Using Risk to Assess the Legal Violence of Mandatory Detention

Immigration mandatory detention is a particularly harsh example of the structural violence embedded in immigration enforcement. It deprives liberty without bond for immigrants with prior crimes, and assigns many individuals to the harsh conditions associated with unnecessary and even wrongful detention. Mandatory detention has been justified on the grounds that mandatory detainees are a danger to public safety. This article puts to the test this presumption of dangerousness among mandatory detainees, and finds, to the contrary, that immigrants with prior charges or convictions are no more dangerous than any other category of individuals in Immigration and Customs Enforcement (ICE) custody. Using the risk classification assessment (RCA) tool, which the author is the first to obtain through the Freedom of Information Act, the article contributes to the growing criticism of mandatory detention, providing evidence that many of those in mandatory detention should probably have never been detained.


Introduction
Marcel, a 29 year-old Paraguayan male, received a 33-day sentence for larceny and was mandatorily detained by Immigration and Customs Enforcement (ICE).Min Soo, a 30 year-old woman from South Korea, who, like Marcel, was mandatorily detained without bond for the offense of larceny.Because law on immigration consequences [1]. 1 that flow from a conviction for larceny 2 is in flux, in certain states Marcel and Min Soo may have a strong challenge to their classification as mandatory detainees.However, since being categorized as "mandatorily detainable", Marcel and Min Soo were ineligible for bond adjudication before an immigration judge [2]. 3Since all mandatory detainees receive a risk classification assessment (RCA), Marcel and Minn Soo were scored and categorized as low to medium risk, which suggests that were they not mandatorily detained by statute, and they should not have been detained at all.
In this article, I examine mandatory detention through the lens of the ICE's risk classification assessment (RCA).The RCA assigns risk scores to those in custody, which are used for detention decisions and security assignments in detention.Even those mandatorily detainable receive an RCA score, revealing detailed information about their criminal backgrounds on the one hand and equities on the other.These factors are calculated to reveal the detainees' risk to public safety and flight.The findings demonstrate that mandatory detainees, such as Marcel and Minn Soo, are not generally high risk, and are no more dangerous than any other group of immigrants in ICE custody who are eligible for bond or released outright, and, thus, like them, mandatory detainees should be eligible for bond, or released outright [3].Additional findings demonstrate that ICE mandatorily detains those who clearly should not be mandatorily detained under the statutes as well as those who arguably do not fall under an enumerated offense [1].
The article also examines mandatory detention through the lens of structural violence 4 to account for the physical and psychological harm that mandatory detention inflicts on noncitizens, overwhelmingly of color.Structural violence is a term ascribed to Johan Galtung in 1969, which refers to an indirect "avoidable impairment of fundamental human needs" ( [5], p. 167).It affects those at the bottom rungs of society, and can be attributed to racism, xenophobia, and nationalism.Noncitizens who find themselves in immigration custody are at this bottom rung.
As applied to law, scholars refer to legal violence to take stock of the violent effects of immigration law on noncitizens [6].Whereas "legal violence" scholarship focuses on the effects of law, this article examines the violence embedded in legal decisions themselves, which creates an unsafe space for noncitizens, and removes options for relief.Violence is compounded by the system of crimmigration law that denies bond and legitimately removes procedural safeguards.
Mandatory detention is a violence that is exerted indirectly through 8 U.S.C.1226, the process for detaining noncitizens in removal proceedings.As discussed below, structural violence is embedded in immigration law, particularly the government's plenary power to detain.Harm is created for the noncitizen along a lengthy process of denying procedural and substantive justice.The violence can be tracked to immigration law's harshly asymmetrical power dynamic, which diminishes the human dignity of noncitizens, as well as to court decisions that "bait and switch" a criminal process for a civil process.Much of the structural violence that diminishes the quality of life for the subset of noncitizens with prior crimes is concealed by assertions of legitimacy and due process, as real world decisions actually occur in private halls and behind concrete prison walls.
The concept of benevolent violence similarly describes inequality with reference to unjust social, political, and economic systems [7].This article recognizes neoliberalism as coinciding chronologically with crimmigration (early 1980s), and informing the development of individualizing and criminalizing strategies of social control, as applied to immigration law, with reference to public law and criminal law more generally.As detention capacity increased during the 1980s, private prisons like the Corrections Corporation of America started detaining immigrants for profit.The profit motive for private firms led to interest groups applying pressure to Congress for more detention facilities.Indeed, from 1985 to 1988, immigration detention bed space nearly tripled, from 6000 beds to 16,000 beds, and exploded during the 1990s with the construction and renovation of new facilities, which provided capacity for immigration authorities to detain thousands of more removable noncitizens.
By the early 2000s, ICE's capacity to detain helped to turn the rationale for detaining to depend less on individualized determinations of dangerousness and more on categorical determinations of aggravated felonies, crimes involving moral turpitude, and legislatively imposed bed quotas, which encourage mass detention.Contrary to the narrative, neither public safety nor assuring appearances at hearings are at risk, but, rather, the rule of law itself.Rule of law norms, such as as proportionality, publicity and accountability, are trampled by mandatory detention practices.
After discussing mandatory detention within the historical context of immigration law, the article will document mandatory detention as an unnecessarily harsh deprivation of individual liberty.Although the loss of liberty applies individually to noncitizens, decisions are categorical and thus take no account of individual characteristics, such as a noncitizen's legal status, risk score or likelihood 4 For discussion of structural violence, please see: [4]. of relief.They profile noncitizens that are overwhelmingly of color and impugn human dignity by detaining without bond.
For purposes of operationalizing the concept, I examine structural violence in terms of a system of law that allows and overlooks unnecessary and unjustified deprivations of liberty according to ICE's own standards and guidelines.Although all noncitizens in mandatory detention have been deprived of liberty without due process of law, I will highlight the subset of those who might not have been detained had they not been subjected to a categorical determination to mandatory detention under 236c.This category of individuals includes those who receive low and medium risk scores, and whose criminal offense is not necessarily mandatorily detainable under 8 U.S.C. 1226.It highlights the injustice of detaining those who are not a high risk and those who were wrongly detained with a flagrant disregard for fundamental rights to be free.The violence is indirect and embedded in the legal process.It exists in the detention of those with prior crimes, particularly those who, even ICE, says it would not mandatorily detain.The structural violence here is located in the lack of accountability for a wrongful detention.It is ostensibly invisible for lack of bond and review authority (aside from a Joseph hearing 5 ).In particular, the article points to a system of immigration law that allows ICE to detain on the basis of a secretive RCA algorithm, including even those it ambiguously and erroneously categorizes as mandatorily detainable.The article recommends ameliorating this structural violence by abolishing mandatory detention.

