A Framework for Understanding Subfederal Enforcement of Immigration Laws
Huyen Pham, A Framework for Understanding Subfederal Enforcement of Immigration Laws
A Framework for Understanding Subfederal Enforcement of Immigration Laws
When thirty-two-year-old Kate Steinle was randomly shot during
broad daylight on San Francisco’s Pier 14, the initial public reaction was
one of shock. When the shooter was determined to be an unauthorized
immigrant who had been deported five times, and had been recently released
by the San Francisco Sheriff’s Department despite an extensive criminal
record, that shock turned to outrage.1 Politicians across the country railed
against San Francisco’s “sanctuary policy,” which limited cooperation
between the city’s law enforcement agencies (LEAs) and Immigration and
Customs Enforcement (ICE), resulting in the immigrant’s release. Senator
David Vitter (R-LA) introduced legislation—ultimately unsuccessful—to
strip federal funding from jurisdictions that do not fully cooperate with
federal immigration enforcement.2 Some states have threatened noncooperative
LEAs with a similar loss of state funding.3
* Professor of Law, Texas A&M University School of Law; A.B. Harvard College; J.D.
Harvard Law School. I would like to acknowledge the helpful comments I received when I
presented a version of this article at the Conference of Asian Pacific American Law Faculty at the
University of California, Davis. I am particularly grateful for the thoughtful feedback I received
from Hiroshi Motomura. I would also like to thank Elizabeth Anderson, Michael Doyle,
Kimberley Gorny, Christina Gutierrez, Jake Jensen, and Jessica Lesnau for their excellent research
1. Evan Sernoffsky & Jaxon Van Derbeken, Pier Shooting Suspect had been Released from
S.F. Jail, SF GATE (July
, 7:11 AM),
2. Mike DeBonis, Senate Democrats Block “Sanctuary Cities” Bill, WASH. POST (Oct. 20,
3. Letter from Greg Abbott, Tex. Governor, to Lupe Valdez, Dall. Cty. Sheriff (Oct. 26,
2015.pdf (demanding that Sheriff Valdez reverse her department’s policy of refusing to
automatically honor all ICE detainer requests or risk losing state funds).
MODELING COOPERATION: A FRAMEWORK
Media outlets, in their coverage of this controversy, reported that
hundreds of subfederal jurisdictions also have sanctuary policies. In doing so,
these media accounts lumped together San Francisco’s very strict
noncooperation policy with other jurisdictions that have much more moderate
policies.4 This lumping together has important implications because it glosses
over the differences in LEA practices and frames the debate as all or
nothing—either LEAs cooperate wholesale with federal immigration
enforcement, or they become “sanctuaries” like San Francisco, avoiding any type
of immigration cooperation whatsoever. In fact, however, LEAs have taken
a variety of paths, with some implementing fairly nuanced cooperation
Drawing on our empirical database of subfederal immigration laws,5
this article seeks to move the debate forward by describing, in detail, the six
different immigration enforcement models that LEAs have implemented. At
the heart of all these models is the question of how much cooperation, if
any, LEAs should offer to federal immigration authorities. Because the
federal government ultimately controls the removal process, and because the
Supreme Court has struck down subfederal attempts to enforce immigration
law outside of the federal scheme,6 cooperation is the key concept for
evaluating LEA responses.
The models, listed in order of most to least cooperation, are: (1) the
unilateral model, where LEAs initiate immigration enforcement, feeding
into federal programs; (2) the 287(g) model, where LEAs are delegated
authority by ICE to act as immigration agents; (3) the jailhouse interview
model, where LEAs voluntarily provide ICE with names of foreign-born
detainees and allow ICE to interview those detainees in LEA facilities; (4)
the detainer model, where LEAs cooperate with federal requests to detain
immigrants and to notify when those immigrants are released from local
custody; (5) the notification-only model, where LEAs will notify ICE when
specific immigrants are released from local custody, but will not detain
immigrants on ICE’s behalf; and (6) the no-cooperation model, where LEAs
do not detain or notify.
As explained in Section II, these different models developed against a
context of changing historical, legal, and technological forces. Some of the
models developed in response to these changes, particularly to changing
federal policy.7 We also see instances where subfederal immigration
initiatives have been engines of change themselves, influencing enforcement
policies at the federal level. The federal government alternately discouraged
and then encouraged LEA immigration cooperation at different points in
time.8 After the 9/11 attacks, when the federal government encouraged LEA
cooperation in earnest,9 the form of that cooperation was more
labor-intensive. Some LEAs signed 287(g) agreements, where their officers are trained
and authorized to act as immigration agents (Model 2); other LEAs agreed
to participate in the Criminal Alien Program (CAP), through which they
allowed ICE agents to interview foreign-born individuals being held in their
jails (Model 3). A handful of jurisdictions have tried to act more
unilaterally, imposing enforcement requirements on their LEAs independent of any
federal programs (Model 1).
With advances in technology and the inclusion of immigration
information in federal crime databases that all LEAs check when they make
arrests, federal-LEA cooperation has become more focused on information
sharing. Under this structure, LEAs inform ICE when they have custody of
someone who allegedly violated immigration laws; until recently, ICE
would also request LEAs to detain immigrants for ICE pickup. Now, after
legal challenges, the Obama administration’s stated policy was to decrease
the use of federal detainer requests,10 but questions remain about whether
that policy will be widely and consistently implemented.11 Models 4, 5 and
6 represent different jurisdictions’ responses to this information-centered
In discussing the varied LEA responses, the normative question
naturally arises: Which model of immigration enforcement should an LEA
embrace? If an LEA with no current immigration enforcement policy were to
decide on a model, which model should it choose? Or, if an LEA wants to
reconsider its current enforcement model, what factors should it consider in
making its decision? The answers to these questions depend on the interests
of individual LEAs—interests that may vary from LEA to LEA. The second
contribution of this article then is to raise important questions that LEAs
should consider in deciding which model is best for them.
1. Who is your community that you are protecting?
2. How important is community communication/trust to your
3. Will immigration cooperation advance your mission of protecting
To help LEAs think through these questions, this article gathers and
synthesizes the relevant empirical information and related research.
Section II of this article describes the six models of LEA cooperation
in more detail, explaining the historical, legal, and technological context of
each model’s development. Section II develops a framework to assist LEAs
in choosing among the cooperation models. In asking the questions above,
this section also analyzes the relevant empirical research and information.
The article ends with some concluding thoughts for LEAs as they consider
MODELS OF COOPERATION
What does it mean to say that an LEA cooperates with federal
immigration enforcement? “Cooperation” is “an act or instance of working or
acting together for a common purpose or benefit . . . [or] more or less active
assistance from a person [or] organization.”12 To develop the different
models of LEA cooperation, I focus on the assistance prong of the definition
and consider what forms of assistance an LEA can offer to federal
immigration authorities. Because LEAs may be motivated by goals that diverge
from federal immigration goals, I do not consider whether LEAs and
federal immigration authorities are acting “for a common purpose or benefit.”
The voluntary nature of cooperation between any particular LEA and
ICE should also be emphasized. Because of Tenth Amendment constraints,
any assistance that ICE receives from LEAs regarding immigration
enforcement has to be voluntary on the part of the LEAs.13 Any federal mandate to
LEAs to assist in immigration enforcement would raise concerns that the
federal government is commandeering state resources in violation of the
12. Cooperation, DICTIONARY.COM, http://www.dictionary.com/browse/cooperation?s=t (last
visited July 18, 2016).
