How Local Structural Conditions Influence the Adoption of 287(g) Restrictionist Immigration Policy

ICE has used their 287(g) program to target immigration enforcement at the county level. This program authorizes local police officers to carry out federal immigration work to meet arrest and deportation quotas. This has eroded civil rights and led to the criminalization of minorities, particularly Latinos. While previous research has examined the theories behind local anti-immigration work and the social factors affecting anti-immigrant policy adoption more generally, no large-scale quantitative analyses have been conducted as to why some counties adopted 287(g) while others chose not to do so. Addressing this gap in the literature, the following study uses a newly gathered data set of all 287(g) agreements signed since 2002. This data set helps to examine geographic variation in local structural forces shaping adoption of the 287(g) program.

CHAPTER I: INTRODUCTION

The Immigration and Customs Enforcement agency, also known as ICE, has brought immigration enforcement to U.S. counties for the past 20 years through their 287(g) program. They did so by promoting local police officers to double as federal agents to perform federal immigration checks. In practice, this has meant the infringing of rights, criminalizing of minorities, and much more under the vague premise that a person looks like or appears to be illegally in the United States. Across the disciplines of sociology, geography, and political science, scholars have researched the intricacies of the consequences of this program. Legal coalitions and pro-immigrant think tanks have helped sue and force restructures of the program.

Sheriffs in U.S. counties are publicly elected officials. This means that locals get to decide who they will support, and this person has the ability to choose whether their county, through their sheriff’s office, will sign onto ICE’s 287(g) program. A myriad of social forces shapes a county’s population and their attitudes towards immigrants, which in turn may influence the local sheriff’s office decision to adopt 287(g) policy. Numerous studies have researched the factors affecting immigration enforcement. These studies range in their theoretical emphases but generally agree on the importance of social forces related to racial and group threat, local socioeconomic conditions, political culture, and local legacies of previous anti-immigrant organizing. And yet, there have been no large scale, quantitative analyses examining 287(g) policy adoption, in particular. Moreover, in U.S. counties across the nation, there is quite a bit of variation in this adoption, which begs the question: Why do we see so much geographic variation in counties’ adoption of restrictionist immigration policies like 287(g)? This study deploys a unique quantitative data set of all 287(g) agreements signed since the first one in 2002 to examine this social phenomenon and shed light on America’s fragmented enforcement of 287(g).

CHAPTER II: LITERATURE REVIEW

The birth of the 287(g) program came through the 1996 Immigration and Nationality Act with the goal of “enhanc[ing] the safety and security of communities by creating partnerships with state and local law enforcement agencies to identify and remove aliens who are amenable to removal from the United States” (ICE 2020). The program at first was an initiative claiming to promote preparedness against foreign terrorist attacks (Dubina 2005). The first negotiations to reach an agreement to implement 287(g) were in the 1990s with Salt Lake City in Utah (Capps et al. 2011). This county opted out due to strong concerns of racial profiling (Capps et al. 2011). The first agreement was signed in 2002 with the state of Florida mere months after 9/11 (Dubina 2005). Within the first five years there were fewer than 10 agreements, also known as Memorandums of Agreement (MoA), signed. In an average year during the Obama administration there were approximately 56 counties signed up; during the last 3 years this decreased to barely above 30. Within the first 4 months of the Trump administration, however, this number doubled to 60. The average year under the Trump administration doubled Obama years’ average to over 100, skyrocketing up to 150 counties signed on to 287(g) as of October of 2020 (U.S. DHS). ICE has widely showcased their annual deportation quotas by targeting, arresting, and deporting four-hundred thousand immigrants per year (Heath 2013). This ever changing and fragmented adoption of the policy, generated my research question: What social factors are associated with a county’s adoption of 287(g)? I expect local structural conditions related to racial and group threat, local socioeconomic conditions, political culture, and local legacies of anti-immigrant organizing will influence 287(g) implementation.

Since its inception, 287(g) was almost immediately met with wide skepticism on a number of fronts. First, there arose the broad legal question of whether it is the local police department’s duty to prosecute federal immigration offenses and whether such departments were capable of doing so. Relatedly, there were concerns about unintended consequences such as mischaracterization of individuals through police bias, and racial profiling (Capps et al. 2011; Coleman and Kocher 2019; 2011). Law professor Shoba Wadhia questions to which extent local and state police can or should enforce immigration laws (2019), and whether such mis-assignation of duty may “otherwise pose a risk to public safety or national security” (p. 31). Such risks might manifest themselves in the form of increased mistrust of police as well as law enforcement’s misapplication of policy and abuse of rights (Rodriguez et al. 2010; Shapira 2013). Additionally, this dynamic is intrinsically racially problematic because:

Lawful or unlawful presence in the United States is a legal distinction that is not discernable to a law enforcement official… Someone could even be a U.S. citizen and not be knowledgeable of citizenship status [so] it is impossible for a law enforcement official to determine a person’s legal status based on her appearance.

Golash-Boza 2015 p.42

As a result, the policy has had the unintended consequences of further racializing local policing and immigration, often to the detriment of local minority populations (Gonzales 2016; Golash-Boza 2014, 2015; Vickers 2019; Conlon and Hiemstra 2016; Donato and Rodríguez 2014). For instance, Walker and Leitner (2011) performed a municipal-level study through textual analysis of policy. They found 287(g) adoption reflected an area’s stance on race relations, national pride, and community understandings of their region. For example their analysis showed how a growing immigrant population is a predictor of local immigration policy adoption as such municipalities are more likely to introduce restrictionist programs.