Background: The Structural Violence of Mandatory Detention
Mandatory detention exists in the legal crosshairs of criminal law and immigration law, referred to as crimmigration [9][10][11][12]. 6Crimmigration consists of a structurally violent space where immigration authorities punish noncitizens, mostly those of color, under the guise of non-punitive civil law.The space is derived from a foundational dynamic of crimmigration law: the state wielding nearly unchecked power over noncitizens holding few procedural rights and fewer substantive rights.This dynamic produces a gross disparity of power over right, and explains the exceptional status of immigration law [14].Noncitizen males of color are the most likely victims of this largely overlooked patch of the criminal justice system.This system punishes individuals for crimes for which they have already been punished-sometimes long ago-with the disproportionate punishment of mandatory detention and removal, only because this intersection of immigration and crime is technically civil not criminal, and thus affords in this instance almost no legal accountability.
The subsection of noncitizens with a prior crime on their record is particularly vulnerable to mandatory detention.For any offense that now counts as an aggravated felony or crime involving moral turpitude, such noncitizens are one interaction with law enforcement away from being torn from families, friends, neighborhoods and jobs, and tossed into mandatory detention, with the high likelihood of being removed from the country.The immigration enforcement regime that developed during the early 2000s and consists of Secure Communities, PEP, 287(g) and a dozen other enforcement programs under ICE ACESS (an enforcement initiative that includes 13 programs) is largely premised upon raising the risk of detention and removal following an unfavorable encounter with law enforcement [15].For this group it can never be said that individuals have "served their time".Every day, in this country, they face the possibility of serving additional time.

The Derivation of Structural Violence in Immigration Law
Structural violence originates in late 19th century legislation that established an immigration system, criminal in function, but civil in form.The system received legitimacy from a set of highly 5 See [8].6 See Koulish, who refers to crimmigration as the name assumed by critical immigration scholars to describe an immigration system that deploys heightened government power and acknowledges few rights and procedural safeguards, in [13].
deferential Court opinions, starting with Chae Chan Ping and Fong Yue Ting that discerned plenary powers in the political branches and then deferred to them.In 1896, the same Court that decided Plessy v Ferguson pinned plenary powers onto decisions to detain immigrants (of color) as part of the civil process [16].Unlike Plessy, overturned in 1954, these cases were not overturned, and thus toxic racism and bigotry continues to poison the process of detention and removal [16].The racist ideology that gave rise to Chinese exclusion continues to deprive individuals of their liberty through Demore v Kim [17], which cited Wong Wing as justification for mandatory detention [18].Although, in recent years, courts in several circuits have sided with detained immigrants who are subjected to mandatory detention for prolonged periods of time, such decisions remain at the margins of a regime deeply steeped in structural violence [3].
The Court envisioned plenary powers as a way for the political branches to expeditiously dispatch with imagined (and real) enemies of the state with impunity. 7Turned inward, the courts extended plenary powers to decisions about whom to detain as part of the removal process, hence ensuring immigration detention would also be off the table in terms of substantive review.Congress had the final say over whom to hold and remove and delegated this plenary power to the executive, where it now resides with the Secretary of Homeland Security.
The oppression of detained immigrants was born of this dynamic.Judicial precedent legitimized state power over immigrants and policies treating noncitizens differently than citizens, and punishing immigrants under the auspices of civil law [19].Thus, political branches could imprison without punishing, and detain without attracting the scrutiny reserved for those in criminal prisons.Legally speaking, mandatory detainees are civil detainees, not criminals, and their detention-opposed to incarceration-is not punishment.Additionally, political branches could categorically detain individuals on substantive grounds that would violate constitutional safeguards in non-immigration contexts.As an outcome, provisions also apply retroactively, and with no statute of limitations.

Mandatory Detention
The mandatory detention provision applies at the outset of the removal processes on two groups of noncitizens: (1) those convicted of certain prior crimes; and (2) those placed into non-judicial removal proceedings (i.e., removed administratively by DHS, rather than after a full hearing before an immigration judge).Additionally, at the end of removal processes, noncitizens are often placed in mandatory detention for 90 days or longer after an immigration judge issues a formal removal order. 8 In 1996, in the Illegal Immigration Reform and Responsibility Act (IIRIRA), Congress mandated that immigration enforcement authorities "shall take into custody" individuals pending removal who have committed a broad category of crimes, "when the alien is released" from criminal custody [22,23]. 9 This mandatory detention provision, U.S.C. § 1226(c), contributes greatly to over-incarceration.It provides for the automatic no-bond detention of persons convicted of certain enumerated offenses of varying degrees of seriousness-ranging all the way from minor offenses like shoplifting or possession of small quantities of drugs to major crimes of violence.It defines some of those crimes as "aggravated felonies" and "crimes involving moral turpitude", terms of art that impose severe immigration 7 Problem was, at the time customs authorities had no real capacity or infrastructure required to get this job done.See ([20], (a)(2)) ("during the [90-day] removal period, the Attorney General shall detain the alien").After a removal order is issued, different statutes and standards apply to ICE's detention decisions (i.e., 8 U.S.C. § 1231 and accompanying regulations).If ICE is unable to remove the individual within 90 days, ICE may conduct a custody review based primarily on flight risk and public safety risk ([21], 4(d)).Following the review, ICE may choose to continue to detain the noncitizen, or choose to release him or her under supervision ([20], (a)(3), (6); [21], pp.4-5).Detention, however, may not constitutionally extend beyond a period "reasonably necessary to secure removal," and the U.S. Supreme Court has held six months to be presumptively reasonable ([18], pp.699-701).However, exceptions exist for individuals deemed to be especially dangerous or security risks ([21], p. 14).