13. But see Kalhan, supra note 7.
14. See Printz v. United States, 521 U.S. 898, 935 (1997) (“The Federal Government may
neither issue directives requiring the States to address particular problems, nor command the
States’ officers, or those of their political subdivisions, to administer or enforce a federal
In developing the different models of cooperation, I drew upon the
Immigration Climate Index (ICI), an empirical database that I codeveloped
to track subfederal immigration laws.15 The ICI tracks laws enacted by
states, cities, and counties from 2005 to 2015 that regulate immigrants
within their jurisdictions. These subfederal laws regulate in many different
substantive areas (e.g., LEA cooperation in enforcing federal immigration
laws; immigrant access to employment, drivers’ licenses, and social welfare
programs; English-only laws; and, laws related to voting and immigrant
legal services).16 For present purposes, I focus on the LEA-related laws
within the ICI to ascertain trends that are helpful in developing the models
For the ICI, we chose 2005 as the start date for our data collection
because that is the year when subfederal governments enacted immigration
regulations in noticeable volumes. The National Conference of State
Legislatures, a clearinghouse for state level legislation, only started compiling
reports on immigration-related laws in 2005. Before that year, state laws
related to immigration were rare and largely limited to the distribution of
state social service benefits.17 Our own tracking of city and county level
laws confirms a similar timeline for the growth of local immigration laws.
What was the impetus for the growth of subfederal immigration
regulations in 2005? Though subfederal governments have always played an
important role in the integration of immigrants within their jurisdiction, the
phenomenon of direct immigration regulation by cities, counties, and states
can be traced to the 9/11 attacks. In June 2002, nine months after the
attacks, Attorney General John Ashcroft invited states to enforce civil
immigration laws as part of “our narrow anti-terrorism mission.”18 This
invitation generated much controversy; previously, the federal government
had forcibly maintained that state enforcement of federal immigration laws
was limited to criminal laws (e.g., human trafficking laws).19 Using their
“inherent authority” as sovereigns, Ashcroft argued that states could also
enforce civil immigration laws (e.g., laws prohibiting visa overstays).20
Encouraged by this federal invitation, subfederal governments have
leapt into the world of immigration enforcement, taking substantially
different paths. The following section describes the six main models that LEAs
have taken. While it is possible for an LEA to embrace more than one
model at a time (e.g., implementing Model 1’s Unilateral Enforcement
while also allowing ICE to do jailhouse checks through Model 3), it’s easier
to understand these LEA actions as separate models.
Model 1: Unilateral Enforcement
The unilateral model refers to LEAs in jurisdictions that impose
affirmative enforcement responsibilities on the LEAs, independent from any
federal enforcement program. The most well-known example of unilateral
enforcement is Arizona. As part of its infamous SB 1070 legislation,
Arizona requires its law enforcement officers to make a “reasonable attempt”
to determine the immigration status of anyone they encounter during any
“lawful stop, detention or arrest” where “reasonable suspicion exists that
the person is an alien and is unlawfully present in the United States.”21 This
provision, known colloquially as the “show me your papers” law, survived
a constitutional challenge in 2010 when the Supreme Court held that the
provision was not preempted by federal law.22 Alabama, Georgia, South
Carolina, and Utah have enacted similar laws.23
It may seem counterintuitive to characterize unilateral initiatives as
involving immigration cooperation when these laws are enacted
independently of any federal enforcement program. However, because the
enforcement impact of these laws is to increase referrals into the removal
system controlled by the federal government, the laws can be thought of as
a form of immigration cooperation, a form of assistance from the LEA to
the federal government. Moreover, because the enforcement under this
model operates independently of federal control (i.e., the federal
government does not create the guidelines or the rules governing enforcement),
this model creates the potential for the most active form of cooperation
among the six models.
Using Arizona’s law as an example, we see that LEA obligations
under Model 1 are potentially broader. First, the LEA verification of a
person’s immigration status (and any subsequent notification and detention)
can occur at a much earlier stage—during a lawful stop or detention if the
21. The law provides in relevant part: “For any lawful stop, detention or arrest made by a law
enforcement official . . . in the enforcement of any other law or ordinance of a county, city or town
or this state where reasonable suspicion exists that the person is an alien and is unlawfully present
in the United States, a reasonable attempt shall be made, when practicable, to determine the
immigration status of the person, except if the determination may hinder or obstruct an
investigation.” ARIZ. REV. STAT. ANN. § 11-1051 (2016).
22. Arizona v. United States, 132 S. Ct. 2492, 2509–10 (2012).
23. ALA. CODE § 32-6-9(b) (1975); GA. CODE ANN. § 17-5-100(b) (2015); S.C. CODE ANN.
§ 17-13-170 (2013); UTAH CODE ANN. § 76-9-1003(2)(b) (2016).
officer has “reasonable suspicion” to believe that the person does not have
legal immigration status. Second, in contrast with Models 3, 4, and 5,
immigration verification by LEAs can happen even if the person is never
arrested.24 For example, if an officer encounters someone who is panhandling
and the officer reasonably suspects that the person is unlawfully present,
then the officer must verify that person’s immigration status, even if the
officer doesn’t arrest him or her.
Because these unilateral laws operate independently from federal
enforcement programs, the federal government can’t place restrictions on
Arizona’s LEAs (e.g., to limit immigration verification to investigations where
the person is suspected of committing serious or violent crimes). Unlike
with Model 2 (the 287(g) model), the federal government can’t choose
which LEAs participate, so the potential number of participants (and the
resulting increased enforcement) could be quite large. The bottom line,
then, is that cooperation under Model 1 requires active, earlier LEA
participation. This model of cooperation could also potentially result in more
notifications, detentions, and, eventually, removals, as more immigrants are
funneled into the immigration enforcement system.25
The long-term horizon for the unilateral model is uncertain. The
legality of the model was upheld by the Supreme Court, and the states that
rushed to enact some version of it currently still have those laws on their
books.26 However, enacting this unilateral model has also extracted costs.
Arizona received a torrential amount of negative publicity for its “show me
your papers” law. Critics charged that the law would result in racial
profiling of Latinos, either because LEAs would be inadequately trained in the
complexities of immigration law and would resort to enforcement against
those who look like immigrants, or because they would abuse their new
powers to carry out anti-Latino agendas.27 Indeed, the Supreme Court, in
upholding the law, left the door open for a subsequent challenge if there
24. Arizona law further requires LEAs to verify the immigration status of all individuals who
are arrested, regardless of whether reasonable suspicion exists. ARIZ. REV. STAT. ANN. § 11-1051
25. However, the enforcement experience of Tucson, Arizona post–SB 1070 suggests that
this provision hasn’t increased dramatically the number of LEA notifications to ICE. See Perla
Trevizo, SB 1070 Leads to Immigration Checks, Few Deportations, ARIZ. DAILY STAR (Dec. 12,
http://tucson.com/news/sb-leads-to-immigration-checks-few-deportations/article_a2610e46-3efd-570a-9f5a-caa3b95c849c.html (finding that from June 201
4 to October 2015
Tucson police ran over 26,000 immigration checks, but only fifty-one checks were done to
comply with the “show me your papers” provision; the other checks were done to comply with another
state requirement that all arrestees be checked for legal immigration status). However, the number
of show-me checks could increase, depending on each LEA’s policies regarding the parameters of
26. ALA. CODE § 32-6-9(b); GA. CODE ANN. § 17-5-100(b); S.C. CODE ANN. § 17-13-170;
UTAH CODE ANN. § 76-9-1003(2)(b).
27. American Civil Liberties Union, ACLU and Civil Rights Groups File Legal Challenge to
Arizona Racial Profiling Law (May 17, 2010),
was evidence that the law was being applied in an unconstitutional way.28
In protest, national organizations and even some states vowed to boycott
Arizona; some reports estimate that the state lost $140 million in tourism
spending in the immediate aftermath of enacting the law.29 Critics also
charge that LEA involvement with immigration enforcement under this
model diverts important enforcement resources away from the investigation
of local crimes.30
Perhaps most detrimental to the long-term prospects of this model is
the federal government’s aggressive expansion of Secure Communities and
similar programs. As described in more detail below, these programs rely
on information sharing between ICE and LEAs, information sharing that
has been devastatingly successful in placing large numbers of immigrants
into removal proceedings.31 A subfederal government that wants to increase
immigration enforcement in its jurisdiction may find that participation in a
federal enforcement program is a lower-cost and lower-profile way to
obtain the additional enforcement it seeks.
