Piecing Together the US Immigrant Detention Puzzle One Night at a Time: An Analysis of All Persons in DHS-ICE Custody on September 22,

Piecing Together the US Immigrant Detention Puzzle One Night at a Time: An Analysis of All Persons in DHS-ICE Custody on September 22, 2012 1 Donald Kerwin Center for Migration Studies With Daniela Alulema (Center for Migration Studies) and Siqi Tu (Center for Migration Studies) Executive Summary This paper analyzes a dataset of every person in the custody of the US Department of Homeland Security Immigration and Customs Enforcement (DHS-ICE or ICE) on September 22, 2012, and compares this data with an earlier analysis of a similar dataset on detainees in DHS-ICE custody on January 25, 2009. DHS-ICE provided the 2012 and 2009 datasets in response to Freedom of Information Act (FOIA) requests from the Boston Globe and Associated Press. The paper sets forth findings related to: (1) the removal adjudication processes to which the detainees were subject; (2) the facilities in which they were held; (3) their length of detention; and (4) their criminal histories, if any. It finds that on September 22, 2012: DHS-ICE held 35,197 people in its custody. 18,470 detainees had pending removal cases, 14,674 had been ordered removed, and 2,053 cases included no information on whether or not the detainee had been ordered removed. Thirty-eight percent of detainees were subject to summary, non-court removal processes. Forty percent of detainees were from the Northern Triangle states of Central America and 34 percent from Mexico, compared to 37 percent from Mexico and 28 percent from Central America on January 25, 2009. Detainees were held in 189 facilities, with 77 percent concentrated in nine states and 51 percent in the four states that border Mexico. DHS-ICE held 67 percent of all detainees in facilities owned and/ or administered by for-profit prison corporations and 90 percent of detainees in the 21 facilities with the largest detention populations. 1 The authors would like to acknowledge and extend their thanks to Mark Noferi for his initial review, assessment, and research on the September 22, 2012 detention dataset. 2015 by the Center for Migration Studies of New York. All rights reserved. JMHS Volume 3 Number 4 (2015): 330-376

Piecing Together the US Immigrant Detention Puzzle Forty-seven percent of detainees had been held for less than 30 days, and 4,179 (12 percent) had been held for more than six months. Of those ordered removed and continuously detained in the interim, 553 persons had been detained for more than six months after being ordered removed, despite being presumptively eligible for release after six months. Sixty-one percent of detainees had criminal convictions, compared to 42 percent on January 25, 2009. Less than 10 percent of all detainees on September 22, 2012 had committed violent crimes and substantial percentages a higher rate than on January 25, 2009 had committed traffic and immigration violations. The paper argues that DHS-ICE should collect extensive data on detainees, make it broadly available on a consistent basis, and become more transparent and accountable to the public. It recommends that: DHS-ICE comprehensively analyze its detention information systems and data collection to ensure that the agency can meet its legal imperatives, operational responsibilities, and detention standards. DHS-ICE initiate a process to engage Congress, other federal agencies, states, scholars, researchers, and non-governmental organizations (NGOs) on the detention information and data that it should collect and make publicly available. DHS-ICE release its Enforce Alien Removal Module (EARM) and the Enforce Alien Booking Modules (EABM) manual that describes the detention information it collects, and that it regularly release extensive data on all those held in its custody. DHS-ICE work with the immigration courts and detainee states of origin to avoid prolonged and indefinite detention in all cases. DHS-ICE collect complete and accurate data and that it not leave data fields blank or fill them out incorrectly. Congress and DHS-ICE substantially expand community-based, supervised release, alternative-to-detention programs. Congress and DHS-ICE reduce, if not eliminate, the role of for-profit entities in administering immigrant detention facilities, and take full responsibility for this core government responsibility. Introduction Over the last 20 years, government research and accountability agencies and nongovernmental organizations (NGOs) have created a substantial body of knowledge on the operation of the US immigrant detention system. By and large, however, policy makers, researchers, the public, and even front-line immigration officials have not had the benefit of information on detainees collected by the US Department of Homeland Security (DHS). In addition, DHS has never released a complete list of the fields of data that it collects on immigrant detainees or the facilities that hold them. Nor has it released even a redacted 331