Following in Walker and Leitner’s footsteps, my study broadens the analysis further, extending it to counties across the United States. While it is true that — at a broad level and over 20 years — local police departments and federal ICE agents have collaborated increasingly through 287(g) (Gonzales 2016), the adoption and enforcement of policies like 287(g) by local law enforcement remains uneven across the United States. This is largely because, under the direction of President Barack Obama, immigration policy underwent significant changes that provided states and counties greater discretion over enforcement and implementation of programs like 287(g) (Gonzales 2016; Michaud 2010). This meant that some areas of the country quickly adopted 287(g) such as Texas, California, and North Carolina. However, many areas went in the opposite direction, decreasing police collaboration with ICE (Gonzales 2016; Pedroza 2019; Michaud 2010). For example, as of 2012, Los Angeles only offers limited cooperation with ICE and police, no longer providing citizenship information from courts or hospitals. Philadelphia and Baltimore are among other metro areas blocking immigration status reporting to ICE in so called sanctuary municipalities (Golash-Boza 2015; Gonzales 2016).

But what explains these differences? Why is it that some local communities moved to adopt 287(g), while others rejected it? Answers to this central question can be grouped into four broad categories: (a) racial threat dynamics related to local demographics, (b) local political culture, (c) local histories of anti-immigrant organizing, and (d) local socio-economic conditions. Below, I review the relevant literature and discuss its relation to cross-county variation in 287(g) adoption.

Racial and group threat

Literature in the group threat tradition demonstrates how conflict between groups — particularly between/ethnic majority and minority groups — arises, often attributing it to: competition over scarce resources and shifts in local demographic conditions. Such shifts can include changes in minority group sizes and the accompanying perceptions of threat that this induces in majority group members (Blalock 1967; Olzak 1990, 1992; Quillian 1995, 1996; Pinderhughes 1993).

An increase in total or segmented increases in population are thought to increase competition and proximity of minorities with majority group members. This social dynamic generates frustration and hostility against minorities specifically when money is limited through this competition (Medoff 1999; Olzak 1992; Quillian 1995; Tolnay and Beck 1995). In other words, worse living conditions — often indicated by lower income and higher unemployment — will damage group relations and push competing groups into heated rivalry.

Following the power-threat theory, areas with a higher percentage and greater influx of immigrants have been shown to engender white backlash (Blalock 1967; Tolnay, Beck and Massey 1989). When the dominant group’s power is perceived to be challenged this results in more antipathy towards immigrants and other challengers such as other people of color (Dixon 2006; Quillian 1995; Semyonov et al 2004). The scholarship behind the importance of the racial and ethnic composition in a given community depicts the effects of minority changes in population and immigration, yet still leaves us to address whiteness. We know that conflict and anti-immigrant sentiments rise as the minority population rises. When the minority population becomes larger than that of the white locals, their perception of dominance lowers (South and Messner 1986; Galster 1990; Clark 1993).

Besides how this affects the white population, Stacey, Carbone-Lopez, and Rosenfeld (2011) show how this finding extends further and is now associated also with Latino immigration not just local people of color. Latinos are more comfortable with the social climate surrounding them when there is more representation at their local leadership level (Martinez 2008).

Anti-immigrant movements are especially active where the percentage of people of color is much lower due to a belief of group superiority (Levine and Campbell 1972; Jacobs and Wood 1999). On the other hand, minorities tend to feel safest in areas where they are a large percentage of the total population (Disha et al. 2011). It is therefore noteworthy that conflict against minorities is more prevalent where not only are minorities a larger part of the population but also there is more flow of them causing a change in racial population (Levine and Campbell 1972; Bobo 1988).

In terms of immigration, scholars believe that a growth in the Latino population has lead to an increase in restrictive immigration control, often stemming from an uneven distribution of economic factors of immigration at the local level (Borjas 2014; Wong 2012; Rugh and Hall 2016). Black U.S. citizens are also a target of the 287(g) program and are therefore less supportive of stricter immigration enforcement unlike white Americans (Chiricos et al. 2014). Drawing on the scholarship of racial and group threat, I expect local demographic conditions to be a determinant in the adoption of ICE’s 287(g) program through the specific measures of white population percentage, and Latino and African American population percentage and percent change.

Local socioeconomic conditions

Following the group-threat paradigm, the study of local immigration enforcement is linked with the theory of resource mobilization. This theory explains that for restrictionist mobilization to take place there must be sufficient resources to be able to afford such initiatives (McCarthy and Zald 1977). This would mean that counties with lower median household income would have fewer resources for new federal immigration policing costs. This theory would expect poorer communities, regardless of demographic, to be less likely to implement the 287(g) program due to its high costs.

