Immigration Enforcement Climate in Georgia and the Mental Health of Latinx Immigrants

Abstract

Introduction.

Immigrant-related social policies and immigration enforcement contribute to a sociopolitical environment that affects immigrants’ health. This exploratory study in six metro-Atlanta counties examined associations among immigrants’ perceived vulnerability to harmful immigrant-related social policies and county-level 287(g) agreements (which facilitate cooperation between local law enforcement and federal immigration authorities), county-level immigration enforcement levels (arrests, community arrests, detainers, and removals), and immigrants’ mental health.

Methods.

Using data from a 2020 study among Latinx parents who were undocumented or members of mixed-status families (N = 140), we merged data on individuals’ perceived policy vulnerability and depressive and anxiety symptoms with county 287(g) status and immigration enforcement levels.

Results.

Perceived policy vulnerability was not associated with county-level 287(g) status or immigration enforcement levels. Greater policy vulnerability and Immigration and Customs Enforcement arrests were associated with higher depressive and anxiety symptoms, but 287(g) status was associated with lower depressive and anxiety symptoms.

Conclusion.

Perceived policy vulnerability, 287(g) status, and immigration enforcement levels do not always align and can have different associations with mental health.

Federal, state, and local-level immigrant-related policies, hereafter immigrant policies, make up a complex, multilevel policy environment that is an important structural driver of health for immigrants.^1–4 Immigrant policies are policies that determine the rights, protections, benefits, and services (e.g., employment, social safety-net programs) made available to immigrants based on the citizenship or immigration status granted to them under federal law.⁵ Immigrant policies in specific domains (e.g., health care, immigration enforcement) may work in varied ways to affect health. For example, immigration enforcement policies such as Section 287(g) agreements and the Secure Communities Program facilitate the engagement of state and local law enforcement in federal immigration enforcement activities (referred to here as federal-subfederal immigration enforcement policies). Section 287(g) agreements enable local law enforcement to enforce federal immigration law including investigation, apprehension, and detention.⁶ The Secure Communities Program allows for data-sharing between local law enforcement and the U.S. Immigration and Customs Enforcement (ICE) to identify undocumented immigrants in local custody and to request, through ICE detainers, that local law enforcement hold noncitizens until they are taken into ICE custody for removal (deportation).⁷ These immigration enforcement policies can threaten the mental health of immigrants by increasing exposure to racialized policing, increasing fear of law enforcement, and threatening family unity through deportation.^8–11 The absence of these federal-subfederal immigration enforcement policies do not always correspond to lower immigration enforcement rates (i.e., arrests, detainers issued, removals) within a locale, in part, because of other state and local immigrant policies and practices that target immigrants. For example, state policies that deny driver’s licenses to undocumented immigrants, coupled with targeted traffic stops near immigrant communities, facilitate the identification, arrest, and removal of some immigrants.^12–14

The interplay between federal-subfederal immigration enforcement policies and immigrant policies in other domains is apparent in metropolitan Atlanta in Georgia. Georgia has one of the most restrictive state-level immigrant policy climates in the U.S., with policies that require the use of E-Verify for employment verification and restrict undocumented immigrants from eligibility for state-funded Medicaid and driver’s licenses.^1,8,15 Simultaneously, variation exists in the county-level adoption of immigration enforcement policies. For example, Fulton County (city of Atlanta) holds sanctuary policies that limit cooperation between local law enforcement and ICE.¹⁶ In contrast, nearby Gwinnett County held a 287(g) agreement until January 2021 and received national attention for the frequency of non-criminal arrests under 287(g).^17–20 Other counties have conflicting policy and enforcement levels. Dekalb County, which has never held a 287(g) agreement and holds sanctuary policies that limit cooperation with ICE in jails,¹⁶ had the one of the highest number of ICE community arrests per population of undocumented immigrants among all U.S. counties from 2017 to 2018.^20,21