Model 2 deputizes LEAs to carry out immigration enforcement tasks
pursuant to a signed agreement between the LEA and ICE. Those
agreements—called 287(g) agreements—are authorized by a provision of the
Immigration and Nationality Act, which provides in relevant part:
[T]he Attorney General may enter into a written agreement with a
State, or any political subdivision of a State, pursuant to which an
officer or employee of the State or subdivision, who is determined
by the Attorney General to be qualified to perform a function of
an immigration officer in relation to the investigation,
apprehension, or detention of aliens in the United States (including the
transportation of such aliens across State lines to detention
centers), may carry out such function at the expense of the State or
political subdivision and to the extent consistent with State and
Model 2 represents a very active form of LEA cooperation, albeit one
that is more firmly under the federal government’s control. Under this
model, local law enforcement officers are trained by ICE in the basics of
immigration law; upon successful completion of the four-week training,
officers are deputized to carry out immigration law enforcement functions, as
described in the Memorandum of Agreement (MOA) signed by the LEA
and ICE. Examples of these functions include interviewing noncitizens to
ascertain their status, entering information into ICE’s case management
system, issuing detainers, making immigration charges, making
recommendations for voluntary departure, making recommendations for bond and
detention, and transferring noncitizens into ICE custody.33 While training
and carrying out their immigration duties, law enforcement officials
continue to be paid by their LEA departments.
There are three types of 287(g) agreements: (1) the task force model,
which grants broader authority to LEAs to conduct immigration
enforcement tasks during their regular law enforcement activities in the field; (2) a
jail model, which assigns ICE-trained LEA officers to prisons and jails; and
(3) a hybrid model, which combines features of both the task force and jail
models.34 The 287(g) agreements are posted on ICE’s website. While the
agreements expire on a certain date, they can also be revoked by either side.
The agreements specify the scope of immigration powers that each LEA is
allowed to exercise.35
In 2009, at its height, the 287(g) program had seventy-seven LEA
participants.36 Currently, there are only thirty-two LEAs participating in the
program.37 The list of current LEA participants includes the Las Vegas
Metropolitan Police Department, the Jacksonville Sheriff’s Office (Florida),
and the Massachusetts Department of Corrections.38 Explanations for its
declining popularity echo the concerns raised about the unilateral
enforcement model. First, signing a 287(g) agreement is an active and very visible
symbol of immigration enforcement cooperation—one that risks alienating
immigrant communities within LEA jurisdictions, which could undermine
an LEA’s ability to protect public safety. Participating in the 287(g)
program also carries financial costs, as LEAs have to pay for their officers’
time both during training and during the actual implementation of their
immigration duties; these costs have led to criticism that the program diverts
crucial LEA resources from the investigation of local crimes.39
From ICE’s perspective, complaints about racial profiling and other
bad behavior by LEA participants has led ICE to revoke some agreements
and make more specific and stringent criteria for the remaining
agreements.40 Another reason for the declining popularity of the 287(g) program
is technology that makes it possible for LEAs to cooperate in less-costly,
less-visible ways. That technology is discussed in more detail in Section
Model 3: Jailhouse Checks
Under this model, LEAs give ICE agents access to their jails to
interview detainees that ICE may be interested in removing. This federal
program, the Criminal Alien Program (CAP), is one of the least understood and
least studied of the federal immigration programs that have been very
successful in removing immigrants. Described as, “the primary mechanism
through which ICE removes people from the U.S. interior,” CAP was
responsible for removing 508,000 people during 2010–2013 alone.41
The focus of CAP is to screen detainees held in prisons and jails to
ascertain immigration status and begin removal proceedings for those
without lawful immigration status. Participating LEAs cooperate by providing
ICE with lists of detainees held in their jails and allowing ICE access (either
physical access or video access) to interview detainees of interest.42 Often,
ICE will ask LEAs to detain the immigrants beyond their release date, so
ICE can pick them up for removal proceedings.43
The jailhouse check model requires less cooperation from LEAs than
the 287(g) model, where LEAs are deputized to carry out immigration
enforcement functions themselves. Yet, participation in CAP necessarily
entangles LEAs with federal immigration enforcement; under the usual CAP
procedures, ICE agents are physically present and active within LEA jails.
This entanglement raises the very real possibility of creating distrust
UNIVERSITY OF ST. THOMAS LAW JOURNAL
tween LEAs and immigrant communities within their jurisdictions, which
could negatively interfere with their ability to protect public safety.44
CAP is also resource intensive for ICE because it requires them to
have agents available to interview detainees of interest (either through
video or phone interviews, or face-to-face interviews that require agent
travel). CAP operates in all federal and state prisons, where its job is to
determine which prisoners should be placed in removal proceedings, either
at the conclusion of their sentences or before their sentences are fully
served. However, perhaps because of the resource-intensive nature of this
cooperation, the program only operates in approximately 300 local jails.45
Model 4: Immigration Detainers
Model 4 (where LEAs honor ICE detention requests and notification
requests) and Model 5 (where LEAs only honor ICE notification requests)
are best understood against the context of technological advances that made
the models relevant. Described in more detail below, those technological
advances have made immigration information more readily available to
LEAs, thus making LEA cooperation in immigration enforcement less
costly and less visible.
The first technological advance was the addition of immigration
records into the National Crime Information Center (NCIC) database, the
federal database that law enforcement officers around the country check to
identify individuals with criminal histories and outstanding warrants.
Starting in 1996, criminal immigration records (e.g., the offense of re-entering
the U.S. after a previous removal) were added. In 2002, the federal
government added civil immigration records as well (e.g., the offense of not
leaving after a final removal order is issued). Shortly after Attorney General
Ashcroft issued his enforcement invitation to LEAs, the addition of civil
immigration records significantly expanded the scope of potential LEA
cooperation. So, now when an LEA checks the NCIC (as they typically do
during a traffic stop or when they book someone after an arrest), they are
also given information about the person’s immigration records. If there is
an immigration “hit,” the LEA is instructed by the system database to
contact ICE to allow ICE the opportunity to file a detainer.46
The second technological advance is inextricably linked with the
advent of the Secure Communities program in 2008. The goal of Secure
Communities was to “identify and remove aliens who pose a threat to public
safety.”47 To do this, the federal government emphasized automated
biometric identification and the sharing of information among the FBI, DHS,
and LEAs. One crucial component to this strategy was to automatically
compare the fingerprint information that LEAs input when they book
someone with the information in DHS’s Automated Biometric Identification
IDENT is used by DHS for a wide range of immigration control
functions and contains records of those who have interacted with DHS, other
agencies, and even other governments. These records include information
from encounters directly related to immigration enforcement (unauthorized
immigrants apprehended at sea or at the border and immigrants suspected of
violating immigration laws), but also from encounters related to more
benign immigration events (applicants for U.S. visas, noncitizens traveling to
and from the United States, and even U.S. citizens applying to participate in
Global Entry and other similar programs).48
If there is a fingerprint match with IDENT records, ICE does further
research with multiple federal databases to determine whether the flagged
individual is removable from the United States. A response with the
individual’s immigration status and criminal history (if any) is shared with the
regional ICE office, the FBI, and the arresting LEA. Until recently, if ICE
wanted to place the arrested immigrant in removal proceedings, it would
often ask the arresting LEA to detain the immigrant beyond his or her
ordinary release date, to give ICE additional time to pick up the immigrant.49
LEAs that honor these detainer requests fall squarely within Model 4.
