | 42.26 |
| Yes | 30.31 | 34.45 | 35.11 | 37.06 | 35.98 | 35.43 | 40.94 | 43.93 | 45.44 | 48.95 | 53.52 | 57.13 | 57.74 |
| Total | 89,386 | 97,158 | 95,220 | 105,209 | 165,694 | 231,258 | 239,083 | 222,819 | 222,579 | 180,091 | 132,750 | 108,416 | 75,093 |
Note. ICE = Immigration and Customs Enforcement; FY = fiscal year; MSCC; Most Serious Criminal Conviction; LPR = lawful permanent resident; CAP = Criminal Alien Program.
Appendix C
Declaration of Conflicting Interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Multidisciplinary Program in Inequality and Social Policy at Harvard University.