Damaged group dynamics from worse economic conditions are linked closely to unemployment. One would theorize that anti-immigrant groups would then be white and poor. Nevertheless, anti-minority organizing is in fact linked to communities of wealthy economic conditions due to very limited proximity to minorities and a tendency to hoard opportunities (Sattin-Bajaj and Roda 2018; Green, McFalls and Smith 2001; Lyons 2007, 2008). In other words, privileged social groups try to control available resources and prevent minorities from accessing them. Nevertheless, when unemployment is greater, this means that a place’s economy is weaker, which means its resource base is weaker with fewer resources to put towards immigration enforcement (Disha et al. 2011; Lyons 2007, 2008). Therefore, resource mobilization theory expects median household income to be positively associated with implementing 287(g), and unemployment rates are negatively associated with taking up local immigration enforcement initiatives like the 287(g) program.

Local political culture

The Immigration Industrial Complex understands the flow of migrant workers as a large profit to American industry (Golash-Boza 2015). This theory proves that there are interests in criminalizing and marginalizing the undocumented through immigration law enforcement and mass anti-immigrant work (Golash-Boza 2015). This structural pattern uses a rhetoric of fear, a confluence of powerful interests, and a discourse of ‘other’-ization (Golash-Boza 2015; Hiemstra 2019; Bearman and Bruckner 2001). Politics also affect the adoption of local immigration policies. Some argue that conservative localities are twice as likely than Democratic localities to generate and pass restrictionist ordinances (Ramakrishnan and Wong 2010; Chiricos et al. 2014; Lewis et al. 2013; Fennely 2006). More importantly, the flagship 2011 study performed by the Migration Policy Institute found that the 287(g) program was used by local elected sheriffs in the interest of their anti-immigrant agendas (Capps et al. 2011). Republican activism engages in anti-immigrant initiatives out of fear for their resources (Beck 2000; Tilly 1978; McVeigh 2009; Van Dyke and Soule 2002). Overall, anti-immigration acts have in the last decade been associated with civil rights organization resources, political opportunity, and affiliation (McVeigh 2003; Disha et al. 2011; Hanes and Machin 2014). Following the political opportunity perspective, I expect the percentage of local Republican voters to be positively associated with 287(g) implementation.

Local legacies of anti-immigrant organizing

Anti-immigrant movements are supported and bolstered by institutions such as ICE and Customs and Border Patrol (Cohen 2020). Political parties “might respond in favor of anti-immigrant sentiment and institutionalize xenophobia with nativist policies” (Filomeno 2017, p. 3). For Filomeno, “partisanship of a local administration … is more than the partisan affiliation and political ideology of local policy-makers; it is their membership in a national political network” and most importantly, “partisanship is a key variable in current explanations for variation in local immigration policy,” affiliated to conservative ideology specifically (p. 26; Campbell et al. 2006).

In areas where the above-mentioned social factors coincide, anti-immigrant hate groups arise mirroring the policy and local interests of their counties. For example, median household income is positively linked with nativist groups since their members are often privileged middle class and non-Latino whites (Doty 2009; McVeigh et al. 2014; Shapira 2013; Skocpol and Williamson 2013; Stewart, Bendall, and Morgan 2015). The historical legacy of anti-immigrant organizing, often through the presence of extreme nativist groups, is expected to be positively associated with a county’s odds of signing onto ICE’s 287(g) program. Put simply, areas with histories of nativist organizing should have a greater chance of reproducing anti-immigrant behaviors and attitudes over time that would also be consistent with the eventual adoption of 287(g) policy later on.

In conclusion, I have laid out the theoretical frameworks surrounding the social factors known to affect an anti-immigrant policy adoption: a) Racial and group threat, (b) local socio-economic conditions, (c), local political culture, and (d) local legacies of anti-immigrant organizing may all affect a county’s odds of implementing ICE’s 287(g) program.

CHAPTER III: METHOD AND DATA

Dependent variable — implementing 287(g)

Using reported Memorandums of Agreements (MoA) data located on ICE’s website from 2002 to 2020, I create a single, aggregate binary dependent variable coded ‘1’ if the county implemented 287(g). My dependent variable includes a reading of every county in U.S. states (3074 observations), excluding territories and independent municipalities. Doing so allows me to assess the local structural conditions shaping 287(g) implementation. I focus on counties because ICE’s 287(g) agreements are signed by local sheriff’s offices, which typically represent single, defined counties. My data includes four cross-sectional readings of the list of active counties per year. I took six readings every three years to account for any patterns in ICE’s misreporting. Using this data-gathering method I created an exhaustive database of all agreements reported since the very first one.

Independent variables

Racial and group threat (%White, %Black, %Latino, % change in Latino, and % change in Black population)

Racial threat can come from the (1) sheer proportion of minorities in an area, and from (2) quick changes in the proportion of the population. I constructed the variables of percentage of Black, white, and Latino, together with percent change of Black and Latino populations from 1990 to 2000, prior to the start time of my dependent variable. Data for the %white, black and Latino variables were derived from the 2000 decennial census. The % changes in the Latino and black populations were calculated by subtracting 1990 decennial census data from the 2000 decennial census data (U.S. Census Bureau). Decennial census data for 1990 and 2000 were used for the purposes of ensuring proper time ordering between my independent variables and the dependent variable. In other words, these independent variables were measured before the start time of my dependent variable. I predict percentage of white population to be negatively associated with 287(g) implementation, while my two percent change variables of Black and Latino should be positively associated with 287(g) implementation. The threat paradigm suggests %Latino and %Black will be positively related to 287(g) adoption.