These technological advances, with their emphasis on information
sharing, make LEA cooperation much less costly and less visible. Instead of
enacting its own enforcement legislation and formulating its own
enforcement procedures (Model 1), spending extensive resources to train its
officers to enforce immigration laws directly (Model 2), or inviting ICE into
its jails (with the resulting entanglement) (Model 3), an LEA now can
cooperate with ICE in ways that are relatively cost-efficient and not highly
visible to the public. These advantages make Model 4 (and Model 5) more
attractive for both LEAs and ICE.
However, as explained in more detail with Model 5, the LEA practice
of honoring ICE detainers has been the subject of legal challenge. As a
result, one of the major changes the Obama administration instituted when
UNIVERSITY OF ST. THOMAS LAW JOURNAL
it replaced Secure Communities with the Priority Enforcement Program
(PEP) was to limit the issuance of detainer requests to, “special
Nonetheless, there are several reasons why Model 4 remains relevant.
First, PEP, with its relevant guidelines and restrictions, is a program created
and administered by the Obama administration. Based on his stated
positions on the campaign trail, President-elect Trump seems unlikely to
continue this limitation on ICE detainer requests.51 Furthermore, even during
President Obama’s administration, civil rights and immigrant advocates
were skeptical that the detainer limitation would be implemented in a
consistent and timely manner.52 Finally, even if PEP’s changes are enacted in
good faith, the exception allowing for detainer requests to be issued under
“special circumstances” (which is not defined by ICE) still presents an
opportunity for LEAs to accept or deny these requests.
Given the uncertainty about the legality of detainer holds and ongoing
policy concerns about LEA immigration cooperation, the various LEAs that
follow Model 4 have been selective about what kinds of detainer holds and
notification requests they will honor. For example, some LEAs only detain
or notify when the individual has been booked for a violent or other serious
Model 5: Notifications Only
LEAs operating under Model 5 cooperate by notifying ICE when
noncitizens of interest to ICE will be released or transferred from LEA
custody. Importantly, under this model, LEAs do not honor ICE requests to
detain nonimmigrants beyond the date that they would otherwise be
released under LEA procedures.
There seem to be several reasons that LEAs are rejecting detainer
requests. Spurred by protests from immigrant and civil rights advocates, some
LEAs refuse to honor detainer requests because of concerns that the
requests were being issued indiscriminately, reaching noncitizens without
criminal records. Indeed, audits of data from Secure Communities showed
50. See Johnson, supra note 10.
51. President-elect Trump has stated that “ICE officers should be required to place detainers
on every illegal alien they encounter in jails and prisons, since these aliens not only violated
immigration laws, but then went on to engage in activities that led to their arrest by police.”
Donald Trump on Immigration, REPUBLICANVIEWS.ORG (Sept. 1, 2015), http://www.republican
52. Tanfani & Linthicum, supra note 11.
53. See, e.g., Letter from Martin O’Malley, Governor, State of Md., to Hon. Gregg
Hershberger, Sec’y, Dept. of Pub. Safety and Corr. Serv. (April 18, 2014), https://immigrantjustice.org/
sites/immigrantjustice.org/files/MD_Baltimore_2014_04_18.pdf (directing the department to only
honor ICE detainer requests involving individuals convicted or charged with a felony offense;
convicted of three or more misdemeanor convictions; convicted of or charged with an offense
involving a significant threat to public safety; or with an outstanding order of removal).
MODELING COOPERATION: A FRAMEWORK
that, contrary to the stated goals of the program, it has resulted in mass
removals of immigrants with no criminal record.54 This particular concern
has been raised from the inception of the Secure Communities program.
A more recent concern for LEAs is the potential legal and financial
liability that comes with honoring detention requests. A number of federal
court decisions have held that LEAs can be held legally liable for the
resulting Fourth Amendment violations when they hold individuals pursuant to
an ICE detainer request.55 Because civil immigration detentions are seizures
subject to Fourth Amendment constraints, immigration detainer holds must
be based on probable cause regarding removability. Thus, when LEAs
detain someone pursuant to an ICE request that lacks probable cause
regarding the individual’s removability, that detention is unconstitutional and
subjects the LEA to legal and financial liability.
Cognizant of these court decisions, a wave of LEAs declared that they
would no longer honor ICE detainer requests. Different advocacy groups
estimate that over 200 jurisdictions currently have policies expressing
opposition to, or limited cooperation with, ICE detainer requests.56
The adverse court decisions and the decisions by many LEAs to refuse
to honor ICE detainer requests, incentivized the Obama administration to
restrict its use of these requests. In announcing the creation of PEP to
replace Secure Communities, DHS Secretary Jeh Johnson made several
important changes to the administration’s use of detainer requests. First,
instead of routinely issuing detainer requests, ICE would ask for
notification instead. Second, if “special circumstances” do justify the issuance of a
detainer request, then ICE must also “specify that the person is subject to a
final order of removal, or there is other sufficient probable cause to find that
the person is a removable alien.”57 The federal government’s stated purpose
in adding this information is to try to address the Fourth Amendment issues
litigated in the detainer cases discussed above.58
54. See Secure Communities, NAT’L IMMIGR. F. (July 11, 2011), https://immigrationforum
.org/blog/secure-communities-2/ (Secure Communities data shows that “ICE has arrested more
individuals with no criminal history than any other category, a total of 61,234 arrests since the
55. See, e.g., Morales v. Chadbourne, 996 F. Supp. 2d 19, 39 (D.R.I. 2014) (holding
detention of an immigrant based on an immigration detainer “for purposes of mere investigation is not
permitted”); Miranda-Olivares v. Clackamas Cty., 2014 WL 1414305, at *11 (D. Ore. Apr. 11,
2014) (holding that the County violated the plaintiff’s Fourth Amendment rights by detaining her
on an immigration detainer request beyond the time she would have otherwise been released).
56. National Immigrant Justice Center, Challenge Unjust Immigration Detainers (last visited
July 19, 2016) (listing three states and 229 city and county policies limiting cooperation with