9
A mandatorily detained noncitizen can request a conviction review hearing in immigration court.See [8].The noncitizen can appeal that bond redetermination to the administrative Board of Immigration Appeals (BIA).See [24].consequences but that include minor, nonviolent offenses10 including simple possession of small quantities of drugs [25].
Since most offenses are state crimes, ICE must ensure that the state offense matches federal definitions before subjecting a noncitizen to mandatory detention.The classification of crimes that may qualify as "aggravated felonies" or "crimes involving moral turpitude" is an evolving area of law.Because of variations in interpretation among circuits, and the way a state law is articulated and applied, in each state very similar offenses are potentially subject to very different immigration consequences.There are no public data reporting the number of individuals mandatorily detained for criminal convictions, but this number is potentially significant given the large number that DHS removes following a criminal conviction ( [26], p. 13, Table 2). 11n a series of recent decisions, the Supreme Court reaffirmed that whether a particular conviction constitutes an "aggravated felony" or a "crime involving moral turpitude" is determined by whether there is a categorical "match" between the elements of the state offense as defined by statute and the "generic" definition of that offense or the way it is commonly understood. 12Using this analysis, the Fourth Circuit has found that many common Maryland offenses are not properly classified as "aggravated felonies" [32].When the state offense is broader than the generic federal definition of the crime, such offenses do not qualify as aggravated felonies or as crimes involving moral turpitude.
As several former INS officials argued in an amicus brief in Demore v Kim, absolutist rulemaking authority under 1226(c) hinders fair and efficient efforts at detention [33].Since those mandatorily detained hold fewer rights than nearly all those similarly situated elsewhere in public law, mandatory detention amounts to a wholesale rejection of a balanced approach (with applicable legal norms) to detention and release determinations that transports it outside public law norms [34,35].Unlike in the criminal justice system, for example, mandatory detention decisions are made without a neutral magistrate, individualized determination with regard to risk of flight or to public safety, clear and convincing evidence in support of a decision to deny release, or any record to explain why release has been denied ( [1]; [33], p. 9).
Crimmigration scholars have emphatically demonstrated that immigration detention functions as punishment [36]; detention conditions are punitive and demeaning [37][38][39]; detention is "indistinguishable from jails" [40], and in fact, frequently takes place in local jails that contract with ICE; and the system was deliberately designed to be punitive [41].It has also been documented that state of the art surveillance technologies have the capacity to constrain the liberty of those who are not detained but are instead placed on ankle bracelets or similar monitoring devices [14,42,43].In addition to extensively documenting the criminal function of immigration detention [44], scholars have also pointed to the failure of courts to explain how this form of detention could possibly not amount to punishment [45].This highlights the structural violence embedded in mandatory detention specifically, as well as in punitive immigration enforcement laws more generally, redirecting attention back to the history of violence and exclusion that the immigration system manifests, beginning with the Chinese exclusion cases.
The dynamic behind mandatory detention seems irresistible to political leaders across the political spectrum, which helps explain why President Obama's ostensibly liberal democratic administration has detained individuals at unprecedented levels.Over two million immigrants have been detained since 2009, about 400,000 individuals annually and about 34,000 on any given day.In fiscal year 2013, U.S. Immigration and Customs Enforcement (ICE) detained nearly 441,000 noncitizens pending deportation proceedings ( [46], p. 5)-the largest number of individuals passing through any U.S. incarceration system, federal or state [47].On any given day, 34,000 individuals were detained, leading Dora Schriro (2009), assistant to former DHS Secretary Janet Napolitano, to observe that ICE has the "largest detention...program in the country" [37].
Even as bipartisan support grows for reductions in criminal incarceration 13 including pretrial detention, 14 no similar support exists to reduce immigration detention, so as a result, ICE's detention rates remain stubbornly high.In fiscal year 2013, ICE detained nearly 80 percent of its arrestees [50], while criminal law scholars criticize jurisdictions in the U.S. pretrial detention system that unnecessarily detains 38 to 42 percent of defendants [51].Although detention numbers dropped in fiscal year 2014 to 425,478, with 33,227 individuals on a given day, the immigration system's starting point continues to be widespread over-detention [38], even more pronounced than criminal pretrial over-detention [52,53].
At almost every turn, ICE interprets mandatory detention provisions as authorizing it to absolutely detain anybody whose conviction even arguably falls under the enumerated offenses, even where the law is evolving, unclear or contradictory, and regardless of how long their detention lasts or is likely to last [1].As detention numbers suggest, it, instead, is simply resolving almost every legal or factual ambiguity in favor of the government, 15 keeping the challenger in mandatory detention unless he or she had essentially already prevailed in their immigration case.
DHS has publicly interpreted the provision mandating "custody" to require detention, i.e., physical incarceration, as a matter of DHS policy [54]. 16Some advocates and elected officials have argued that "custody" encompasses alternative measures such as electronic monitoring, based on U.S. criminal law precedents [56,57].DHS has also interpreted the provision "when the alien is released" to require detention at any time after release from criminal custody-even years later-rather than only at the time of actual release [58,59].That interpretation has been heavily litigated in federal courts, with some division among circuits as to whether ICE should take custody immediately following a criminal conviction or whether it may do so many years-sometimes decades-after the person has been released from incarceration17 [60].
With mandatory detention, hundreds of thousands of immigrants who ordinarily would have qualified for release on bond or other conditions now end up in immigration detention for the duration of their immigration proceedings.Because of the complexity of immigration cases that involve charges based on prior state crimes, these proceedings routinely take up to or longer than a year to resolve [1].Additionally, the complexity of reconciling state criminal offenses with the INA adds to the challenge of making individual determinations effectively, and leads to a systematic bias in favor of detention.These structural realities of mandatory detention inflict severe harm on immigrants, their families, and their communities, at nearly every turn.