Journal on Migration and Human Security version of the Enforce Alien Removal Module (EARM) and the Enforce Alien Booking Modules (EABM) manual, which reportedly describes the information DHS collects and tracks on detainees, the officials responsible for entering and screening this data, and those who can access this information. DHS implausibly claims that this manual is law enforcement sensitive. Although hardly sufficient, DHS s Immigration and Customs Enforcement (DHS-ICE or ICE) agency has released two datasets with distinct sets of data fields in response to Freedom of Information Act (FOIA) requests by Associated Press reporter Michelle Roberts and Boston Globe reporter Maria Sacchetti. The datasets provide a rare glimpse of the US immigrant detention system on the nights of January 25, 2009 and September 22, 2012. Roberts and Sacchetti relied on the 2009 and 2012 datasets, respectively, to report on long-term detention (Roberts 2009; Sacchetti 2014). An analysis of the 2009 dataset by the Migration Policy Institute concluded that ICE did not appear to track all of the information it would need to assess eligibility for release or compliance with its detention standards (Kerwin and Lin 2009, 5). An examination of detainees on a particular night or nights offers a limited and potentially skewed view of the detention system. For example, it yields a higher percentage of long-term detainees than would an analysis of a dataset of all DHS detainees over the course of a year. However, a point-in-time analysis can produce valuable information on the demographic characteristics of detainees; the types and concentration of detention facilities; length of detention for persons in different removal processes; and the criminal backgrounds of detainees. Coupled with other information, single-day datasets also offer insight into the effect of particular enforcement programs and patterns of immigration violations. Analyses of multiple point-in-time datasets can suggest long-term trends in detention and removal. The 2009 and 2012 datasets contain some overlapping fields, which allow for comparison of the detainee populations on two days, 44 months apart. This paper will summarize key findings from an analysis by the Center for Migration Studies of New York (CMS) of the September 22, 2012 dataset, and will compare these findings with those from the analysis of the January 25, 2009 dataset. The paper will begin with a description of the DHS-ICE ENFORCE database and a summary of the fields contained in the 2009 and 2012 datasets. It will then set forth findings and will conclude with policy recommendations. 332 The ENFORCE Database DHS-ICE operates the Enforcement Integrated Database (EID), which is a shared common database repository with information on the investigation, arrest, booking, detention, and removal of persons encountered by ICE, Customs and Border Protection (CBP), and US Citizenship and Immigration Services (USCIS) (DHS 2011). DHS personnel previously accessed the data in EID using the ENFORCE Alien Detention Module (EADM), and ENFORCE Alien Removal Module (EARM), but EADM has been merged into EARM (ibid.). 2 In recent years, EARM has added information on electronic bond processing, 2 EADM and EARM replaced the DHS Deportable Alien Control System (DACS) on September 30, 2007 (DHS-OIG 2009). DACS was created, in part, to enable the Immigration and Naturalization Service (INS)

Piecing Together the US Immigrant Detention Puzzle select ICE detention policies (like telephone calls following transfer), medical screening, eligibility for alternative-to-detention (ATD) programs, and criminal history (ibid., 9-12). DHS-ICE has also added to EARM a new tab on its risk classification assessment (RCA). The RCA guides ICE determinations on eligibility for release, ATD programs, and placement within the system based on flight risk and public safety concerns (DHS 2012, 3-4). On August 22, 2012, the Boston Globe submitted a FOIA request for data on every detainee in ICE custody, including information on: name and date of birth (or age) of each detainee; country of origin; criminal convictions; most serious criminal conviction; total number of days in ICE custody; and date of final removal order. ICE initially responded to the request with date of birth and other data fields, but subsequently informed the Boston Globe that it had provided some inaccurate information and had improperly provided the date of birth information. It subsequently released eight fields of data for the night of September 22, 2012: Citizenship ; Criminal Yes/No (convicted of a crime); Most Serious Criminal Charge (based on National Crime Information Center (NCIC) categorization); 3 Detention Facility ; V_Days_In_Custody (days in custody from most recent apprehension); Yes/No ; Date (the date the removal order became administratively final); and Case Category (the type of removal process and posture of case). ICE indicated that the data came from its Enforcement and Removal Operations (ERO) division. However, it did not provide a comprehensive list of data fields within this dataset or an explanation of the fields for which data was provided. Nor would ICE-ERO officials subsequently clarify ambiguous fields, including case category codes or indicate the DHS officials (likely deportation officers or immigration enforcement agents) who entered the data. CMS has pieced together some of this information from the analysis of the 2009 dataset and from other sources. The case category field included a series of codes and descriptors which speak to the type of removal proceeding and posture of the case. Analysis of the information captured and Department of Justice (DOJ) to count and report on those it had under docket control and to track non-citizens, both detainees and non-detainees, from the point that they were initially encountered by DHS officials until their departure or the closure of their cases. 3 The NCIC is the Federal Bureau of Investigation s clearinghouse and database of crime data, which can be accessed by federal, state, local, and tribal law enforcement agencies. The most serious criminal charge does not necessarily reflect the crime which serves as the ground for removability (Kerwin and Lin 2009, 23). 333