57. See Johnson, supra note 10.
UNIVERSITY OF ST. THOMAS LAW JOURNAL
Model 6: No Cooperation
This model is perhaps the easiest to understand: LEAs under this
model do not cooperate with federal immigration enforcement in any form,
including the forms discussed in previous models (unilateral enforcement,
deputization, jailhouse checks, detainers, or notification). Their reasons for
noncooperation echo LEA concerns that have been raised about
immigration cooperation, generally. First and foremost, LEAs express unease that
entanglement with immigration enforcement will destroy the trust
developed with immigrant communities and undermine their ability to protect
public safety.59 Second, noncooperative LEAs worry that immigration
enforcement will violate the civil rights of immigrants and groups associated
with immigrants (through racial profiling or unconstitutional detentions).60
Finally, LEAs are concerned about the diversion of scarce local resources to
pay for immigration enforcement (e.g., unreimbursed detention expenses,
time LEA officers spend in immigration enforcement duties, and possible
legal liability associated with certain forms of enforcement).61
San Francisco is, perhaps, the most famous jurisdiction popularly
associated with a noncooperation policy. The policy changed several times
before reaching its current version. In 1989, the city and county of San
Francisco declared itself, through an ordinance, to be a “City and County of
Refuge.”62 That ordinance prohibits city and county employees from
assisting in the enforcement of federal immigration laws and from gathering or
disseminating the immigration status of anyone living in the city or county,
unless required by federal or state law.63 The city amended its code again in
2013 to prohibit its LEAs from honoring detainer requests.64 The San
Francisco Sheriff’s Department, whose primary function is to operate the county
jails, enacted policies that correspondingly limited its deputies’ authority to
cooperate with federal immigration agencies.65
If we stopped the story in 2013, San Francisco would be the classic
example of a noncooperation jurisdiction. Its policies, both citywide and
LEA-focused, were specific in their refusal to cooperate with federal
immigration enforcement. After Kate Steinle was shot in July 2015 by Francisco
Sanchez, an unauthorized immigrant who had been released by the San
Francisco Sheriff’s Department, the city faced immense pressure to change
its noncooperation policy.66 Sanchez, who had been deported five times
previously and had seven felony convictions (five of which were
immigration-related), had been turned over to the sheriff’s department by ICE for an
outstanding drug warrant. According to media reports, ICE had requested
notification if the sheriff’s department released Sanchez.67 When the San
Francisco district attorney declined to prosecute Sanchez for the decade-old
marijuana possession charge, the sheriff’s office released him without
notifying ICE, in compliance with its noncooperation policy. After his release,
Sanchez shot Steinle in an incident that he described as an accident.68
Shortly after the shooting, the San Francisco Board of Supervisors (its
city council) passed a nonbinding resolution to reaffirm its status as a
sanctuary city and rejected cooperation with the federal government’s Priority
Enforcement Program.69 However, when newly elected Sheriff Vicki
Hennessey pushed for more discretion to cooperate with ICE and threatened to
unilaterally enact policies enabling the department to do so, the Board of
Supervisors approved compromise amendments.70
Now, San Francisco’s laws allow the sheriff’s department to honor
detainer requests if the individual has been convicted of a violent felony
within the past seven years, and a magistrate judge has determined that
there is probable cause to believe that the individual is guilty of a violent
felony, based on current charges.71 Similarly, the sheriff’s department may
honor an ICE notification request if the individual has been convicted of a
violent felony within the past seven years (or has been convicted of a
serious felony in the past five years, or has been convicted of three separate
enumerated felonies—other than domestic violence—within the past five
years), and a magistrate judge has determined that there is probable cause to
believe that the individual is guilty of an enumerated felony, based on
current charges.72 Immigrant advocates strongly pushed for the added
involvement of a magistrate judge in both processes, believing that the magistrate’s
UNIVERSITY OF ST. THOMAS LAW JOURNAL
independent determination would help protect the due process rights of
immigrants who are the potential targets of immigration cooperation.73
That San Francisco, with its history of liberal politics and immigrant
advocacy, allows for limited immigration cooperation raises questions
about whether Model 6 actually exists. As we move into a policy discussion
about the type of cooperation model that LEAs should have, it’s important
to consider a strict no-cooperation model among the available options for
several reasons. First, different stakeholders in the debate about
immigration cooperation believe that a strict no-cooperation model is being
implemented in the United States. Based on their rhetoric, some politicians seem
to believe this model exists, and based on news articles, some media outlets
seem to believe this, too.74 Second, even if they do not believe that a
noncooperation policy currently exists, some stakeholders (including immigrant
advocacy groups) believe that such a policy should exist.75 Considering this
model among our available options enables us to weigh the costs and
benefits of adopting a noncooperation policy.
FRAMEWORK FOR DECISION-MAKING
Given the variety in the models of cooperation, which model should an
LEA choose? This analysis starts with the baseline assumption that LEAs
should make their own decisions, based on their particular demographics,
political constraints, and values. This Section seeks to provide a framework
for that decision-making by suggesting questions that LEAs should ask
when choosing among models. The three basic questions are:
1. Who is your community that you are protecting?
2. How important is community communication and trust to your
3. Will immigration cooperation advance your mission of protecting
Defining the Protected Community
The threshold task for an LEA that is choosing among cooperation
models is to define the community that it seeks to protect. LEAs commonly
define the communities under their jurisdiction by geographical boundaries
MODELING COOPERATION: A FRAMEWORK
(individuals living within set county, city, or state lines). Yet, to make a
wise decision about immigration cooperation, an LEA needs to know
detailed demographic information about the residents within its jurisdiction.
How many foreign-born residents live in the jurisdiction? How many
unauthorized immigrants live there? In the LEA’s jurisdiction, how large are the
populations of Asians, Latinos, and other groups commonly associated with
When an LEA is considering the impact of its cooperation policy, it’s
important to recognize that the impact will go well beyond the unauthorized
immigrants within its jurisdiction. Demographic data shows that many
immigrant families are mixed status—meaning some family members have
authorized immigration status (often children born in the United States who
are natural-born citizens) while other members have unauthorized status.77
When an immigrant is deported as a result of LEA immigration
cooperation, the deportation not only hurts the affected immigrant, but also his or
her family members, some of whom are likely to have lawful status.
Additionally, the federal databases used in immigration cooperation
schemes often have erroneous information, raising the stakes for affected
groups within the LEA’s jurisdiction. Studies by outside groups and by the
federal government itself have pointed out the patterns of errors in this
information.78 The implications of these errors are that people who have not
committed immigration offenses are being wrongfully questioned and
detained.79 Finally, the pernicious effects of racial profiling, a concern with all
immigration cooperation models, are not limited to immigrants, of course.
Latinos, Asians, and other individuals associated with immigrants can be
victims of racial profiling, even if they are U.S. citizens.
The larger the size of these affected communities, the more ripple
effects immigration cooperation will have within a jurisdiction. Accordingly,
an LEA that has large populations of immigrants and other
immigrant-asso76. Online data sources can provide some of this demographic information quickly and
accurately. See, e.g., U.S. Immigrant Population by County and State, MIGRATION POLICY INST., http://
(last visited July 17, 2016).
77. The PEW Research Center estimates that, in 2010, an estimated nine million people were
living in “mixed status” families. A Nation of Immigrants: A Portrait of the 40 Million, Including
11 Million Unauthorized, PEW RES. CTR. (Jan. 29, 2013), http://www.pewhispanic.org/2013/01/
78. Based on data from 2002 to 2004, the Migration Policy Institute reported that NCIC’s
rate of false positives was forty-two percent overall and as high as ninety percent for some
individual law enforcement agencies. “‘The incredibly high number of false positives in the database
means that police resources, which are always stretched thin, are being wasted on detaining
immigrants and non-immigrants alike who haven’t done anything wrong,’ said MPI President Demetrio
Papademetriou.” Press Release, Migration Policy Inst., MPI Report Shows Database Errors
Plague Federal Efforts to Induce Immigration Enforcement by Local Police; Report Offers First
Look at Police Activity Since September 11 (Dec. 8, 2005), http://www.migrationpolicy.org/news/
UNIVERSITY OF ST. THOMAS LAW JOURNAL
ciated communities should carefully consider the negative impact of
choosing an active model of immigration cooperation.
B. The Role of Community Trust
The second question an LEA needs to consider is the relative
importance of community trust and communication to its policing model. If the
affected populations are sizeable (see Question 1 above), then an LEA
needs to carefully consider whether communication and trust with those
populations is important to its policing model. As discussed in more detail
below, immigration cooperation may undermine the ability of LEAs to
protect their communities.
A 2012 study by researchers at the University of Illinois at Chicago
shows that immigrants who associate their LEAs with immigration
enforcement are less willing to report crimes or cooperate as witnesses, for fear of
exposing themselves or their family members.80 The survey interviewed
over 2000 Latinos living in different counties across the United States with
different immigration statuses (U.S. born, naturalized, and unauthorized).
The increased reluctance to cooperate with law enforcement was widely
shared by Latinos, regardless of immigration status, and was particularly
heightened in counties like Maricopa County, Arizona, where LEAs have
been actively involved in immigration enforcement.
More specific findings from the survey include:
Forty-four percent of respondents stated that they are less likely to
contact an LEA if they have been victims of a crime because they
are afraid that the LEA will use the contact to inquire about their
immigration status or the immigration status of people they know.
Forty-five percent reported that they are less likely to volunteer
information about crimes; similarly, forty-five percent said that they
are less likely to report a crime because they are afraid the LEA
will inquire about their immigration status or the immigration status
of people they know.