Some existing studies of immigration enforcement focus on single components of immigration enforcement, such as the presence of enforcement policies,^9,22 enforcement rates,²³ or perceived policy climate.²⁴ Rarely are all three of these measures simultaneously collected over different policy environments to examine their associations with one another or their effects on health. Thus, we used data from a community-based participatory research (CBPR) study on immigrant policy stressors, COVID-19 stressors, and the mental health of Latinx parents to examine if perceived vulnerability to harmful immigrant policies (hereafter policy vulnerability) aligned with one county-level immigration enforcement policy and four measures of immigration enforcement levels. Further, we examined associations between these three factors—perceived policy vulnerability, 287(g) policy, and immigration enforcement rates—and participants’ depressive and anxiety symptoms. We hypothesized that policy vulnerability would be higher among individuals living in counties with active 287(g) agreements and where immigration enforcement levels were higher. We also hypothesized that policy vulnerability, greater immigration enforcement, and the presence of 287(g) agreements would be each be associated with poorer mental health.

Methods

Study and sample.

This analysis draws data from a broader mixed-methods CBPR study conducted in 2020 to examine COVID-19-related stressors, Georgia immigrant policy environment, and the mental health of Latinx parents who were undocumented or part of mixed-status families (i.e., families with members of different citizenship or immigration statuses) living in the metropolitan Atlanta area in Georgia. The study was conducted in partnership with Ser Familia, a long-standing, Latinx-led organization that provides mental health services and psychoeducational programs to Latinx families. Participants were recruited from Ser Familia clients at the time of their visit to one of six metro-Atlanta sites. Adults were eligible to participate if they were 18 years old or older, a parent or caregiver, Spanish-speaking, and undocumented or a member of a mixed-status family. Recruitment and survey administration took place between August and October 2020. Surveys were administered orally, in Spanish, by bicultural and bilingual Ser Familia staff, and answers were recorded in REDCap or on paper surveys. Paper surveys were double-entered into REDCap and verified by two independent research team members. Of the 145 parents who completed the survey, 140 with information on county of residence were included in this analysis.

Data.

For this exploratory analysis, we merged survey data from the above-described mixed-methods study with data on county-level 287(g) agreement status and immigration enforcement levels from external sources. The surveys captured information on participants’ perceived policy vulnerability and their self-reported depressive and anxiety symptoms. We assessed policy vulnerability using the Perceived Immigration Policy Effects Scale (PIPES).²⁵ Participants were asked to reflect on their “experiences and feelings about the immigration policy context” and were asked how often (1=never through 5=always) they experienced 24 items across four domains of discrimination, social exclusion, threat to family, and children’s vulnerability. This analysis examined only the social exclusion (five items; α =.76) (e.g., “Do you fear being deported or detained?” and threat to family (three items; α = .81) (e.g., “Do you worry about family separation due to deportation?”) subscales to focus on vulnerabilities hypothesized to be especially tied to local immigration enforcement policies. We summed the values of item responses to create separate, continuous social exclusion and threat to family scores, with higher scores indicating greater policy vulnerability. Depressive symptoms were assessed using the Center for Epidemiologic Studies Depression 10-item scale (CES-D 10).^26,27 Items were summed and examined as a continuous total score (max possible = 30). Anxiety symptoms were assessed using the General Anxiety Disorder-7 scale (GAD-7).^28,29 Items were summed and examined as a continuous total score (max possible = 21).

Survey measures were merged with data on county-level 287(g) agreement status and immigration enforcement levels based on participants’ county of residence. Active 287(g) status was assessed as of October 2020, prior to the November elections (which yielded changes in national, state, and county governments). County-level immigration enforcement levels were captured using TRAC Immigration’s data on the 2017 number of ICE arrests and community arrests, detainers at county facilities, and removals under the Secure Communities Program.^30–32 Each of the four measures of immigration enforcement were calculated as the one-year rate per 100,000 county population. Data for 2017 were used because it was the most recent year with a full year of data for all four enforcement measures (e.g., arrest data are only available through May 2018).

Analysis.