Journal on Migration and Human Security in this field reflects changes to the US immigration system brought about by the Illegal Immigration Reform and Immigrant Responsibility Act of 1996 (IIRIRA). Prior to IIRIRA, removable non-citizens fell into two general categories: the deportable and the excludable. EARM has retained this nomenclature. Non-citizens who had entered the country pre- IIRIRA had greater constitutional rights and were placed in deportation proceeding, while those stopped at a port of entry or apprehended near a border were deemed to be excludable based on the legal fiction that they had not entered the country. IIRIRA combined deportation and exclusion into a uniform removal process and replaced the term entry with admission, which it defined as a lawful entry after inspection and authorization. 4 Thus, post-iirira, non-citizens who illegally enter the country are deemed inadmissible because they have not been inspected and authorized. Immigration law has also retained separate grounds of inadmissibility (which apply to those not admitted) and deportability (which apply to those admitted). 5 Non-citizens who are subject to grounds of inadmissibility bear the burden of showing that they are admissible. By contrast, the government bears the burden of showing that non-citizens are deportable. IIRIRA also created several summary and expedited removal processes which do not include an appearance before an immigration judge (IJ) (ACLU 2014, 15-30). These informal processes have been applied to a sharply escalating number and percentage of persons facing removal since 1996. Expedited removal, for example, applies to non-citizens who attempt to enter the United States without entry documents or by using improper documents, who are arrested within 100 miles of US land and coastal borders, and who have been present in the United States for less than two years. This process accounted for 193,032 removals or 44 percent of all removals in 2013 (Simansky 2014, 5). Reinstatement of removal applies to persons who have been ordered removed or have departed voluntarily while under a removal order, and who illegally re-enter the United States. 6 This process accounted for 170,247 removals or 39 percent of all removals in 2013 (ibid.). removal applies to non-citizens who are not lawful permanent residents (LPRs) and who have committed an aggravated felony. There were 127,376 administrative removals from FY 2003 to FY 2013 (Rosenblum and McCabe 2014, 23). Stipulated orders of removal, which must be signed by an IJ, can nonetheless occur without a hearing when a non-citizen admits the factual claims against him or her, concedes deportability or inadmissibility, and waives his or her right to appeal the removal order (ACLU 2014, 29). 7 DHS does not report on the number of stipulated removals each year. EARM s case category field includes codes that capture these distinct removal processes, the case posture and, in some instances, its disposition. Brief descriptions of the codes follow: 334 4 Immigration and Nationality Act (INA) 101(a)(13)(A). 5 INA 212; INA 237. 6 Non-citizens may challenge a determination that they are subject to reinstatement of removal in a written or oral statement to an immigration officer. 8 CFR 241.8. 7 8 CFR 1003.25(b).