Seventy percent of unauthorized immigrants said that they are less
likely to contact LEAs if they are victims of a crime because of
increased LEA cooperation with immigration law enforcement.
Twenty-eight percent of U.S. born Latinos said that they are less
likely to report a crime because they are afraid the LEA will inquire
about their immigration status or the immigration status of people
80. NIK THEODORE, DEP’T OF URBAN PLANNING AND POLICY, UNIV. OF ILL. AT CHI.,
INSECURE COMMUNITIES: LATINO PERCEPTIONS OF POLICE INVOLVEMENT IN IMMIGRATION
7 (May 2013
81. Id. at 4–5.
Based on these results, the authors of the study concluded that LEA
involvement with immigration law enforcement has increased Latino fear
of, and isolation from, their local LEAs. This sense of isolation has many
Latinos feeling less secure in their homes and neighborhoods, as they are
reluctant to contact LEAs if they are victimized by criminals.82
If an LEA employs a community policing model, it may be reluctant to
become entangled with immigration enforcement. Community policing is a
model of law enforcement that depends on developing community
partnerships.83 Though LEAs differ in how they develop these partnerships, the
essential components consist of “develop[ing] positive relations with the
community, . . . involv[ing] the community in the quest for better crime
control and prevention, and . . . pool[ing] their resources with those of the
community to address the most urgent concerns of community members.”84
To do these things, it is essential that an LEA maintains trust and candid
communication with the residents living in its jurisdiction. As the
University of Illinois at Chicago survey demonstrates, that trust and
communication will not be forthcoming from many Latinos if an LEA becomes
involved with immigration enforcement.
At a minimum, the principles of community policing, together with the
conclusions of the University of Illinois at Chicago survey, should raise red
flags for LEAs contemplating the adoption of a more visible and active
cooperation model (Models 1–3).
C. Does Cooperation Advance Public Safety?
This question is at the heart of an LEA’s decision-making and
essentially requires an LEA to balance the potential costs and benefits of
immigration cooperation. The question assumes that an LEA’s primary mission
is to protect public safety and presents a framework for deciding whether
immigration cooperation advances public safety in its jurisdiction (and if so,
which model of cooperation).
The benefits of cooperation are increased immigration enforcement
and increased removals of immigrants. Our national experience with the
Secure Communities program clearly demonstrated that information sharing
between LEAs and ICE leads to increased deportations.85 If those who are
deported have committed dangerous or serious crimes, then public safety is
arguably enhanced when these persons are removed from the United States.
82. Id. at 17.
83. The second commonly identified requirement of community policing is problem solving:
“the process through which the specific concerns of communities are identified and through which
the most appropriate remedies to abate these problems are found.” BUREAU OF JUSTICE
ASSISTANCE, UNDERSTANDING COMMUNITY POLICING: A FRAMEWORK FOR ACTION 13 (Aug. 1994),
85. U.S. Immigration and Customs Enforcement, supra note 47.
UNIVERSITY OF ST. THOMAS LAW JOURNAL
The costs of cooperation include the reduced trust and communication with
immigrant communities (discussed earlier), fiscal costs to pay for
enforcement, and possible legal liability. Depending on the political currents within
an LEA’s jurisdiction, immigration cooperation may also bring political
costs or benefits.
To accurately weigh the costs and benefits of immigration cooperation,
LEAs should consider these questions:
What kinds of crime are your enforcement priorities?
Will immigration cooperation decrease those types of crimes?
Looking at the different LEA models, will certain types of
cooperation be more effective in decreasing crime than others?
In the LEA’s community, does cooperation bring political benefits
What are the fiscal and legal costs of cooperation?
With finite resources, LEAs have to prioritize the types of offenses
they will enforce. The specific crimes that LEAs are responsible for
enforcing depends, of course, on their jurisdictions’ laws, but for police
departments, sheriff’s offices, and state troopers, the list of crimes is likely to
include crimes against persons, crimes against property, traffic offenses,
and controlled substance violations.
Where do immigration offenses, and specifically, civil immigration
offenses, fit in on that list? Examples of civil immigration offenses that may
be enforced through immigration cooperation include overstaying a
temporary visa,86 or staying in the United States after a removal order has been
issued.87 With the possible exception of Model 2, where LEAs are
deputized to enforce immigration laws, LEAs are not directly enforcing civil
immigration laws through immigration. However, their assistance (through
granting ICE access to their jails, providing notifications of release, or
honoring detainer requests) in Models 1, 3, 4, and 5 result in additional
enforcement of civil immigration laws.
Some LEAs may place a high priority on assisting with civil
immigration law enforcement. In explaining their support for immigration
cooperation, some LEA leaders argue that all violations of law need to be punished
in order to preserve the rule of law.88 Other LEA officials have suggested
that enforcing the nation’s immigration laws, and thus, securing the
country’s borders, furthers public safety.89 If an LEA places a high priority on
MODELING COOPERATION: A FRAMEWORK
immigration law enforcement relative to the enforcement of other types of
laws, then it would favor Models 1–3 (and, perhaps, Model 4 as well) for its
active forms of cooperation.
Other LEAs may place a lower priority on immigration enforcement,
especially as compared with other law enforcement responsibilities that the
LEAs may already have. If an LEA regards immigration cooperation as
having little or no priority, then it could simply choose not to cooperate
with any immigration enforcement (Model 6). However, for reasons
explained below, a pure no-cooperation policy is difficult to maintain because
of intense political opposition to such a policy.
An LEA that doesn’t prioritize immigration law enforcement may find
it easier to embrace Model 5. Not only is Model 5 more politically
palatable, but it can also provide synergy between criminal law enforcement and
immigration cooperation. For example, if an LEA prioritizes laws
prohibiting violent or dangerous crimes, it can set up internal criteria so it only
honors ICE detainer or notification requests targeted at violent or dangerous
crimes. There are many different possible criteria: an LEA could choose to
honor ICE requests only for persons who have been charged with violent or
dangerous crimes, only for persons with previous convictions for violent or
dangerous crimes, only for persons with recent convictions, regardless of
the nature of the underlying crimes, or only when a neutral party (like a
magistrate judge) has determined that probable cause supports a criminal
By setting its own criteria, an LEA can ensure that its immigration
cooperation efforts also enhance its primary goal of enforcing laws
targeting more traditional crimes (like homicide or robbery).
Another important consideration for LEAs is to weigh the political
impact of adopting any specific model of cooperation. LEAs do not work in a
political vacuum. In most jurisdictions, sheriffs are elected and police chiefs
are appointed by mayors, who are themselves elected.91 Immigration
cooperation is a very politically volatile topic, and while LEAs should not be
guided solely by political concerns in choosing a model of cooperation,
LEAs should be sensitive to the political impact of their decisions. In some
jurisdictions, adopting an active and visible model of cooperation may bring
political benefits; in other jurisdictions, the same policy might bring
condemnation from civil rights and immigrant advocacy groups.
Similarly, given the rise of anti-immigrant sentiment, a strong
noncooperation policy can provoke threats to defund LEAs at the state or
federal level. Even San Francisco—a very pro-immigrant city, in a
pro90. San Francisco’s current policy requires involvement of a magistrate judge before its
LEAs can cooperate with immigration law enforcement. See S.F., CAL., ADMIN CODE ch. 12H
§§ 12H.2, 12I (2016).
91. JEFFREY J. NOBLE & GEOFFREY P. ALPERT, MANAGING ACCOUNTABILITY SYSTEMS FOR
POLICE CONDUCT: INTERNAL AFFAIRS AND EXTERNAL OVERSIGHT 176 (2009).
immigrant state—faced political pressure to allow exceptions in its
noncooperation policy.92 Vicki Hennessey, who was elected sheriff of San
Francisco shortly after the Kate Steinle shooting, had made a campaign
promise to modify the sheriff department’s no-cooperation policy. Once in
office, Hennessey claimed an electoral mandate and pushed the city’s Board
of Supervisors to change the city’s laws to allow her department to have
more discretion to contact ICE.93 The experience of San Francisco suggests
that a strict Model 6 no-cooperation policy is politically difficult to sustain.