We conducted exploratory bivariate analyses using simple generalized linear regression models to examine associations between PIPES social exclusion and threats to family subscales and the 287(g) status and immigration enforcement levels in their county of residence. We also examined whether a county’s 287(g) status or each measure of immigration enforcement was associated with depressive symptoms and general anxiety disorder symptoms. Statistical significance was assessed at an alpha <.1 because of the small sample size. Analyses were completed using SAS version 9.4.³³

Results

Demographic characteristics, policy vulnerability, and mental health scores of the sample (n = 140) are presented in Table 1. In brief, 84% of participants were from Mexico or Central America and, on average, participants had lived 14 years in the U.S. Since living in the U.S., 48% had family or friends who had been detained and 19% had family or friends who had been deported by immigration enforcement agents. The average PIPES social exclusion and threat to family subscale scores were 13.1 (standard deviation (SD) = 4.5) and 11.1 (SD = 3.7), respectively. These scores were similar, though slightly higher, than in the original PIPES study.²⁵ The average CES-D 10 score was 11.9 (SD = 6.0), corresponding to high symptomatology,²⁶ and the average GAD-7 score was 5.9 (4.9), corresponding to mild symptomatology.²⁸

Table 1.

CHARACTERISTICS OF THE ANALYTIC SAMPLE (N = 140)

	N	Percent	Mean (SD)	Range
Female sex	124	88
Age (in years)			40 (8)	20, 70
Marital status
Currently married	86	61
Separated, divorced or widowed	39	28
Never married	15	11
Annual household income <$30,000	121	86
Country or region of origin
Mexico	77	55
Central America	41	29
South America	19	14
Othera	3	2
Length of U.S. residence (in years)			14 (8)	1, 38
Have family or friends detainedb	67	48
Have family or friends deportedb	26	19
Perceived policy vulnerability
PIPES: Social exclusionc			13 (5)	5, 25
PIPES: Threat to familyc			11 (4)	3, 15
Depressive symptoms (CES-Dd 10 score)			12 (6)	0, 27
Anxiety symptoms (GAD-7 score)e			6 (5)	0, 21

Notes

^aOther countries of origin include the U.S. (including Puerto Rico) or other countries outside of the Americas.

^bHave family or friends detained or deported since the respondent had been living in the U.S.

^cPIPES = Perceived Immigration Policy Effects Scale

^dCES-D = Center for Epidemiological Studies Depression

^eGAD-7 = General Anxiety Disorder-7

Six counties were represented in the study, and most participants (81%) resided in Cobb and Gwinnett counties. Half of the counties held 287(g) agreements and of these, all had 2017 total ICE arrest, detainer, and removal rates above Georgia averages for each metric (Table 2). Among the non-287(g) counties, only Dekalb County had ICE arrest and removal rates above the Georgia average. Dekalb County’s ICE community arrests (173.2 per 100,000) were approximately 28 times greater than the community arrest rates in the next highest county in the study (Gwinnett: 6.2 per 100,000) and were among the highest levels of community ICE arrests per population of all U.S. counties in 2017.^20,21 This warranted that we explore models with and without the 13 Dekalb County residents in our sample.

Table 2.

COUNTY 287(G) STATUS AND IMMIGRATION ENFORCEMENT LEVELS PER 100,000 COUNTY POPULATION FOR SIX GEORGIA COUNTIES

County	287(g) Statusa	ICE Arrestsb	ICE Community Arrestsb	Detainersb	Removalsb
Dekalb	No	301c	173c	54	34c
Forsyth	No	9	0	37	36c
Fulton	No	8	3	44	17
Cobb	Yes	82c	5	99c	42c
Gwinnett	Yes	183c	6	216c	51c
Hall	Yes	189c	0	246c	87c

Notes

^aCounty 287(g) status as of October 2020.

^bImmigration enforcement measures reported for 2017.

^cAnnual rate above Georgia: ICE Arrests (73.7 per 100,000), ICE Community arrests (16.1 per 100,000), Detainers (68.1 per 100,000), Removals (25.5 per 100,000).