Local resources and socio-economic conditions (median household income, unemployment)

I also consider how variation in local resources and socio-economic conditions could affect variation in counties’ 287(g) implementation. The theoretical framework of resource mobilization explains that for mobilization such as 287(g) to occur, there must be sufficient resources. If localities, such as the counties in my study, do not have enough resources as measured by low employment and low median household income, then it is unlikely a sheriff’s office has resources to deal with additional costs of taking up federal immigration enforcement through 287(g). Additionally, areas with rising unemployment, as previously detailed, will have a population more likely viewing racial minorities as competitors for jobs and threatening. Therefore, median household income and unemployment serve as specific measures to test the variation in 287(g) implementation. I gather this data from the 2000 census such as indicated in Table 1 as a snapshot prior to the start of 287(g). I predict local median household income to be positively association with 287(g) implementation. Unemployment should be negatively associated with 287(g) adoption.

Local legacies of anti-immigrant organizing (the count of extreme nativist groups)

Areas with a prior history of nativist organizing are more likely to be supportive of 287(g) policy adoption. Areas with more hate groups are also likely to have more extreme nativist groups. Prior presence of nativist groups can appear to be associated to with increased odds in adoption 287(g) in those counties. Areas with a history of nativist organizing are likely to continue to promote nativist policies. These counties with a history of anti-immigrant organizing will more than likely have the organizing and resource structures in place to promote nativist policies like 287(g). I used 2008–2009 data of active extreme nativist groups from the Southern Poverty Law Center’s “Intelligence Report” (2010). These groups can react to local migration flows but also mirror already-existing anti-immigration ordinances. I created a continuous variable that serves as the count for number of extreme nativist groups present in each county. I expect the count of Extreme Nativist Groups to be positively associated with the implementation of 287(g).

Local political culture (%republican voters)

The Republican Party represents many conservative values. Conservative supporters of Republican candidates tend to align with anti-immigrant initiatives. Therefore, the percentage of republican voters can tell us a lot about the potential level of anti-immigrant antipathy in a county. In this case, I measure an area’s level of conservative culture by the percentage of the population that voted for the Republican candidate in the 2000 presidential election, which is the most recent presidential election prior to the measurement of the dependent variable. This data comes from MIT’s Election Data and Science Lab (2018). I predict that the percentage of republican voters will be positively associated with implementing 287(g).

Analytic strategy

I used a logistic regression analysis to test my hypothesis using the statistical software Stata 13. My statistical model is adequate for this analysis because when researchers analyze quantitative data with dichotomous categorical dependent variables, logistic regression analysis is considered appropriate (Agresti 2002; Allison 1999). Results report the odds ratio for a one-unit change in each independent variable.

Analysis

The model analyses the validity of the claims of my hypothesis. The independent variables were analyzed in the order that the research questions lead the research process. In other words, demographic variables were analyzed first which led to testing other social predictors. The model incorporates independent, dependent, and control variables to measure the propensity of adopting the 287(g) program. Table 2 shows the descriptive statistics of the array of variables:

Control variable (total population). All such models control for total population. Doing so helps to check the validity of other populations’ associations on the dependent variable.

CHAPTER IV: RESULTS

From the descriptive statistics depicted in Table 2, the dependent variable of 287(g) has a mean of 0.05. This means that 5% of counties adopted 287(g). One can notice noteworthy facts such as a mean of extreme nativist groups in U.S. counties between 2007 and 2008 near 1 with a value of 0.76. The Latino and Black population represent a 6.24 and 8.5 percentage respectively of the U.S. population. The white only population has a mean percentage of 81.75. The mean of the change in the Latino population is 1.69 percent increase which is higher than the 0.12 percent increase in the Black population prior to the start of the 287(g) program. These general statistics inform us with a broad brush of what the environment was prior and during ICE’s 287(g) program.

I used the “logistic” command in Stata 13 to test the odds of the logistic regression model. The analysis provides odds-ratios for each variable. I can interpret these by first understanding the nature of the model. The odds ratio Wald chi-square of 208.29 with a p-value of 0.00001 shows that my model is overall better than one without the included predictors. From observing the p-values labeled as “P > | z |” in Table 3, one can notice variable results that are and are not statistically significant: smaller or equal than 0.05, and larger than 0.05 respectively. The variables of Unemployment Rate, Total Population, %Latino, %change Latino, %change Black, and Republican Voters are all statistically significant, with the latter five listed being only marginally so since their associated p-value is smaller than 0.1.

Racial and group threat (%White, %Black, %Latino, % change in Latino, and % change in Black population)

The percent Latino variable has an odds ratio coefficient of 1.31 and is marginally statistically significant with a p-value of approximately 0.1. The magnitude of the coefficient is particularly large. In fact, a one percentage point increase in the Latino population is associated with a roughly 30 percent increase in the odds of 287(g) implementation. This aligns with theories of racial threat that suggest larger Latino populations will engender greater perceptions of threat, which in turn will lead to greater efforts at socially controlling them through, for instance, restrictionist policies like 287(g).