ICE's Risk Classification Assessment (RCA)
ICE's "Risk Classification Assessment" (RCA) initiative, from the outset, was designed to advance Administration initiatives to reduce and better tailor its use of detention, and increase use of alternatives to detention (ATD), as part of its broader effort to make immigration detention "truly civil" [62]. 18RCA is nonetheless administered to all those in custody, even those mandatorily detainable, and documents the irony of a risk based system that over-detains due to the absolutist categories of mandatory detention, thus ignoring the purpose of a risk tool to reduce and better tailor its use of detention.
ICE's RCA is a computerized algorithm that ICE uses in its book-in process to assess two primary factors: Risk to public safety, and risk of flight from proceedings [64]. 19ICE officers collect criminal and immigration history throu gh records checks, and family and personal data, such as local ties, family history, residency history, or substance abuse, through an intake interview ( [65], p. 6).The algorithm then recommends detention or release, the amount of bail (if any), and detention or supervision levels. 20ICE has never released its methodology nor business rules.
ICE conducts the RCA on all noncitizens that enter its custody, including those apprehended by other DHS agencies or components.The only significant exception, as of late 2012, was that ICE officers were not required to complete an RCA for any alien for whom detention is mandatory and whose departure or removal will likely occur within five days. 21After an RCA recommendation is made, an ICE supervisor issues a final determination as to custody status.
Properly used, the RCA can tailor detention to individuals who are high risk.The RCA assesses two factors-risk to public safety, and risk of flight from immigration court proceedings-and produces an assessment of high, medium or low for each risk factor.RCA does so based on extensive data collected in the RCA process [66], 22 particularly regarding special vulnerabilities, if any exist ( [66], pp.12-13). 23ata on public safety, which comprises the salient variables for dangerousness, is likely more reliable than data on flight risk.ICE's public safety assessment is based mainly on "static" (i.e., previously-existing) information in criminal databases, which provide criminal history (i.e., records of all criminal charges, dispositions and sentences), 24 open wants or warrants, supervision history (such as bond breaches or supervision violations), and disciplinary infractions [67]. 25The one exception is ICE's evaluation of gang affiliations (also termed "Security Threat Group" status), based on information gathered by an ICE officer during the intake interview ( [69], pp.7-8; [70], p. 8, footnote 7).From "static" criminal history, the RCA also assesses whether mandatory detention for a prior crime under INA § 236(c) applies, which trumps the RCA's detention recommendation. 26.
The flight risk assessment, conversely, is based mainly on "dynamic" information, collected by an officer through ICE's intake interview.This includes local ties (such as a U.S. citizen spouse or child), family history, residency history, substance abuse history, work authorization, legal representation, and other factors (exceptions include immigration violation history, history of absconding, and pending USCIS benefit applications) [65].Some criminal researchers have found static information more predictive than dynamic information [72]. 27Moreover, noncitizen individuals might plausibly under-report some of these factors-particularly the presence of family members, due to fear of their deportation.
Based on an overall weighing of these two factors, the RCA then produces one of four recommendations to: (1) detain without bond; (2) detain but with eligibility for bond; (3) defer the decision to the ICE supervisor; or (4) release.Mandatory detention results in an RCA recommendation to detain without bond. 28For those in mandatory detention, the RCA produces one of five security level recommendations: (1) low security; (2) low-medium security; (3) medium security; (4) medium high security; and (5) high security.

Methodology
The analysis in this article stems from 485 Risk Classification Assessment (RCA) Detailed Summaries, summarizing risk assessments conducted by ICE's Baltimore Field Office from 1 March to 17 June 2013, that we received pursuant to a Freedom of Information Act (FOIA) request submitted on 4 April 2013. 29

Mandatory Detainees and Over Detention
The Baltimore sample reinforces national data about ICE over-detention.ICE detained 82% (399 of 485), and released 18% (86 of 485).Almost half (47%) of the sample (227 out of 485 cases) were mandatorily detainable regardless of the RCA's assessment of particular public safety and flight risks, as shown in Table 1.The high detention rate was at least in part the result of mandatory detention provisions, as shown in Table 2, and otherwise of a risk assessment tool highly skewed towards detention.Mandatory detention though was perhaps the most important factor that contributed to high detention rates in the Baltimore sample.Of the 334 individuals for whom the RCA issued a detailed summary and identified as being in a particular type of proceeding (judicial, administrative, expedited, or reinstatement), nearly two-thirds (63%, or 209) were mandatorily detainable. 30turpitude is "quite complex") ([25]; [71], pp.1488-89).Even when mandatory detention applies, RCA apparently performs the risk assessment so as to recommend security levels within detention (i.e., low/medium/high-security). 27 For a comparison of static to dynamic factors, see [73]. 28In August 2013, ICE streamlined the RCA process in expedited removal cases by generating an automatic detain decision, thus allowing field officers to skip the those submission and approval steps ([66], p. 13). 29 .I believe these are the first publicly available samples of the RCA.I coded and inputted the information in the 485 samples into a database. 30We use the phrase mandatorily "detainable" rather than mandatorily detained because ICE supervisors released a handful of these individuals that the RCA identified as mandatorily detainable.ICE classified 30 percent (101 of 334) of the arrestees as mandatorily detainable for a prior crime in identifiable removal proceedings: 72 in judicial proceedings, and 29 in non-judicial administrative removal proceedings. 32 Although in theory, mandatory detention statutes were intended for only the most dangerous individuals, this is not how the system operates in practice.First, of the individuals mandatorily detainable for prior crimes, only 25% (25 of 101) were high risk to public safety, 58% (59 of 101), were medium risk, and 17% (17 of 101) were low risk to public safety, even by ICE's own risk assessment tool, which is already biased towards high-risk assessments. 34 Second, were it not for mandatory detention statutes and allegations, according to ICE's own guidelines, 5% of those mandatorily detained would have been released outright, and 48% (48 of 101) would have been deferred to the supervisor for what amounts to a discretionary decision about detention.In other words, if ICE limited detention to those most dangerous, about 50% of those mandatorily detainable could be released from physical detention.
Third, the individuals mandatorily detainable without bond for prior crimes collectively were virtually indistinguishable regarding risk from others detained or released in judicial proceedings.RCA assessed the vast majority of discretionary cases as being medium or high flight risks, medium or high public safety risks, or both.Just 23 percent of discretionary cases were classified as high public safety risks (59 of 258 cases) and 39 percent were classified as medium public safety risks (101 of 258 cases). 35  32 This does not include one individual listed as with a "bag and baggage" letter, and two individuals listed as "other".This does include the three individuals ICE released. 34See [3] for discussion about how ICE arrives at this determination, and why ICE's risk classifications are somewhat questionable. 35The individuals listed as "bag and baggage" tended to skew towards higher flight risk and lower public safety risk compared to the individuals listed as in judicial proceedings.Of the 107 "bag and baggage" individuals, RCA assessed 78 percent (83 of 107 cases) as high flight risk, 21 percent (22 of 107 cases) as medium flight risk, and 2 percent (2 of 107 cases) as low flight risk; while RCA assessed 22 percent (24 of 107 cases) as high public safety risk, 27 percent (29 of 107 cases) as medium public safety risk, and 50 percent (54 of 107 cases) as low public safety risk.The higher flight risk assessments may be Many mandatory detainees are not a very high risk because their crimes are relatively minor and they have significant equities.Mandatory detainees are held without bond regardless of risk to public safety [43].Most are assessed as a medium public safety risk by the RCA tool, which only rarely-seventeen percent of the time-classifies anyone as a low public safety risk [3].
Since 2009, the Administration has repeatedly stated that immigration detention was not designed for avoiding flight risk [37,74].Indeed, those mandatorily detainable in judicial proceedings were collectively lower flight risks than others in judicial proceedings.
However, corresponding risk assessments for flight were more weighted in the higher risk categories: 31% (31 of 101) were high risk for flight, 54% (54 of 101) were medium risk, and 16% (16 of 101) were rated a low risk of flight.In comparison, RCA assessed a majority of discretionary cases as high flight risks (54 percent, 140 of 258 cases), and almost all of the remainder (42 percent, 108 of 258 cases) were classified as medium flight risks. 36Not only should individuals not be detained for flight risk per design, perhaps both they and the taxpayer would be better served by some less intrusive, less costly alternative to detention in those cases.