Regardless of the model chosen, an LEA should expect strong political
reactions and be prepared to clearly communicate its reasoning for choosing
The last factor for LEAs to consider is the financial costs of
immigration cooperation. One cost—the threat of state or federal governments
cutting funding for noncooperation policies—has already been mentioned.94
Other costs are directly related to cooperation, as LEAs that cooperate with
immigration enforcement do so largely on their own dime. When LEA
officers assist in immigration cooperation, their salaries are paid for by the
LEA, not by the federal government.95 Moreover, when an LEA honors an
ICE detainer request, the expenses it incurs to detain the individual are not
reimbursed by ICE.96 Finally, in light of recent litigation, an LEA that
honors a detainer request may also be financially liable for violating detainees’
The decision by an LEA to define its role in immigration enforcement
is an important and complicated task. The complexity arises, in part, from
competing (and often contradictory) demands made by different
constituencies, changes in federal policies, and changes in technology that can reshape
enforcement tools. As it considers this decision, an LEA should recognize
that there is substantial variety in the cooperation models available to it.
This article defines these models based on current federal programs and
If new technologies in immigration enforcement are developed, or if
the federal government changes its cooperation programs, then the specific
details of these models will, of course, change. However, the policy
questions raised in this article should remain relevant. In this complicated area
of policy and politics, LEAs are well-served by knowing their own
communities and their own law enforcement goals. With this knowledge and
awareness, LEAs can be more proactive in defining their own cooperation
model, resulting in policies that reflect their enforcement priorities and are
more beneficial and consistent for the communities within their
* * *
4. Todd Beaman , Kate Steinle's Mother: Sanctuary Cities' Law Has “Gotten Totally Crazy” , NEWSMAX (July 21 , 2015 ), http://www.newsmax.com/Newsfront/Liz-Sullivan -KathrynSteinle- sanctuary-cities/ 2015 /07/21/id/658298/ (grouping San Francisco' s sanctuary law with other jurisdictions' laws that limit cooperation with federal immigration authorities).
5. Huyen Pham & Pham Hoang Van , Measuring the Climate for Immigrants: A State-byState Analysis, in STRANGE NEIGHBORS: THE ROLE OF STATES IN IMMIGRATION POLICY 21-39 (Carissa Byrne Hessick & Gabriel J. Chin eds., 2014 ); Huyen Pham & Pham Hoang Van , StateCreated Immigration Climates and Domestic Migrants , 38 HAW. L. REV. 181 ( 2016 ).
6. Arizona v. United States , 132 S. Ct . 2492 , 2509 - 10 ( 2012 ) (striking down three Arizona laws that made it a state crime to be unlawfully present or to work without authorization and that authorized Arizona LEAs to make warrantless arrests of people suspected to be removable).
7. See Anil Kalhan , Immigration Policing and Federalism Through the Lens of Technology, Surveillance, and Privacy , 74 OHIO ST. L .J. 1105 , 1108 - 09 ( 2013 ) (arguing that a technologybased shift toward automated immigration policing disrupts the prevailing federalism equilibrium by mandating state and local participation in immigration policing).
8. See infra note 20 (opining that local police may enforce civil but not criminal provisions of the Immigration & Nationality Act) .
9. John Ashcroft, Att'y Gen., Prepared Remarks on the National Security Entry-Exit Registration System (June 6 , 2002 ), https://www.justice.gov/archive/ag/speeches/2002/060502agpre paredremarks.htm.
10. See Memorandum from Jeh Charles Johnson, Secretary, U.S. Dep't of Homeland Sec., to Thomas S. Winkowski, Acting Director, U.S. Immigration & Customs Enforcement , et al. (Nov. 20 , 2014 ), https://www.dhs.gov/sites/default/files/publications/14_1120_ memo_prosecutorial_dis cretion.pdf (stating ICE will not seek immigration detainers except under “extraordinary circumstances”).
11. Joseph Tanfani & Kate Linthicum , New Federal Effort to Deport Criminal Immigrants Draws Local Skepticism, L.A. TIMES ( July 16 , 2016 , 6 :00 AM), http://www.latimes.com/nation/lana-immigrants-detainers-20150512-story.html.
15. Pham & Hoang Van , Measuring the Climate for Immigrants, supra note 5; Pham & Hoang Van , State-Created Immigration Climates and Domestic Migrants, supra note 5.
16. Pham & Hoang Van , Measuring the Climate for Immigrants, supra note 5 , at 17.
17. Most of these pre-2005 state laws were reacting to federal welfare reform, the Illegal Immigration Reform and Immigrant Responsibility Act, which prohibited the distribution of welfare benefits to most immigrants. E-mail from Ann Morse, Program Dir ., Immigrant Policy Project , Nat'l Conference of State Legislature, to author (Aug. 12 , 2009 , 11 : 47 AM ) (on file with author).
18. Ashcroft , supra note 9.
19. Memorandum from the Off . of Legal Counsel of the U.S. Dep't of Justice, on Assistance by State and Local Police in Apprehending Illegal Aliens to the U.S. Attorney for the S. Dist . of Cal. (Feb. 5 , 1996 ) (opining that local police may enforce civil but not criminal provisions of the Immigration & Nationality Act) .
20. Ashcroft , supra note 9.
28. Arizona v. United States , 132 S. Ct . 2492 , 2509 ( 2012 ).
29. Mark Berman , Arizona's Experience with Controversial Laws and Boycotts, WASH . POST (Feb. 26 , 2014 ), https://www.washingtonpost.com/news/post-nation/wp/2014/02/26/arizonas -ex perience-with-controversial-laws-and-boycotts/?utm_term= . d517ebdf88b6 .
30. ACLU, supra note 27.
31. See discussion infra Section II .D.
32. Immigration and Nationality Act § 287 (g), 8 U.S.C. § 1357 (g) ( 2012 ).
33. RANDY CAPPS ET AL., MIGRATION POLICY INSTITUTE, DELEGATION AND DIVERGENCE: A STUDY OF 287(G) STATE AND LOCAL IMMIGRATION ENFORCEMENT 12- 13 ( 2011 ).
34. Id . at 14-15.
35. Memorandum of Agreement from Todd Entrekin , Sheriff, Etowah Cty . Sheriff's Off. , to Thomas D. Homan, Exec. Assoc. Dir. Enf't and Removal Operations, U.S. Immigration and Customs Enf't, Dep't of Homeland Sec. (June 8, 2016 ), https://www.ice.gov/doclib/foia/memoran dumsofAgreementUnderstanding/r_287getowah. pdf. In response to community complaints that Phoenix Sheriff Joe Arpaio was engaging in racial profiling and otherwise violating civil rights laws, ICE in 2011 revoked the 287(g) agreement it had previously signed with the Maricopa County Sheriff's Office. Ray Stern, Feds Pull 287(g) Authority From Maricopa County Jails Because of Civil Rights Violations, PHOENIX NEW TIMES (Dec . 15, 2011 , 1 :00 PM), http://www .phoenixnewtimes.com/news/feds-pull-287 - g -authority-from-maricopa-county-jails-because-ofcivil-rights-violations-6631025.
36. Jennifer M. Chaco ´n, A Diversion of Attention? Immigration Courts and the Adjudication of Fourth and Fifth Amendment Rights, 59 DUKE L .J. 1563 , 1582 n. 88 ( 2010 ).