The percent White variable has an odds ratio of 1.27. For every unit increase in %White, the odds of implementing this program increase by 27 percent. This positive association runs counter to the hypothesized negative association; however, the result is not statistically significant. Percent Black has an odds ratio of 1.31. Like our %Latino predictor, it is positively associated with the odds of 287(g) adoption in a county; it therefore is in line with the expected theoretical prediction; however, it is not statistically significant.

Percent change in the Latino population has an odds ratio of 0.54, meaning it displays a negative association with the dependent variable. For every one unit increase in %change in Latino population in a county, this would decrease the odds of 287(g) adoption by 46%. The magnitude of the coefficient is quite strong, and the finding is marginally statistically significant since its p-value is under 0.1. I predicted for this association to be negative. Previous scholarship has argued that when a minority population grows rapidly at a time, this community becomes more politically powerful and influential, and our more able to protect their interests (South and Messner 1986; Galster 1990; Clark 1993). This would explain the nature of this association from the rapidly growing Latino population could help to guard against the adoption of restrictionist anti-immigrant policies like the 287(g) program.

Percent change in the Black population, similarly to %change Latino, also has a negative association with the odds of 287(g) implementation. Its coefficient is 0.47 meaning that for every unit increase, it would decrease the odds of signing onto the program by 53%. This finding is also marginally statistically significant with a p-value of 0.059; not far from the commonly accepted value of 0.5 to be considered significant. While the association runs counter to the hypothesize relationship, there is perhaps, a good theoretical explanation for these findings. Building on how minority population’s power and community engagement rises as their representation increases, this may be similar with the black population. More diverse areas may operate as a safeguard for minorities through their organizing. These findings suggest that an increase would mean making white inhabitants have more diversity in their county, possibly making it less likely to implement the 287(g) program.

Local resources and socioeconomic conditions (median household income, unemployment)

The effect of median household income on 287(g) policy adoption is virtually non-existent, as well as statistically insignificant. This result is understandable since having the resources to organize is not limited by income and often can depend on the history of previous anti-immigrant initiatives as a local tendency.

The unemployment rate, on the other hand, is strongly and negatively associated with 287(g) adoption. In Table 3 we see how for every one unit increase in unemployment rate, there is 22% decrease in the odds of a county implementing 287(g). This result is in line with the hypothesized association discussed previously. As unemployment increases, the area’s economy is weaker, therefore limiting their available resources to take up federal immigration enforcement duties such as the 287(g) program. This finding is supported by the resource mobilization theory.

Local legacies of anti-immigrant organizing (extreme nativist groups)

The predictor variable of Extreme Nativist Groups has and odds ratio of 1.14. This coefficient shows a positive association between the presence of extreme nativist groups in a county with the odds of that county implementing ICE’s 287(g) program. Therefore, every one unit increase in the presence of extreme nativist groups increases the odds of 287(g) adoption by 14%. While this variable is not statistically significant, the result is in the expected direction.

Local political culture (%republican voters)

The variable percent Republican Voters has an odds ratio of 1.02. Though less than in other predictor variables, Republican Voters have a weak positive association with the odds of affecting 287(g) adoption in a county. This result aligns with the political opportunity theory and general understandings of local political culture. This evidence supports the hypothesis and is marginally statistically significant with a p-value of 0.058.

Control variable

My control variable, Total Population, returns an odds ratio of 1. This means that for a one unit increase in Total Population the odds of 287(g) implementation increase by a very small amount. While statistically significant with a p-value of 0.031, there is not real theoretical interest in this finding.

CHAPTER V: DISCUSSION AND CONCLUSION

The 287(g) program has turned federal immigration enforcement into a local threat to immigrants for the past 20 years. This has created an array of issues such as racial profiling of people of color and criminalizing of these communities. Sheriff’s offices in U.S. counties have had to decide whether to sign onto this ICE program through their Memorandums of Agreement. One would wonder what would motivate local enforcement departments to want to adopt this program. What local structural factors are associated with a county’s adoption of 287(g)? Theoretical explanations focused on: (a) racial threat dynamics related to local demographics, (b) local political culture, © local legacies of anti-immigrant organizing, and (d) local socio-economic conditions.

Based on these data, geographic variation in 287(g) implementation is driven largely by racial threat. The odds of implementing this program is tied specifically to the size of the Latino population. The scholarship of racial threat helped predict a positive association between this implementation and rise in the Latino population, which my data analysis supports. The rapid growth of the minority population appears to have an insulating effect, in this study, rather than a threatening effect. In other words, the higher representation, above a certain degree, secures minorities’ political presence as a community at the county level.

Local political culture is also greatly relevant to the odds of a county choosing to adopt 287(g). Areas with greater percentages of Republican voters are more likely to implement policies like 287(g) that are aimed at socially controlling minority populations. Of course, the outcome is also shaped by the available local resource base of a county. This is indicated by the results linked to the local unemployment rate. Whether Republican voters are likely to adopt 287(g) may crucially hinge on whether they have the local resources to do so. Clearly, these results indicate, a certain level of local resources is necessary for anti-immigrant and restrictionist mobilization.