Tenuous Mandatory Detention Classifications
At least some of those classified by the RCA tool as mandatorily detainable appear to have substantial legal challenges to that classification.This is because under existing case law, a number of state offenses categorically do not qualify as aggravated felonies or as crimes involving moral turpitude [1].
In some cases the misclassification was clear-for example, a person was convicted only of second-degree assault and the Fourth Circuit had already ruled that that was not an aggravated felony [1].In other cases, there was no in-circuit precedent directly on point for that particular crime, but a legal analysis based on the approach required by the Supreme Court-matching the elements of the state offense to the generic definition of the crime-would result in the same conclusion.For example, the Maryland theft statute should not qualify as an aggravated felony under this analysis, as the BIA has now twice found in unpublished opinions; but there is no precedential case directly on point and ICE, at least for some time, continued to place persons convicted of the most minor theft offenses in mandatory detention [1].
What this shows is that: (1) in some instances ICE is not performing due diligence to ensure that those it classifies as mandatory detainees are actually so as determined by existing case law directly on point, and is detaining under § 1226(c) persons who should actually be provided a bond hearing under § 1226(a); and (2) ICE detains anybody who could arguably fall under the purview of the statute even where that determination is tenuous or a substantial legal argument to the contrary exists.Rather than because individuals with a "bag and baggage" letter have a final removal order, and thus are assumed to have less incentive to appear for removal.Additionally, ICE may have re-apprehended many of these individuals after absconding.being weighted in favor of the minimal process of a bond hearing, ICE and the RCA's determinations are thus systematically weighted in favor of over-detention.This is exactly the reverse of the situation in the criminal justice system, where the presumption is that everybody is entitled to bond hearing and most actually qualify for release on bond [1].
Applying the categorical approach based on Fourth Circuit case law applicable to Maryland, we have found that potentially 42 percent 37 of those mandatorily detained for prior crimes could have had legal challenges to that classification and should therefore have been subject only to discretionary detention under § 1226(a) [1]. 38 The RCA appears to incorrectly categorize several Maryland criminal offenses within the federal category for mandatory detention [1].For example, the RCA categorized Maryland theft convictions as mandatorily detainable in the 4th Circuit even though that classification is highly tenuous under existing precedent [2]. 39ICE similarly erred in mandatorily detaining individuals with convictions for second-degree assault [75] and possession of cocaine [76].
The category of tenuous and questionable mandatory detention classifications consisted of individuals who were generally not high risk.Most involved minor non-violent crimes, many had family or community ties, and most were assessed as either a medium or a low risk to public safety.
Instead, individuals in this category were detained regardless of their public safety risk: 25 percent (10 of 40) were assessed a low risk to public safety; 60 percent (24 of 40) a medium risk to public safety; and only 15 percent (6 of 40) a high risk to public safety.These ratings are almost identical to those within the data set who were subject to discretionary detention and then released from custody: 25 percent low; 58 percent medium; and 17 percent high [3].Thus, most of those incorrectly mandatorily detained could have been released, saving significant resources and avoiding high personal costs for them and their family members [1].
In reality, a full consideration of equities would likely result in a greater percentage of low risk classifications-especially since the RCA algorithm takes no account of various relevant factors such as the age of the conviction.Thus, the conviction may have occurred two or more decades prior to the person's coming into ICE custody, but the RCA tool takes no account of that fact in its assessment.Others are very young or have family and community ties and relatively minor crimes, but are still rated as medium risk for flight and public safety.Although it is difficult to uncover the full range of a person's equities using just the RCA tool, at least a few cases appear even on that cursory analysis to qualify for a lower rating.For example:

