The effects of immigration enforcement on traffic stops: Changing driver or police behavior?

• Published date: 01 August 2023
• Author: Britte van Tiem
• Date: 01 August 2023
• DOI: http://doi.org/10.1111/1745-9133.12625

DOI: ./- .

ORIGINAL ARTICLE

The effects of immigration enforcement on

traffic stops: Changing driver or police

behavior?

Britte van Tiem

Department of Criminology, University of

Pennsylvania, Philadelphia, Pennsylvania

Correspondence

Britte van Tiem, Department of

Criminology, University of Pennsylvania,

 McNeil Bldg,  Locust Walk,

Philadelphia, PA.

Email: bvantiem@sas.upenn.edu

Abstract

Research Summary: This research asks whether jail-

based immigration enforcement leads to the profiling

of Hispanics by municipal police. I leverage a natural

experiment to examine the effects of (g) jail partner-

ships on traffic stops and arrests by municipal police

in North Carolina in the late s. I find that stops

of Hispanic drivers fell in the wake of (g) agree-

ments, and show that this fall was driven by changes in

Hispanic road use. While I cannot unambiguously dis-

entangle police and driver behavior, I find no evidence

that municipal police officers increased pretextual stops

and arrests of Hispanic drivers.

Policy Implications: While there is existing evidence

of racial profiling under (g) agreements by sher-

iff’s deputies, this behavior does not appear to have

extended to nonsignatory municipal police agencies.

Despite this, the signing of (g) agreements appears

to have prompted Hispanic outmigration and changes

in driving behavior, adding to a growing body of evi-

dence on the ripple effects of local-federal immigration

enforcement partnerships in Hispanic communities.

This serves as a reminder that changes in demographic

composition of the immigrant population may medi-

ate the effects of immigration enforcement efforts on

other outcomes.

Criminology & Public Policy. ;:–. ©  American Society of Criminology. 457wileyonlinelibrary.com/journal/capp

458  TI EM

KEYWORDS

immigration enforcement, jails, police, profiling, race, traffic

(g) jail partnerships between County Sheriff’s Offices (CSOs) and Immigration and Cus-

toms Enforcement (ICE) became a prominent and controversial feature of the immigration

enforcement landscape in the late s. Under these partnerships, sheriff’s correctional deputies

conducted immigration screening and enforced select aspects of federal immigration law when

arrestees were booked into jail. In counties where (g) jail partnerships were in place, law

enforcement officers on patrol indirectly became part of the immigration enforcement landscape

because decisions to stop and arrest individuals in jurisdictions with active (g)programs largely

determined who would be drawn into the federal deportation pipeline (Motomura,). This led

many to worry that (g) partnerships would incentivize street-level law enforcement officers to

stop and arrest potentially undocumented individuals for pretextual reasons, so that they could

be delivered to the doorstep of the county jail for processing (Capps et al., ; Coleman, ;

Gill et al., ; Provine et al., ; Rodriguez et al., ; Stuesse & Coleman, ;Varsanyi

et al., ). Traffic enforcement quickly came to be understood as the prime stage for the profil-

ing of Hispanic drivers, and observers linked (g) jail agreements to the sharp increase in ICE

removals for traffic offenses in the mid-s (Capps et al., ; Moinester, ). In two cases,

Justice Department Investigations concluded that the CSO had indeed unlawfullytarg etedLatino

drivers (Perez, ,).Concerns about racial profiling however extended to other law enforce-

ment agencies as well, because sheriffs’ correctional deputies would check the immigration status

of all people checked into the local jail, regardless of whether a sheriff’s deputy, police officer, or

state trooper had made the initial arrest. Whether the signing of (g) jail agreementsaffected the

stop and arrest behavior of officers in nonsignatory law enforcement agencies however remains

an open empirical question (Coleman, ; Moinester, ; Treyger et al., ).