Piecing Together the US Immigrant Detention Puzzle [1A] Voluntary Departure Un-Expired and Un-Extended Voluntary Departure. This code covers detainees granted voluntary departure who have not overstayed the period of voluntary departure. Case category code [1B] covers persons whose voluntary departure period has been extended. Case category code [1C] covers persons who overstayed their voluntary departure period. [2A] Deportable Under Adjudication by IJ. This code covers detainees in pending deportation proceedings. [2B] Deportable Under Adjudication by BIA. This code applies to detainees who are appealing their removal orders to the Board of Immigration Appeals (BIA). [3] Deportable. This code covers detainees with a final order of removal, including those pursuing petitions for review in federal court. 8 The dataset lists 102 persons in this category who do not have a final order. This logical impossibility might best be explained by simple error. [5B] Removable ICE Fugitive. This code covers detainees who were ordered removed via normal removal proceedings and subsequently absconded and were rearrested. [5C] Relief Granted Withholding of Deportation/Removal. This code covers detainees granted withholding of removal. [5D] of Deportation/Removal Deferred Action Granted. This code applies to detainees granted deferred action. [5E] Relief Granted Extended Voluntary Departure. This code covers detainees granted extended voluntary departure. 9 [5F] Unable to Obtain Travel Document. This code covers detainees ordered removed for whom a travel document cannot be obtained. [8A] Excludable/Inadmissible Hearing Not Commenced. This code covers detainees who have been served with a Notice to Appear (NTA) for removal proceedings, but whose NTAs have not yet been lodged with an immigration court or whose master calendar (preliminary removal) hearings have not yet taken place. [8B] Excludable/Inadmissible Under Adjudication by an Immigration Judge. This code covers detainees whose removal proceedings are pending, but who did not lawfully enter after inspection and authorization. [8C] Excludable/Inadmissible Issued. This code covers detainees who received a final order, including those pursuing a petition for review in federal court. This category included 160 persons who did not receive a final order, which appears to be another mistake. [8D] Excludable/Inadmissible Under Adjudication by BIA. This code covers detainees whose appeals of removal orders are pending with the BIA. [8E] Inadmissible ICE Fugitive. This code covers detainees who were ordered removed based on a ground of inadmissibility, and who absconded and were rearrested. [8F] Expedited Removal. This code applies to detainees in expedited removal who are not referred for a credible fear interview, both those with and without a final order of removal. [8G] Expedited Removal Credible Fear Referral. This code applies to detainees 8 ERO headquarters tracks petitions for review, but absent a stay of removal, persons pursuing petitions for review can be removed. ERO normally contacts the local ICE Office of the Chief Counsel before removal in the case of persons who have filed a petition for review. 9 For a description of extended voluntary departure and other US temporary protection programs, see Kerwin (2014, 55). 335

Journal on Migration and Human Security in expedited removal who were referred for credible fear interviews. While cases are supposed to remain in the [8G] category until they receive political asylum or a final order of removal, 1,244 cases in this category had received a removal order. In 71 instances, there is no indication whether the detainee had been ordered removed or not. [8H] Expedited Removal Status Claim Referral. This code covers detainees who claim to have lawful immigration status and, if so, should not be in expedited removal. [8I] Inadmissible ICE Fugitive Expedited Removal. This code covers detainees in the expedited removal process who absconded after a credible fear interview. [9] VR Under Safeguards. This code applies to detainees granted voluntary departure who are being held in custody until they are repatriated. [10] Visa Waiver Deportation/Removal. This code covers detainees from visa waiver countries who waived their right to contest removal. 10 [11] Deportation/Removal. This code covers detainees subject to administrative removal because they are alleged to have committed an aggravated felony and not to be LPRs. 11 Some persons in this category may be eligible for a reasonable fear interview by an asylum officer to determine whether they will be referred to an immigration court to seek withholding of removal. 12 [12] Judicial Deportation / Removal. This code covers detainees who were ordered deported by a federal judge in criminal proceedings. 13 [13] Section 250 Removal. This code covers indigent detainees who seek to be removed at the government s expense. 14 [14] Crewman, Stowaways, S-Visa Holders, 235 (c) cases. This code covers detainees who were crewman, stowaways, criminal or terrorist informants, or persons ordered removed on security and related grounds. [16] Reinstated. This code covers persons who were ordered removed or who departed voluntarily while under a removal order, and who illegally reentered. It also applies to persons subject to reinstatement of removal who request political asylum and, as a result, are eligible for a reasonable fear screening. The remaining fields in the September 22, 2012 dataset require less explanation. They include the detainee s citizenship; Criminal Yes/No (which captures convictions, not criminal charges); Most Serious Criminal Charge (based on NCIC classifications); 15 Detention Facility ; V_Days_In_Custody from the date of the most recent apprehension; 16 Final 336 10 Under INA 217(b), to participate in the visa waiver program, a foreign national from a reciprocating or designated country must waive the right to review or appeal an inadmissibility determination or to contest removal. However, they can seek political asylum. 11 INA 238(b). 12 8 CFR 238.1(f)(3). 13 INA 238(c)(1). 14 INA 250. 15 This category does not necessarily reflect the crime which serves as the grounds for removability or the basis of detention (Kerwin and Lin 2009, 23). 16 In the past, DHS-ICE often underestimated the days in custody set forth in this field by treating a transfer as a release for statistical and reporting purposes; that is, days in custody reflected only the days in the facility in question, not continuous days in custody. This problem has reportedly been corrected.