Although producing many novel insights about variation in 287(g) implementation, this study is not without limitations. The results of this study focus on the analysis of secondary quantitative data. It, therefore, misses out on more micro level social factors that may affect 287(g) implementation. Another methodology, perhaps more qualitative in nature (like interviews) could provide the sort of refined data necessary to tease out these social processes. Additionally, my data largely stems from the U.S. Census Bureau decennial census. Like all census data, it is limited by the circumstance of who could and could not answer the questions, although it should be reasonably comprehensive. Additionally, the most novel aspect of this study’s data is the newly gathered archive of all 287(g) agreements signed since the very first one. Using the WaybackMachine from archive.org is the most reliable way, through an algorithm, of gathering unbiased snapshots.[1] Nevertheless, the data of my dependent variable is limited by the many mistakes in reporting on the ICE website. No patterns were found of which time of the year updates would be made. Throughout the process of gathering this data I found almost 20 mistakes that resulted in missed values. These missed values ranged from non-existent names of counties to allocating the wrong state next to a county’s name. It is worth noting that I have treated 287(g) policy adoption as something that occurs exclusively at the county level. However, this can be problematic since the 287(g) program has only more recently become more of a county level decision making process after their latest MoA remodel. By extension, many different organizations could adopt the program that were not county’s sheriff offices.

Nevertheless, this study provided much needed insight into what social factors affect the odds of a county implementing ICE’s 287(g) program. One of the benefits of such a study is that the results can be used to lay the ground work for fruitful policy work in the future. Such policy can keep ICE accountable to their misreporting of what should be publicly available data. Not only so, but it promotes an environment of accountability to their actions. My results have implications such as recognizing the value of minority representation but also recognizing the power in anti-restrictionist organizing. Anti-immigrant practices such as 287(g) can be fought against through policy. New policies can help by creating the infrastructure through national support for areas were minority population and therefore power is very low. Future research will want to refine the analyses conducted here. In particular, a more micro-level study will be able to help flesh out the social processes linking together the local structural conditions highlighted here with the adoption of policies like 287(g).

[1] ICE does not provide a comprehensive list of counties that adopted the 287(g) program across all years. Therefore, I had to use the WayBack time machine (archive.org) to be able to go back in time and gather screen shots of earlier periods from 2002 up to the present. This site uses an algorithm to take a large number of screen shots when it detects a change in the website. This makes the information easy to track back throughout the research process.

REFERENCES

Bearman, Peter and Hannah Bruckner. 2001. “Promising the Future.” American Journal of Sociology 4:859–912.

Beck, E. M. 2000. “Guess Who’s Coming to Town.” Sociological Focus 33:153–73.

Bobo, Lawrence. 1988. “Group Conflict, Prejudice and the Paradox of Contemporary Racial Attitudes” in pp. 85–114 Eliminating Racism, Katz, Phyllis and Dalmas Taylor (eds.). Plenum Press.

Borjas, George J. 2014. Immigration Economics. Cambridge, MA. Harvard University Press.

Blalock, Hubert M. 1967. Toward a Theory of Minority-Group Relations. New York: Jon Wiley and Sons.

Campbell, Andrea L., Cara Wong, and Jack Citrin. 2006. “Racial Threat, Partisan Climate, and Direct Democracy.” Political Behavior 28:129–50.

Capps, Randy, Marc R. Rosenblum, Cristina Rodríguez, and Muzaffar Chishti. 2011. Delegation and Divergence: A Study of 287(g) State and Local Immigration Enforcement. Washington, D.C. Migration Policy Institute.

Chiricos, Ted, Elizabeth K. Stupi, Brian J. Stults, and Marc Gertz. 2014. “Undocumented Immigrant Threat and Support for Social Controls.” Social Problems 61(4):673–92.

Clark, W.A.V. 1993. “Neighborhood Tipping in Multiethnic/Racial Context.” Journal of Urban Affairs 15:161–72.

Cohen, Elizabeth F. 2020. Illegal: How America’s Lawless Immigration Regime Threatens Us All. New York, NY. Basic Books.

Coleman, Mathew and Austin Kocher. 2011. “Detention, deportation, devolution and immigrant incapacitation in the US, post 9/11.” The Geographical Journal. 177(3):228–237.

Coleman, Mathew and Austin Kocher. 2019. “Rethinking the ‘Gold Standard’ of Racial Profiling: §287(g), Secure Communities and Racially Discrepant Police Power.” American Behavioral Scientist. 63(9):1185–1220.

Conlon, Deirdre and Nancy Hiemstra. 2017. Intimate Economies of Immigration Detention: Critical perspectives. New York, NY. Routledge.

Cornelius, W. 2010. Preface. In Taking Local Control: Immigration Policy Activism in U.S. Cities and States, ed. M. Varsanyu, 73–96. Stanford, CA: Stanford University Press.

Disha, Ilir, James Cavendish and Ryan King. 2011. “Historical Events and Spaces of Hate.” Social Problems 58(1):21–46.

Dixon, Jeffrey C. 2006. “The Ties that Bind and Those that Don’t.” Social Forces 84(4):2179–2204.

Donato, Katharine M., and Leslie Ann Rodríguez. 2014. “Police Arrests In A Time Of Uncertainty: The impact of 287(g) on arrests in a new immigrant gateway.” American Behavioral Scientist 58(13):1696–1722.

Doty, Roxanne L. 2009. The Law into Their Own Hands. Tucson, AZ: The University of Arizona Press.