Mean scores for each PIPES subscale were not statistically different in 287(g) counties compared with non-287(g) counties (social exclusion: mean [SD] = 12.9 [4.3] vs. 14.1 [5.6]; threat to family: 11.2 [3.5] vs. 10.3 [4.4]). Exclusion of Dekalb County residents did not meaningfully change the results. No association was observed between the two PIPES subscales and any of the four immigration enforcement measures (e.g., arrests and social exclusion: B(95% confidence interval (CI)): .0002 (−.01, .01); arrests and threat to family: B(95% CI): −.003 (−.01, .01); regression results not shown in tables).

Higher PIPES subscale scores were associated with higher depressive and anxiety symptoms (Table 3). Total and community ICE arrests were associated with very small increases in depressive symptoms and anxiety symptoms. No association was observed with ICE detainers or Secure Community removals. In contrast, living in a 287(g) county was associated with lower mean depressive symptoms. When Dekalb County residents were excluded, patterns of lower mental health symptoms in 287(g) counties remained, but the associations were no longer statistically significant (B_{depressive symptoms}(95% CI): −1.7 (−6.0, 2.6); B_{anxiety symptoms}(95% CI): −2.1 (−5.4, 1.2)).

Table 3.

ASSOCIATIONS BETWEEN PERCEIVED POLICY VULNERABILITY, COUNTY 287(G) STATUS, IMMIGRATION ENFORCEMENT LEVELS, AND DEPRESSIVE AND ANXIETY SYMPTOMS AMONG LATINX PARENTS (N = 140)^a

	Depressive Symptomsb		Anxiety Symptomsb
	B	95% Confidence Interval	B	95% Confidence Interval
PIPES: Social exclusionc	.46	.26, .65*	.34	.17, .50*
PIPES: Threat to familyc	.19	−.07, .46	.26	05, 48*
County 287(g) Statusd	−2.90	−.12, −5.69*	−3.66	−1.37, −5.94*
ICE arrestse	.01	−.001, .03*	.01	.003, .03*
ICE community arrestse	.02	.001, .04*	.03	.01, .04*
Detainerse	−.002	−.02, .01	−.003	−.02, .01
Removalse	−.01	−.12, .11	−.02	−.11, .18

Notes

^aAll models fit independently.

^bDepressive symptoms assessed using the CES-D 10; Anxiety symptoms assessed using the GAD-7.

^cPIPES = Perceived Immigration Policy Effects Scale

^dCounty 287(g) status as of October 2020.

^eEnforcement measures reported in 2017 number per 100,000 county population.

^*p <.1

Discussion

We conducted an exploratory study of Latinx parents who were undocumented or members of mixed-status families living in six counties in the metro-Atlanta area to examine associations between perceived immigrant policy vulnerability and county-level immigration enforcement policy, immigration enforcement levels, and mental health. We did not observe an association between policy vulnerability and either county’s 287(g) status or immigration enforcement levels, suggesting that these components of the immigration enforcement climate might have independent effects on health. We observed that higher policy vulnerability and ICE community arrests were associated with modestly higher depressive and anxiety symptoms. In contrast, having an active 287(g) agreement was associated with lower depressive and anxiety symptoms, which was contrary to our hypothesis. In analyses excluding a non-287(g) county with especially high ICE arrest rates, this association was no longer statistically significant.

Our findings did not support our hypothesis that perceived policy vulnerabilities related to social exclusion and threat to family (measured using PIPES) would be associated with county immigration enforcement levels and policies. Although we selected PIPES subscales with items that assessed vulnerabilities related to mobility in the community, detention, or deportation, it may be that in answering the full PIPES measure, which had no specific prompt for types of policies (PIPES prompt: “immigration policy context”), participants may have also reflected on the entire immigrant policy environment. Thus, the PIPES measure could have captured a range of restrictive immigrant policies, including state and federal policies and policies in other domains (e.g., social safety net). Examples of these policies emerged from qualitative findings from frontline workers in the full mixed-methods study. Workers emphasized the impact of the exclusion of undocumented and/or mixed status families from the Coronavirus Aid, Relief, and Economic Security (CARES) Act of 2020, which provided financial assistance to address economic instability caused by the COVID-19 pandemic³⁴ and Georgia’s pre-existing restrictive driver’s license laws.³⁵ Additionally, they expressed concern about the new Public Charge Rule that went into effect in February 2020, which assessed an individual’s likelihood of receiving of public benefits in the future based on their past and present financial status.^35,36