Piecing Together the US Immigrant Detention Puzzle Order Yes/No (whether the detainee received a final removal order); and Date, (the date the removal order became administratively final). Findings Congress has mandated that DHS maintain at least 34,000 immigration detention beds each night. 17 To meet this requirement, DHS-ICE operates its own facilities, and additionally contracts to use facilities owned by: (1) the federal Bureau of Prisons; (2) for-profit prison corporations; and (3) states and localities that frequently sub-contract with for-profit prisons to operate or administer their jails and prisons. Many DHS-ICE contractual and sub-contractual agreements guarantee that private correctional agencies will be paid for a minimum number of detainees per night, no matter how many people the facilities actually hold (DWN and CCR 2015). In this way, the contracts incentivize filling detention facilities each night (ibid., 6-7). On September 22, 2012, DHS-ICE held 35,197 persons in its custody, compared to 32,000 on January 25, 2009. The number of persons in custody on September 22, 2012 exceeds DHS-ICE s average daily bed use in 2012, but tracks the overall expansion of the detention system between 2009 and 2012 (Chart 1). Chart 1. Average Daily Population of Detainees, FY 1994-2013 Sources: Siskin 2004, 2012; DHS-OIG 2014. One would expect that the process for determining whether a non-citizen should be removed would take longer on average than the act of physically removing a person ordered removed. If this were the case, then the vast majority of detainees would be in removal proceedings and a far smaller percentage would have been ordered removed. Yet on September 22, 2012, 18,470 detainees had pending cases, 14,674 had received final orders of removal, 17 Consolidated Appropriations Act, 2014, Pub. L. No. 113-76, 128 Stat. 5 (2014), 251 337

Journal on Migration and Human Security and the final order field was blank in 2,053 cases. 18 Thus, a surprisingly high percentage (44 percent) of the 33,144 detainees for whom the final order field was completed had received final orders. Of the 2,053 cases which had no final order information, 2,009 had been detained for less than six months, 35 for between six months and a year, and nine for more than one year. The January 25, 2009 dataset included 18,690 detainees in removal proceedings, 10,873 who had received a final order of removal, and 2,437 for whom this field was left blank. Of the 29,563 cases for which the final order field was completed, 37 percent had received a final order. The 2012 dataset unlike the 2009 dataset included the removal case category field. Not surprisingly, most detainees on September 22, 2012 (61 percent) were in formal removal proceedings before an IJ or had been ordered removed by an IJ. These included: 7,326 (20.8 percent) who had not been admitted and had pending exclusion proceedings; 4,795 (13.6 percent) in deportation proceedings before an IJ; 3,081 (8.8 percent) who had received a final order of removal in exclusion proceedings; 2,985 (8.5 percent) in exclusion proceedings whose hearings had not yet begun; 2,354 (6.7 percent) who had received a final order of removal in deportation proceedings; 511 (1.5 percent) with appeals in deportation cases before the BIA; and 380 (1.1 percent) with appeals in exclusion cases before the BIA. However, consistent with expanded use of informal removal processes post-iirira, 38 percent of detainees were in summary, expedited, and administrative removal processes, including: 5,866 (16.7 percent) in the reinstatement of removal process; 3,492 (9.9 percent) in expedited removal; 3,332 (9.5 percent) who were subject to expedited removal but had been referred for credible fear interviews; and 650 (1.9 percent) in the administrative removal process. In both 2009 and 2012, immigrant detainees originated from the great majority of nations in the world, although citizens of Mexico, Central American and Caribbean nations predominated. Detainees in the 2009 dataset came from 177 nations, with the largest population (37 percent) from Mexico and 28 percent from Central America. On September 22, 2012, DHS-ICE detained persons from 177 nations, with 40 percent from the Northern Triangle states of Central America and 34 percent from Mexico (Table 1). 19 The decrease in the percentage of Mexican detainees between 2009 and 2012 is consistent with the dramatic drop in arrivals to the United States from Mexico between 2000 and 2012 and the decrease in the Mexican US unauthorized population from 2010 to 2013 (Warren and Kerwin 2015, 87-89). The increase in Central American detainees from 2009 to 2012 is 338 18 The 14,674 figure includes five cases in which a final order of removal date had not been entered, but the final order code indicates that a final order had been entered. 19 The nationalities of 15 detainees were listed as unknown.