Dubina, Mark F., Special Agent Supervisor, Tampa Bay Regional Operations Center, Florida Department of Law Enforcement, Regional Domestic Security Task Force (testimony). The 287(g) Program: Ensuring the Integrity of America’s Border Security System Through Federal-State Partnerships [2005] 109th Cong., 1st Session (House Committee on Homeland Security, Subcommittee on Management, Integration, and Oversight), p.17. https://www.govinfo.gov/content/pkg/CHRG-109hhrg28332/pdf/CHRG-109hhrg28332.pdf.

Farris, Emily M., and Mirya R. Holman. 2016. “All Politics Is Local? County Sheriffs And Localized Policies Of Immigration Enforcement.” Political Research Quarterly 70(1):142–154.

Fennely, K. 2006. State and Local Policy Responses to Immigration in Minnesota. New York, NY: Century Foundation.

Filomeno, Felipe Amin. 2017. Theories Of Local Immigration Policy. Palgrave Macmillan.

Galster, George C. 1990. “White Flight from Racially Integrated Neighborhoods in the 1970s.” Urban Studies 27:385–99.

Golash-Boza, Tanya Maria. 2014. Exclusionary Immigration and Citizenship Policies. In Race and Racisms: A Critical Approach.

Golash-Boza, Tanya Maria. 2015. Immigration Nation: Raids, Detentions, And Deportations In Post-9/11 America. New York: Routledge.

Gonzales, Roberto. 2016. Lives in Limbo: Undocumented and Coming of Age in America. Oakland, Calif.: University of California Press.

Green, Donald, Laurence McFalls, and Jennifer Smith. 2001. “Hate Crime.” Annual Review of Sociology 27:479–504.

Hanes, Emma and Stephen Machin. 2014. “Hate Crime in the Wake of Terror Attacks,” Journal of Contemporary Criminal Justice 30(3).

Heath, Brad. 2013. “Immigration Tactics Aimed at Boosting Deportations.” USA Today, February 17.

Hiemstra, Nancy. 2019. Detain and Deport: The Chaotic U.S. Immigration Enforcement Regime. Athens, GA. University of Georgia Press.

ICE. 2020. Department of Homeland Security (Agreements of Cooperation in Communities to Enhance Safety and Security) section 287(g).

Jacobs, David and Katherine Wood. 1999. “Interracial Conflict and Interracial Homicide?” American Journal of Sociology 105(1):157–90.

Levine, Robert A. and Donald. T. Campbell. 1972. Ethnocentrism. New York: Wiley.

Lyons, Christopher J. 2007. “Community (Dis)Organization and Racially Motivated Crime.” American Journal of Sociology 113(3):815–63.

Lyons, Christopher J. 2008. “Defending Turf.” Social Forces 87(1):357–385.

Lewis, P., D. Provine, M. Varsanyi, and S. Decker. 2013. Why Do (Some) City Police Departments Enforce Federal Immigration Law? Political, Demographic, and Organizational Influences on Local Choices. Journal of Public Administration Research and Theory, 23: 1–25.

Martinez, Lisa M. 2008. “The Individual and Contextual Determinants of Protest among Latinos.” Mobilization 13(2):189–204.

McCarthy, John D. and Mayer N. Zald. 1977. “Resource Mobilization and Social Movements.” American Journal of Sociology 82:1212–41.

McVeigh, Rory, Michael R. Welch, and Thoroddur Bjarnason. 2003. “Hate Crime Reporting as a Successful Social Movement Outcome.” American Sociological Review 68(6):843–67.

McVeigh, Rory. 2009. The Rise of the Ku Klux Klan. Minneapolis, MN: University of Minnesota Press.

McVeigh, Rory, Kraig Beyerlein, Burrell Vann, Jr., and Priyamvada Trivedi. 2014. “Educational Segregation, Tea Party Organizations and Battles over Distributive Justice.” American Sociological Review 79(4):630–52.

Medoff, Marshall H. 1999. “A Theoretical and Positive Analysis of Hate and Hate Crimes,” American Journal of Economics and Sociology, 58(4):959–73.

Michaud, Nicholas D. 2010. “From 287 (G) To SB1070: The Decline Of The Federal Immigration Partnership And The Rise Of State-Level Immigration Enforcement.” Arizona Law Review 52(4):1083–1133.

MIT Election Data and Science Lab, 2018, “County Presidential Election Returns 2000–2016”, https://doi.org/10.7910/DVN/VOQCHQ, Harvard Dataverse, V6, UNF:6:ZZe1xuZ5H2l4NUiSRcRf8Q== [fileUNF]

National Police Research Platform. 2016. “2016 National Survey of Law Enforcement Officers.” Pew Research Center.

Olzak, Susan. 1990. “The Political Context of Competition.” Social Forces 69:395–421.

Olzak, Susan. 1992. The Dynamics of Ethnic Competition and Conflict. Palo Alto, Calif.: Stanford University Press.

Pedroza, Juan. 2019. “Where Immigration Enforcement Agreements Stalled: The Location Of Local 287(G) Program Applications And Inquiries (2005–2012).”

Pinderhughes, Howard. 1993. “The Anatomy of Racially Motivated Violence in New York City.” Social Problems 40: 478–92.

Quillian, Lincoln. 1995. “Prejudice as a Response to Perceived Group Threat.” American Sociological Review 60(4):586–611.