In this paper, I investigatethe effects of (g) agreements on traffic stops and arrests by munic-

ipal police in North Carolina. Using traffic stop data for -, I leverage the timing of

(g) agreements and cross-group differences in exposure to identify effects in an event-study

framework. Because the effects of (g) agreements are unlikely to be contained to the counties

that signed them, and because other immigration-related policy changes affected counties across

North Carolina in the period under study,I rely on placebo tests rather than on cross-county com-

parisons to distinguish results in (g) counties from statewide trends. I show that municipal

traffic stops of Hispanic drivers fell after (g) agreements were signed. Using population data

and data on traffic crashes, I show that this fall in Hispanic traffic stops is largely explained by a

change in road use by Hispanic drivers. Notably, arrest rates of Hispanic drivers remained rela-

tively constant, and there is no evidence to suggest that a change in the risk distribution of stops

masks an increase in pretextual arrests. Absent much better data on road use, I cannot unam-

biguously disentangle officer from driver behavior. Overall, however, results do not support the

idea that (g) jail enforcement partnerships resulted in increased pretextual stopsand arrests of

Hispanic drivers by municipal police.

This paper proceeds as follows. Section discusses (g) jail enforcement partnerships and

the profiling concerns they generated. In Section , I review how extant theoretical and empirical

work informs how we might expect (g) jail enforcement partnerships to affect traffic stops and

outcomes, and discuss implications for identification. In Section , I consider the implementation

 TI EM 459

of (g) in North Carolina, and place it within the context of a wider immigration enforcement

shock affecting North Carolina during the period under study. Section describes the data and

identification strategy.Sections and describe results for stop and arrest outcomes, respectively,

and Section discusses those results.

1287(g) AND PROFILING CONCERNS

In , Congress enacted Section (g) of the Immigration and Nationality Act as part of the

Illegal Immigration Reform and Immigrant Responsibility Act. Section (g) enabled federal

actors to delegate select immigration enforcement powers to state and local officials, providing

the basis for what came to be known as (g) partnerships. In the mid-s, congressional

appropriations for the (g) program increased rapidly as federal, state, and local authorities

looked for ways to combat the perceived threats posed by illegal immigration.In  alone,

(g) partnerships identified , individuals for removal (OIG, ). This paper focuses on

one type of (g) partnership, known as the Jail Enforcement Model (JEM) of (g), which

enabled sheriff’s correctional deputies to conduct postarrest immigration screening and enforce

select aspects of federal immigration law.After attending a -week ICE training course at a Fed-

eral Law Enforcement TrainingCenter, newly designated Jail Enforcement Officers (JEOs) could

screen the immigration status of arrestees booked into the county jail, issue detainers, and pre-

pare documents for removal proceedings. These new powers allowed Sheriffs to respond directly

to public safety and immigration concerns in the communities that elected them, while ICE saw

its enforcement capacity expanded.

Critics of the program were quick to flag that a large share of the individuals processed in jail

under (g) were first arrested by police officers working for nonsignatory agencies. While jail-

based (g) partnerships formally changed nothing about police officers’ powers to stop or arrest,

many worried that (g) agreements would encourageoff icers with anti-immigrantviews to stop

and arrest additional people they suspected to be undocumented. Such fears reflected a keen

appreciation of the fact that the discretion to arrest was “the discretion that matters” once (g)

agreements were in place (Motomura, ). Concerns about pretextual stops and arrests were

particularly pronounced in counties where the Sheriff appeared to set all arrested undocumented

migrants on a path to removal, including individuals arrested for minor criminal offenses and

traffic violations (Capps et al., ). Protracted concerns about deportations pursuant to minor

(traffic) violations and racial profiling associated with the program contributed to a dwindling of

active (g) agreements following the turn of the decade (Capps et al., ;GAO,; Kandel,

;Pham,).

Two prior studies haveassessed the impact of immigration enforcement partnerships on police

behavior using quasi-experimental methods.Treyger e t al. () study the impact of the activa-

tion of Secure Communities, another jail-based immigration enforcement partnership, on arrest

rates. Unable to distinguish arrestees of Hispanic ethnicity,the authors look at patterns in arrests

for minor crimes, and at arrests of White suspects relative to Black suspects, finding null results in

both pursuits. Coon () looks at the impact of gon Hispanic arrest rates in Frederick County,

Maryland. He compares arrests in the Frederick County Sheriff’s Office (FSCO), which signed a

(g) agreement, and the Frederick PoliceDepartment (FPD), which did not.He finds evidence

that the FCSO’s arrests of Hispanics increased relative to those of the FPD. Notably, he also finds

that arrests of Hispanics fell in both jurisdictions, which the author suggests is likely to reflect

efforts by Hispanic individuals to avoid contact with law enforcement. Such law enforcement