‚
Juan (May #235) is a 25-year-old Guatemalan man who had been living at the same address with immediate family members for at least six months and who had a spouse, child, or other family member in the local community.He was mandatorily detained pursuant to a conviction for burglary, categorized as a crime involving moral turpitude.The sentence was less than a year, which indicates a relatively minor or first-time offense.Juan was rated a medium risk to public safety and a medium risk of flight.
‚ Josue (June #491) is a 19-year-old from Trinidad, Tobago who was subjected to mandatory detention for selling a controlled substance, a nonviolent offense.He was under 21 and had been 37 This figure 42% is the result of legal analysis of statutes within Maryland and case law within the 4th circuit compared to federal immigration law. 38The limitations of this type of analysis with many unknown variables means that this number is only about potential legal challenges, and not necessarily indicative of the number of people who would not ultimately be found subject to mandatory detention based on their offenses.But the number illustrates the lack of due process inherent in making such a determination without providing a real opportunity for individuals to dispute their classification.It also illustrates that arguable cases are always placed in mandatory detention rather than provided a bond hearing. 39See [2] ("Omargharib's 2011 conviction for grand larceny, in violation of Va.Code Ann.§ 18.2-95, was not a 'theft offense' under the INA").Applying a similar analysis, the BIA has now twice found [cite cases] in unpublished decisions that MD theft is not categorically an aggravated felony.The exact same reasoning applies to the question of whether it is a CIMT.Thus, although there is no controlling precedent exactly on point, an application of the categorical approach mandated by the Supreme Court would lead to the conclusion that those persons are not subject to mandatory detention.
living with immediate family members at his address for at least six months.He also was rated a medium risk for public safety and a medium risk for flight.

Discussion and Recommendations
Although measuring structural violence in mandatory detention is an imprecise science, we have found that ICE detains those who are not high risk and those whose state crimes arguably fail to match federal immigration categories.From studying ICE's use of risk assessment in Baltimore, I draw several conclusions about ICE's use of detention (although caution should be used in drawing national conclusions from one city). 40 First and foremost, mandatory detention provisions prevent the ICE risk assessment from facilitating reductions in detention or better tailoring to public safety risk.ICE's factual assessment that its detainees primarily pose flight risk-not public safety risk, which DHS historically has cited to justify detention-validates arguments for alternatives to detention (ATD) that ensure appearance at proceedings at lower cost.ICE's discretionary decisions have not reflected that potential to use ATD, nor have ICE's decisions for those with special vulnerabilities.Indeed, if similar nationwide, ICE's factual assessments would undercut Congress' rationale for its national detention bed minimum of "not less than 34,000 detention beds daily" [77].

Mandatory Detention Provisions Prevent ICE's Risk Classification Assessment from Reducing Detention Rates and Tailoring Detention to Public Safety Risk
Mandatory detention provisions prevent ICE's risk assessment from reducing the rates of individuals detained, and helping ICE to tailor detention to public safety risks, as criminal risk assessments have done. 41Along these lines, ICE's enforcement memoranda have long prioritized mandatory detainees alongside public safety threats, in recognizance of applicable law. 42That said, DHS has some ability to modify the mandatory detention provision for prior crimes, and more limited ability to modify its processes for those seeking asylum relief in non-judicial removal proceedings, once mandatory detention no longer applies. 43

Prior Crimes
If DHS or the BIA (DOJ) reinterpreted the term "custody" in 8 U.S.C. § 1226(c) to encompass alternatives to detention ( [55], p. 128, n. 669), ICE could likely release into supervision a significant number of low-and medium-public safety risks with prior criminal convictions, without increases in crime or flight from proceedings.Indeed, ICE officers may already be exercising this discretion on a case-by-case basis, as Baltimore's ICE office released three of these individuals. 44  40 For example, it is likely that an ICE field office closer to the southern border would detain a higher percentage of those in nonjudicial proceedings. 41ICE officials have already identified mandatory detention laws as an impediment to expanding ICE's alternatives to detention programs ( [50], p. 28). 42See [78] ("as a general rule, DHS detention resources should be used to support the enforcement priorities noted above or for aliens subject to mandatory detention by law").This is not a change.See ([74], p. 3) (Morton memos also stated that detention resources should be used "for aliens subject to mandatory detention by law").Additionally, a 2004 ICE policy memo, still in effect, places mandatory detainees first among detention priorities, above other high priorities such as "national security interest aliens", "aliens who exhibit specific, articulable intelligence-based risk factors for terrorism", "aliens who present an articulable danger to the community", or "suspected alien and narcotics smugglers."[? ].Moreover, "[i]n the case of mandatory detention", the 2004 memo directs guidelines to be heeded "strictly", with ICE managers given no discretion to deviate (unlike other categories) [? ]. 43 Regarding post-removal order mandatory detention, it is difficult to draw conclusions from the Baltimore data, because of the small sample size of those within 90 days of their removal order. 44Additionally, ICE memoranda direct officers to consult with ICE counsel when encountering an individual who is mandatorily detainable but has a special vulnerability ( [78], p. 5).If release is not possible, it is unclear what purpose the consultation would serve.
Those mandatorily detained pending judicial proceedings would be logical targets for release.These individuals were almost indistinguishable regarding public safety risk from those released in judicial proceedings.Many individuals also self-reported strong U.S. family, residency and community ties, diminishing their flight risk.Correspondingly, many may have stabilized their lives and strengthened those U.S. ties since their crime that triggered mandatory detention, as a majority (60 percent) were mandatorily detained for a crime unrelated to their recent ICE encounter. 45dditionally, according to other studies, a significant minority of mandatory detainees (for a prior crime) has a chance to defeat deportation, further reducing their risk of fleeing before a hearing. 46f ICE began to release lower public safety threats that are now mandatorily detained pending judicial proceedings, it could impact significant numbers-perhaps tens of thousands in a year. 47oing so could free up resources for ICE to detain more significant public safety threats, or use for other means, and also minimize humanitarian concerns based on family separation, particularly applicable to these mandatory detainees for prior crimes. 48