Quillian, Lincoln. 1996. “Group Threat and Regional Change in Attitudes toward African-Americans.” The American Journal of Sociology 102(3):816–60.

Ramakrishnan, K. and T. Wong. 2010. Partisanship, Not Spanish: Explaining Municipal Ordinances Affecting Undocumented Immigrants. In Taking Local Control: Immigration Policy Activism in U.S. Cities and States, ed. M. Varsanyu, 73–96. Stanford, CA: Stanford University Press.

Rodríguez, Cristina, Muzaffar Chishti, Randy Capps, and Laura St. John. 2010. A Program in Flux: New Priorities and Implementation Challenges for 287(g). Washington, DC: Migration Policy Institute.

Rugh, Jacob S. and Matthew Hall. 2016. “Deporting the American Dream: Immigration Enforcement and Latino Foreclosures.” Sociological Science 3:1053–1076.

Sattin-Bajaj, Carolyn, and Allison Roda. 2018. “Opportunity Hoarding in School Choice Contexts: The Role of Policy Design in Promoting Middle-Class Parents’ Exclusionary Behaviors.” Educational Policy 34(7):992–1035.

Shapira, Harel. 2013. Waiting For José: The Minutemen’s Pursuit Of America. Princeton: Princeton University Press.

Semyonov, Moshe, Rebecca Raijman, Anat Yom Tov, and Peter Schmidt. 2004. “Population Size, Perceived Threat, and Exclusion.” Social Science Research 33(4):681–701.

South, Scott J. and Steven F. Messner. 1986. “Structural Determinants of Intergroup Association.” American Journal of Sociology 91(6): 1409–30.

Southern Poverty Law Center. 2011. ‘“Nativist Extremist” groups 2010’, The Southern Poverty Law Center Intelligence Report, 141.

Skocpol, Theda and Vanessa Williamson. 2013. The Tea Party and the Remaking of Republican Conservatism. Oxford, England: Oxford University Press.

Stacey, Michele, Kristen Carbone-Lopez, and Richard Rosenfeld. 2011. “Demographic Change and Ethnically Motivated Crime.” Journal of Contemporary Criminal Justice 27(3):278–298.

Stewart, Julie, Michele Enciso Bendall, and Charlie V. Morgan. 2015. “Jobs, Flags, and Laws: How Interests, Culture, and Values Explain Recruitment into the Utah Minuteman Project.” Sociological Perspectives 58:627–48.

Taxin, Amy, “Immigrants with Old Deportation Orders Arrested at Check-Ins,” Associated Press, June 18, 2017, www.apnews.com.

Thompson, Daniel. 2019. “How Partisan Is Local Law Enforcement? Evidence from Sheriff Cooperation with Immigration Authorities” American Political Science Review forthcoming.

Tilly, Charles. 1978. From Mobilization to Revolution. Boston, MA: Addison-Wesley Publishing.

Tolnay, Stewart E., and E. M. Beck. 1995. “A Festival of Violence: An Analysis of Southern Lynchings, 1882–1930” History: Reviews of New Books 24(3):107–108.

Tolnay, Stewart, E. M. Beck, and James L. Massey. 1989. “The Power Threat Hypothesis and Black Lynching” Social Forces 67:634–41.

U.S. Bureau of the Census. 1990. 1990–1999 Decennial Census of the Population: Profile of General Demographic Characteristics. Washington, DC: U.S. Department of Commerce.

U.S. Bureau of the Census. 2000. 2000–2009 Decennial Census of the Population: Profile of General Demographic Characteristics. Washington, DC: U.S. Department of Commerce.

U.S. Bureau of the Census. 2000. American Community Survey. Washington, DC: U.S. Department of Commerce.

U.S. Department of Homeland Security. 2019. “Delegation of Immigration Authority Section 287(g) Immigration and Nationality Act; Participating Entities; Chart.” U.S. Immigration and Customs Enforcement.

Van Dyke, Nella and Sarah Soule. 2002. “Structural Social Change and the Mobilizing Effect of Threat.” Social Problems 49(4):497–520.

Vickers, T. (2019). Borders, Migration and Class in an Age of Crisis: Producing Immigrants and Workers. Bristol University Press.

Wadhia, Shoba Sivaprasad. 2019. Banned: Immigration Enforcement In The Time Of Trump. New York: New York University Press.

Walker, Kyle, and Helga Leitner. 2011. “The Variegated Landscape of Local Immigration Policies in the United States.” Urban Geography 32(2):156–78.

Ward, Matthew. 2013. “Mobilising ‘Minutemen’: Predicting Public Support for Anti-Immigration Activism in the United States.” Sociological Research Online 18:4: 195–212.

Ward, Matthew. 2017. “Opportunity, Resources, and Threat: Explaining Local Nativist Organizing in the United States.” Sociological Perspectives 60:3: 459–478.

Wong, Tom K. 2012. “287(g) and the Politics of Interior Immigration Control in the United States: Explaining Local Cooperation with Federal Immigration Authorities.” Journal of Ethnic and Migration Studies 3:5: 737–756.

Mario Marset Ehrle
A Thesis Submitted to the Honors College of The University of Southern Mississippi in Partial Fulfillment of Honors Requirements
December 2020