Piecing Together the US Immigrant Detention Puzzle consistent with the steep rise in the number of migrants, including unaccompanied children, fleeing violence and poverty in Northern Triangle states from 2009 to 2014 (Carlson and Gallagher 2015, 133-134; DHS-CBP 2015). Table 1. Top 10 Countries of Origin for ICE Detainees on September 22, 2012 Citizenship Number of detainees Percentage of Detainee Population (35,197) Mexico 11,885 33.8 El Salvador 5,275 15.0 Honduras 4,594 13.1 Guatemala 4,284 12.2 Jamaica 874 2.5 Dominican Republic 637 1.8 Ecuador 471 1.3 Haiti 369 1.0 India 360 1.0 Cuba 329 0.9 Source: CMS analysis of DHS-ICE dataset for September 22, 2012. The Privatization of the US Immigrant Detention System Over the last six years, the Obama administration has pursued administrative reforms of the US immigrant detention system. A centerpiece of this initiative has been strengthened oversight of DHS contract facilities. The 2009 and 2012 datasets indicate the facility in which each detainee was held. One way that DHS-ICE has attempted to tighten accountability has been to reduce the number of contract facilities. On January 25, 2009, ICE detained immigrants in 286 facilities, with 68 percent in nine states (California, Arizona, New Mexico, Texas, Arkansas, Louisiana, Alabama, Georgia, and Florida) and 47 percent in the four US states that border Mexico (Kerwin and Lin 2009, 8-9). By September 22, 2012, the number of contract facilities in use had fallen, and ICE detention populations had become even more concentrated. In particular, ICE held immigrants in 189 facilities, with 77 percent of detainees in nine states (Texas, California, Arizona, Georgia, New Jersey, Louisiana, Florida, Washington, and Pennsylvania) and 51 percent in the four states bordering Mexico. The composition of the facilities used by DHS-ICE changed little between January 25, 2009 and September 22, 2012. On January 25, 2009, ICE held 68 percent of detainees in local jails and state prisons pursuant to inter-governmental service agreements (IGSAs), 17 percent in for-profit contract detention facilities (CDFs), 10 percent in its own service processing centers (SPCs), 3 percent in shelters and other alternative facilities, and 2 percent in federal prisons (Kerwin and Lin 2009, 8). 20 20 By 2014, privately owned facilities held 19 percent of immigrant detainees (USCCB-MRS and CMS 2015, 13). 339

Journal on Migration and Human Security On September 22, 2012, ICE held nearly 71 percent of detainees in jails and prisons under IGSAs, 16.5 percent in CDFs, 10.9 percent in SPCs, 1.4 percent in federal prisons, and less than 1 percent in shelters and other alternative facilities (Figure 1). The IGSA category included 57 facilities, holding 14.8 percent of all detainees, which were subject to intergovernmental agreements (IGA) with the US Marshals Service (USMS). USMS IGA facilities are owned and operated by state and local governments or private entities under agreement or contract with USMS to house federal prisoners until they are acquitted or convicted. (GAO 2013, 10). ICE detainees at these facilities may be commingled with or held separately from federal prisoners (ibid.). Figure 1. Percentage of Detainees by Type of Facility, September 22, 2012 Source: CMS analysis of DHS-ICE dataset for September 22, 2012. Given that the CDF category does not account for the operation by for-profit corporations of state and local facilities that are under IGSAs with the federal government, that category significantly understates the degree to which the detention system has been privatized. On January 25, 2009, for example, for-profit prisons owned and/or operated 75 percent of the beds in the 17 most populated detention facilities, which held 51 percent of all detainees (Kerwin and Lin 2009, 14-15). On September 22, 2012, for-profit prisons owned and/or operated 90 percent of the beds in the 21 facilities that held 52 percent of all detainees (Appendix A; Figure 2). 21 340 21 CMS identified detention facility operators based on an ICE document that included all ICE authorized facilities as of December 2, 2013 (DHS-ICE 2013a).