Non-Judicial Removals
As DHS shifts enforcement resources to the border and removes more through non-judicial processes, mandatory detention through these processes may increasingly prevent DHS from allocating detention resources to public safety risks. 49Moreover, DHS likely cannot reinterpret these Congressional provisions to encompass alternatives to detention at the outset.Both the expedited removal and reinstatement provisions mandate "detention", not custody.
A key question is whether Congress should continue non-judicial mandatory detention for arrestees whom ICE generally assesses as high flight risks, but low public safety risks.For example, RCA in Baltimore classified 100 percent of those in expedited removal as "high" flight risk and 88 percent as "low" public safety risk. 50The disparity was not quite as pronounced for those in reinstatement, who previously spent some time in the U.S. 51  Flight risk alone can be addressed through supervision and monitoring, while public safety risk cannot.In 1996, Congress imposed mandatory detention in the expedited removal process because of concerns that border crossers without documentation would not show for hearings if released [87].Detention prevents flight and ensures successful removal, of course.However, since 1996, supervision techniques including electronic supervision have emerged that also show successful prevention of flight (if not 100 percent), as discussed below ( [36], pp.1512-15; [88]).

ICE Assesses Its Detainees as Primarily Posing Flight Risk, Which Alternatives to Detention Can Mitigate at Lower Cost
More broadly, if ICE's nationwide factual assessments are similar to those in Baltimore (as ICE's detention outcomes were), strong arguments exist that ICE could increase its use of lower-cost alternatives to detention (ATD).RCA assessed the vast majority of the Baltimore arrestees (not just those in expedited removal) as flight risks, but not serious public safety risks.Put another way, ICE primarily used detention to ensure that immigrants appeared for removal processes, rather than to prevent immigrants from committing crimes.ICE's Baltimore results contradict ICE's publicly stated rationales for detention, which have long focused on public safety.For example, ICE Director Sarah Saldaña recently stated that ICE has "focused our detention and removal resources on public safety and national security threats to ensure we are doing everything we can to keep America safe" [89]. 52ICE's detention outcomes in Baltimore also contravene one of the purposes of RCA-to encourage alternatives to detention for those not dangerous ( [47], p. 20).Indeed, ICE's recent budget request reiterates ICE's goal to "[c]ontinue to prioritize aliens for ATD who present the highest risk of flight" ([90], p. 66).Flight risk is a valid purpose for detention, and detention prevents flight at a 100 percent rate. 53However, if those released will not threaten public safety, the question is whether ICE can use alternative mechanisms to achieve nearly the same results with far lower financial, human and social costs.ICE's alternative to detention program, called the Intensive Supervision Appearance Program ("ISAP"), has demonstrated remarkably high success in ensuring appearance at court hearings.From fiscal years 2011 to 2013, 95 percent of participants in ISAP's "full-service" program, which involves supervision, technology monitoring (either electronic GPS tracking or phone reporting), periodic visits and case management, appeared at their scheduled removal hearings ( [50], pp.30-31). 54Concurrently, DHS' Inspector General reported that in FY 2012, only 4.9 percent of ISAP participants absconded, and 4 percent were arrested by another law enforcement agency ([66], p. 6).These success rates echo earlier findings regarding alternatives to detention. 55  removing those with a criminal history.See ([26], p. 11).Among the general population, though, research has showed immigrants to generally have lower crime rates than the native-born [86]. 52ICE's FY 2016 budget request reiterates ICE's goal to "focus costly detention space on criminal and priority aliens" ( [90], pp. 64-66).ICE has aligned its capabilities with immigration enforcement priorities and policies so that "mandatory and priority individuals (such as violent criminals and those who pose a threat to national security) are kept in detention". 53Expedited removal provisions were enacted largely in response to flight risk concerns.See [91] (the U.S. estimated that only 50 percent of released noncitizens appeared for proceedings, and only 11 percent of those ordered removed but not detained complied with removal). 54ISAP also has a "Technology-only" program, which involves only electronic GPS tracking or phone reporting.ICE has not had data to track the success of the "technology-only" program.In the full-service program, 52 percent are monitored by telephone and 48 percent by GPS, while in the "technology-only" program, 97 percent are monitored by phone, and 3 percent by GPS ([50], p. 26). 55An earlier U.S. study from 1997 to 2000 showed that 92% of criminal aliens released under supervision attended all of their hearings, and concluded that "mandatory detention of virtually all criminal aliens is not necessary."([92], pp.33, 36).Even 82% of criminal aliens released on recognizance without supervision appeared, as did 77% of those released on bond.
The same study showed that 93 percent of asylum seekers in expedited removal with credible fear, whom were provided intensive supervision, appeared at their immigration proceedings ( [92], pp. 10, 27).As to asylum seekers, the Vera study also showed 84 percent compliance by asylum seekers even with minimal supervision without potential redetention, and 78 percent compliance by those simply released without supervision [92].The study thus concluded that "[a]sylum seekers do not need to be detained to appear...They also do not seem to need intensive supervision."( [92], p. 31).
The Obama Administration is attempting to increase its use of ATD.ICE requested $122 million for alternatives in fiscal year 2016-an increase over past years, albeit still much less than its $2.4 billion request for detention [90]. 56If alternatives to detention were implemented more broadly-and in conjunction with reductions in detention-there would be significant cost savings [55]. 57A U.S. GAO report found the average daily cost of the ISAP program to be $10.55 per day, compared to the $159/day estimated cost of detention ( [50], p. 18).One independent report found that ICE could save $1.4 billion if it could place every detained noncitizen not convicted of a serious crime into ATD (leaving aside the bed minimum requirement) [40,93].Immigrant advocates have also argued that ATDs impose far less social burdens on immigrants and their families ( [26], p. 46).