Our findings that higher policy vulnerabilities were associated with higher depressive and anxiety symptoms was consistent with our hypothesis. Findings for immigration enforcement levels were mixed. U.S. Immigration and Customs Enforcement community arrest rates were associated with very small increases in depressive and anxiety symptoms, but other measures of total ICE arrests, detainers, and removals were not associated with mental health. This association with community arrests may be because community arrests take place at home, work, or during travel²¹ and are thus more visible to the community than other immigration enforcement activities that take place once a person has been taken into custody by law enforcement (e.g., issue of detainers).

Lastly, our finding that mental health scores of participants living in non-287(g) counties were lower than those of participants living in 287(g) counties was contrary to our hypothesis. Other research has found that 287(g) agreements and the Secure Communities Program were associated with poorer self-rated health and higher psychological distress among immigrants in some or all noncitizen households.⁹ Our findings may be due to the high arrest rates in Dekalb County, a non-287(g) county. In analyses without Dekalb County, the unexpected direction of the association remained, but was no longer statistically significant. Because of our small sample of Dekalb County residents, we cannot make a definitive conclusion that arrest rates are confounding the 287(g) and mental health association.

Strengths and limitations.

Our findings are exploratory and are based on a small sample of Latinx individuals living in six counties in Georgia in a study that was not designed to examine associations between immigration enforcement levels or policies. Although we compare policy vulnerability with county 287(g) status and immigration enforcement levels, we are limited in our ability to examine this correlation across more counties within and outside of the state. We were unable to assess exposure to 2020 levels of immigration enforcement because full-year data were not available for all four measures in 2020 or any more proximal year than 2017. Comparisons between 2017, 2018, and partial-year 2019 TRAC data on detainers and removals show comparable or increasing levels of enforcement, and partial-year ICE arrest data (through May 2018) indicate that arrest levels would have been similar to or greater than levels in 2017. We also acknowledge that interruptions in enforcement may have occurred in 2020 because of the COVID-19 pandemic, but the previous three years of enforcement are likely to have had lasting effects on immigrant communities.

Despite these limitations, our study is strengthened by the collection of data on perceived policy vulnerability and the examination of associations with local immigration enforcement policy and immigrant enforcement levels. This is especially valuable because perceived policy vulnerability is not widely assessed in studies among immigrants. Our findings are also limited by the likely cross-county exposures that immigrants face in their day-to-day lives. For example, some participants resided in non-287(g) counties but sought services from offices located in 287(g) counties and could have faced the threats of policies and immigration enforcement in those counties. In studies with larger sample sizes, opportunities may exist to compare the effects of policy and enforcement levels on individuals who do and do not regularly cross county lines.

Lastly, our study is strengthened by its timing. This study took place before the 2020 election that yielded changes to immigrant policies, including the ending of 287(g) agreements in two counties in this study. Additionally, this study took place during the first year of the COVID-19 pandemic in the U.S. and still found associations between mental health and different components of the immigration enforcement climate. This underscores how influential immigration enforcement policies and activities are on undocumented and mixed-status families, even amid other unprecedented threats to health and livelihood.

In this exploratory study, we observed that perceived vulnerability to harmful immigrant policies, county-level immigration enforcement policies, and immigration enforcement levels do not always align with one another. Further, we observed that each of these components had different effects on mental health. Future studies of immigration enforcement policies and health should incorporate components of perceived vulnerability, enforcement policies, and enforcement levels because they may differentially contribute to health. Policymakers, especially those in locales that have ended or let 287(g) agreements lapse, should be aware that this is one step toward promoting health equity and that it is necessary to address other restrictive, non-enforcement immigrant policies and practices.