37. U.S. IMMIGRATION & CUSTOMS ENFORCEMENT , Fact Sheet : Delegation of Immigration Authority Section 287(g) Immigration and Nationality Act , https://www.ice. gov/287g (last viewed on Jul. 18 , 2016 ).
38. Id .
39. CAPPS ET AL., supra note 33 , at 28-29.
40. Id . at 4.
41. GUILLERMO CANTOR ET AL., AM. IMMIGRATION COUNCIL , ENFORCEMENT OVERDRIVE : A COMPREHENSIVE ASSESSMENT OF ICE'S CRIMINAL ALIEN PROGRAM 4-6 (Nov . 2015 ), https://www .americanimmigrationcouncil.org/sites/default/files/research/enforcement_overdrive _a_compre hensive_assessment_of_ices_criminal_alien_program_final .pdf.
42. ANDREA GUTTIN , IMMIGRATION POLICY CTR., THE CRIMINAL ALIEN PROGRAM: IMMIGRATION ENFORCEMENT IN TRAVIS COUNTY , TEXAS 5 ( 2010 ), https://www.americanimmigrationcoun cil.org/sites/default/files/research/Criminal_Alien_Program_ 021710 .pdf.
43. Id .
44. Mirela Iverac , New York Works Too Closely with ICE , Critics Say , WNYC NEWS (June 8 , 2011 ), http://www.wnyc.org/story/140853-ice-rikers/.
45. GUTTIN, supra note 42, at 4-6.
46. See U.S. DEP'T. OF HOMELAND SEC., PRIVACY IMPACT ASSESSMENT FOR ALIEN CRIMINAL RESPONSE INFORMATION MANAGEMENT SYSTEM ( ACRIME ) 4 - 5 ( 2010 ), https://www.dhs.gov/ xlibrary/assets/privacy/privacy_pia_ice_acrime.pdf; Nat'l Crime Info. Ctr. , NCIC 2000 Operating Manual: Immigration Violators File , at 195 ( 2000 ).
47. U.S. Immigration and Customs Enforcement , Secure Communities, https://www.ice.gov/ secure-communities #tab1 (last viewed Aug . 15 , 2016 ).
48. U.S. DEP'T OF HOMELAND SEC., PRIVACY IMPACT ASSESSMENT FOR THE AUTOMATED BIOMETRIC IDENTIFICATION SYSTEM (IDENT ) 11 - 15 ( 2012 ), https://www.dhs.gov/sites/default/ files/publications/privacy-pia -nppd-ident-06252013.pdf; OFF. OF INSPECTOR GEN ., U.S. DEP'T OF HOMELAND SEC., OPERATIONS OF UNITED STATES IMMIGRATION AND CUSTOMS ENFORCEMENT'S COMMUNITIES , 4 - 5 ( 2012 ), https://www.oig.dhs.gov/assets/Mgmt/2012/OIG_ 12 - 64 _ Mar12 .pdf.
49. U.S. Dep't Of Homeland Sec., supra note 48 , at 6-8.
59. See Iverac, supra note 44.
60. See , e.g., ACLU, supra note 27.
61. Id .
62. S.F. , CAL., ADMIN. CODE ch. 12H, § 12H.1 ( 1989 ).
63. Id .
64. S.F. , CAL., ADMIN. CODE ch. 12I, § 12I.I ( 2013 ).
65. Memorandum from Ross Mirkarimi , Sheriff, S.F. Sheriff 's Dep't, to All Sworn Pers ., S.F. Sheriff 's Dep't (Jan . 28, 2015 ), http://www.catrustact.org/uploads/2/5/4/6/25464410/sf_sher iffs_dept_ice_policy_3 _2015.pdf (informing personnel that the department will no longer honor ICE detainers submitted on I-247 forms); Memorandum from Ross Mirkarimi , Sheriff, S.F. Sheriff 's Dep't , to All Pers., S.F. Sheriff 's Dep't (Mar . 13, 2015 ), http://www.catrustact.org/uploads/2/ 5/4/6/ 25464410/ice_contact,_ signed.pdf (informing personnel that the department will have only limited contact and exchange of information with ICE).
66. DeBonis, supra note 2.
67. Some media accounts state that ICE had asked the sheriff's department to detain Sanchez, rather than merely notify ICE of his release . Lee Romney , In Face of Criticism, S.F. Supervisor Reaffirms City's Sanctuary Policies , L.A. TIMES (Oct. 25 , 2015 , 5 :20 PM), http://www .latimes.com/local/california/la-me -sf-sanctuary-20151026-story .html.
68. Sernoffsky & Van Derbeken, supra note 1.
69. S.F. Bd . of Supervisors Res . 389 - 15 (Oct. 20 , 2015 ), http://www.sfbos.org/ftp/ uploadedfiles/bdsupvrs/resolutions15/r0389- 15 .pdf.
70. Emily Green , SF Supervisors OK Compromise Sanctuary City Legislation , S.F. CHRON . (May 24, 2016 , 9 :51 PM), http://www.sfgate.com/bayarea/article/SF-supervisors -OK-compromise-sanctuary-city-7943757 .php.
71. S.F. , CAL., ADMIN. CODE ch. 12H, §§ 12H.2 , 12I ( 2016 ).
72. Id .
73. Id .
74. Cal Thomas, Kate Steinle's Death and Sanctuary Cities: What Good are our laws without Enforcement? , FOX NEWS OPINION (July 13 , 2015 ), http://www.foxnews.com/opinion/2015/ 07/13/there-should -be-no-sanctuary-from-law .html.
75. See , e.g., Oversight Hearing on New York City's Policy Regarding Local Law Enforcement and Immigration Before the Subcomm. on Immigration, Border Sec. and Claims of the H. Comm. on the Judiciary, 108th Cong . ( 2004 ) (Statement of the ACLU) , https://www.aclu. org/ aclu-statement-new-york-citys-policy-regarding-local-law-enforcement-and-immigration-housejudiciary (arguing that “[a]ny benefits that might result from enlisting state and local police to enforce complex immigration laws would be far outweighed by the serious consequences of such a change .”)
86. Immigration and Nationality Act § 222 (g), 8 U.S.C. § 1202 (g) ( 2016 ).
87. Immigration and Nationality Act § 212 ( a)(9)(A)(ii)(I), 8 U .S.C. § 1182(a)(9)(A)(ii)(I) ( 2016 ).
88. See , e.g., Craig E. Ferrell Jr. , Immigration Enforcement: Is It A Local Issue?, THE POLICE CHIEF , Feb . 2004 , at 36 (“ Many local law enforcement executives can support this position because persons who are in the country illegally have violated the law and should be treated in the same fashion as other criminals .”).
89. See id.
92. Romney , supra note 67.
93. Green , supra note 70.
94. See Abbott, supra note 3.
95. For example, local officers training and implementing 287(g) agreements are paid by their LEAs . AM. IMMIGRATION COUNCIL , THE 287(G) PROGRAM: A FLAWED AND OBSOLETE METHOD OF IMMIGRATION ENFORCEMENT (Nov . 29, 2012 ), https://www.americanimmigrationcoun cil.org/sites/default/files/research/287g_fact_sheet_ 11 - 2012 _ 0 .pdf.
96. MATTHEW SEAMON , CATHOLIC LEGAL IMMIGRATION NETWORK, INC., THE COST OF STATE & LOCAL INVOLVEMENT IN IMMIGRATION ENFORCEMENT 3 ( June 2014 ), https://cliniclegal .org/sites/default/files/cost_of_involvement_in_immigration_enforcement_version_ 5_mm.pdf. Representative costs include utilities and pay for guards, administrators, cooks, and other personnel needed to operate the detention center .
97. See Morales v. Chadbourne , 966 F. Supp . 2d 19 (D.R.I . 2014 ); Miranda-Olivares v . Clackamas Cty., 2014 WL 1414305 (D. Ore. Apr . 11 , 2014 ).