Conclusions
Mandatory detention is a form of structural violence.It is an extraordinary deprivation of liberty that provides neither bond nor relief.Such deprivation is reinforced by a system of law that treats noncitizens worse than citizens and provides weak civil law safeguards against immense enforcement power.The violence is made even harsher by ICE's erroneous calls for mandatory detention that go enforced without review.Such harm is generally limited to those without counsel and of color.
RCA provides DHS with the means for tailoring detention to those who score as a high risk to public safety.With the increasing use of alternatives to detention for flight risk, there is no reason to imprison individuals who score a high flight risk.With RCA and ATD, a properly tuned algorithm can help ICE to detain those who score a high risk to public safety, and no one else.In other words, with RCA, there is no need for mandatory detention.
With RCA data, I have been able to examine the consistency of RCA mandatory detention decisions to find the deprivation of liberty process to be surprisingly ad-hoc and arbitrary.Further, RCA sheds light on the actual dangerousness of mandatory detainees.The presumption of dangerousness is unrelated to the actual dangerousness of individuals who are mandatorily detained.First mandatory detainees are no more dangerous, or risky, that any other immigrant in immigration custody.Second, the RCA findings reveal that individuals may sometimes mistakenly be classified as being subject to mandatory detention [1].
These findings also reinforce the argument increasingly made in circuit courts as they order bond hearings for those mandatorily detained for long periods of time.In Rodriguez v Robbins, the 9th Circuit said the government must bear the brunt of establishing flight risk and dangerousness to protect immigrants from erroneous deprivations of liberty.More recently Lora v Shanahan in 2015 in the 2nd Circuit created a bright line rule that requires the government to give individuals a bond hearing before an immigration judge within six months of the start of detention.According to the court, the government must meet the burden of showing "clear and convincing evidence of those individuals are flight risk or dangerous to the community".
Since the RCA shows many individuals in mandatory detention are not risky, perhaps mandatory detention's days are numbered.
RCA is likely to be the tool with which to assess dangerousness for mandatory detainees within six months of detention.I recommend ICE re-examine the RCA algorithm to ensure that individuals in mandatory detention indeed should be there by law.The six-month rule from Lora and Rodrigues, notwithstanding, it also seems imperative for ICE to reconcile the administration of mandatory detention with actual dangerousness prior to a mandatory bond hearing at six months.A good way to do this would be for Congress to remove mandatory detention from the books and replace it with discretionary detention-based on RCA-for all individuals in ICE custody. 56See [90] (requesting an increase of $28 million over the President's budget for FY 2015, to increase the daily population in ATD from 27,219 participants at the end of FY 2014 to 53,000 participants). 57See also ([52], pp.1372-74) (similar costs of criminal justice monitoring).
As much as mandatory detention is also a tool of institutional violence: a creature of plenary power, with gross asymmetries of power between the state and the individual, it will likely be left to the political process-not the courts-to mandate shifts in policy as part of comprehensive immigration reform.
With such findings, it is also reasonable to suggest that since ICE is now in possession of technologies that can assure immigrants appear in court and for removal, and can tailor physical detention to individuals who really are dangerous, some courts have begun to claim that mandatory detention is unnecessarily punitive.

8
50  of 107 individuals are listed with a history of absconding, and of those 50, 47 are assessed as high flight risk.Comparatively, of the 125 individuals in judicial proceedings, RCA assessed 31 percent (39 of 125 cases) as high flight risk, 62 percent (78 of 125 cases) as medium flight risk, and 6 percent (8 of 125 cases) as low flight risk; while RCA assessed 22 percent (27 of 125 cases) as high public safety risk, 54 percent (67 of 125 cases) as medium public safety risk, and 25 percent (31 of 125 cases) as low public safety risk.Only 2 of the 125 individuals are listed with a history of absconding.36The individuals listed as "bag and baggage" tended to skew towards higher flight risk and lower public safety risk compared to the individuals listed as in judicial proceedings.Of the 107 "bag and baggage" individuals, RCA assessed 78 percent (83 of 107 cases) as high flight risk, 21 percent (22 of 107 cases) as medium flight risk, and 2 percent (2 of 107 cases) as low flight risk; while RCA assessed 22 percent (24 of 107 cases) as high public safety risk, 27 percent (29 of 107 cases) as medium public safety risk, and 50 percent (54 of 107 cases) as low public safety risk.The higher flight risk assessments may be because individuals with a "bag and baggage" letter have a final removal order, and thus are assumed to have less incentive to appear for removal.Additionally, ICE may have re-apprehended many of these individuals after absconding.50 of 107 individuals are listed with a history of absconding, and of those 50, 47 are assessed as high flight risk.Comparatively, of the 125 individuals in judicial proceedings, RCA assessed 31 percent (39 of 125 cases) as high flight risk, 62 percent (78 of 125 cases) as medium flight risk, and 6 percent (8 of 125 cases) as low flight risk; while RCA assessed 22 percent (27 of 125 cases) as high public safety risk, 54 percent (67 of 125 cases) as medium public safety risk, and 25 percent (31 of 125 cases) as low public safety risk.Only 2 of the 125 individuals are listed with a history of absconding.
ICE provided 505 RCA Detailed Summaries to the authors through a non-adversarial FOIA process, in a series of four productions from September 2013 to June 2014.All were from ICE's Baltimore Field Office (which spans the state of Maryland), in four batches labeled "March 2013", "April 2013", "May 2013", and "June 2013" (ICE represented that the last batch was incomplete).After excluding duplicates and incomplete samples, 485 samples remained.98 percent of these samples represent RCAs conducted from March 1 to 17 June 2013 (leaving aside four conducted between December 2012 and February 2013, and five conducted from 17 June to 15 July 2013).Based on this, and conversations with ICE's FOIA officer, we believe they reflect the totality of the risk assessments that ICE's Baltimore Field Office conducted from March 1 to 17 June 2013.ICE also informed us at one point that it would provide several thousand pages of records from the New York Field Office.The RCA Detailed Summary is printed out from ICE's Enforcement Case Tracking System (ENFORCE)-the primary case management system for ICE's Enforcement and Removal Operations (ERO) office-through the ENFORCE Alien Removal Module (EARM)

Table 1 .
Detention Outcomes for Mandatorily Detainable and Discretionary Cases.

Table 2 .
Detention Outcomes for Mandatorily Detainable and Discretionary Cases, ICE Baltimore Field Office Sample, Spring 2013.
Source: Author analysis of risk classification assessment (RCA); Detailed Summaries from ICE's Baltimore Field Office, March through June 2013, provided by ICE through FOIA, on file with the author.