Piecing Together the US Immigrant Detention Puzzle Figure 2. Presence of Private Contractors in 21 Largest Detention Facilities, September 22, 2012 Facilities by Type Privately Owned or Operated Facility Source: CMS analysis of DHS-ICE dataset for September 22, 2012. Moreover, for-profit prison agencies owned and/or operated 67 percent of all detention beds on September 22, 2012 (Figure 3). With the opening of immense new private facilities in 2014 and 2015, these numbers have likely increased since 2012. Figure 3: Incidence of Privately Owned and/or Operated Facilities, September 22, 2012 Privately Owned or Operated Facility * The others and not known category include staging facilities, hospitals, hold rooms, and medical centers, whose operators could not be determined. On September 22, 2012, 112 detainees were held in these facilities. Source: CMS analysis of DHS-ICE dataset for September 22, 2012. 341

Journal on Migration and Human Security Length of Detention Length of Detention Post-Removal Order The US Supreme Court has prohibited the indefinite detention of non-citizens who have been ordered removed. In its 2001 decision in Zadvydas v. Davis, the court held that six months after a final order of removal, the government must either release a detainee or demonstrate a significant likelihood that the detainee will be removed in the reasonably foreseeable future. 22 If the government cannot meet this burden, the detainee must be released. In 2003, the court extended this decision to inadmissible persons who are ordered removed. 23 In combination, these decisions have led to a dramatic decrease in the number of post-removal order, long-term detainees (Kerwin and Lin 2009, 18, note 36). However, these decisions do not apply to persons in pending proceedings; i.e., who have not yet been ordered removed. In a 2003 decision, the court upheld mandatory detention for what it assumed to be the brief period necessary to complete removal proceedings. 24 These decisions raise two principal empirical issues: whether DHS is removing detainees within six months of their removal orders, and whether the detention of persons with pending removal proceedings is, in fact, brief. The analysis of the September 22, 2012 dataset reveals, as did the 2009 dataset, that while in most cases detention is relatively brief, large numbers of detainees have been confined for prolonged periods on any given night, including 553 for more than six months after being ordered removed. 25 On September 22, 2012, 6,298 detainees had received a final order and had been detained since receiving the order (Table 3). These detainees had, on average, been detained for 64 days following receipt of their removal order. 26 More specifically, 5,018 detainees (80 percent) had been detained following receipt of a removal order for less than 90 days; 727 (12 percent) had been detained for between 90 days and six months; 409 (6 percent) had been detained for between six months and one year; and 144 (2 percent) had been detained for one year or more (Table 3). Of the 10,771 post-removal order detainees on January 25, 2009, 8,513 (79 percent) had been detained post-removal order for less than 90 days; 1,266 (12 percent) for 90 days to six months; 676 (6 percent) for six months to one year; and 316 (3 percent) for one year or more (Kerwin and Lin 2009, 17). 342 22 Zadvydas v. Davis, 533 US 678 (2001). 23 Clark v. Martinez, 543 US 371, 378 (2005). 24 Demore v. Kim, 538 US 510 (2003). 25 The overall number of detainees who were detained on September 22, 2012 more than six months after receiving their removal order both those continuously in detention since being ordered removed and those who had not been continuously detained since being ordered removed was 951. 26 A second group of 8,371 detainees on September 22, 2012 had been ordered removed but had not been continuously detained since receiving their removal order. This could have occurred for different reasons. Some could have been ordered removed in absentia and later been arrested and detained. Others could have been ordered removed, absconded, and subsequently been rearrested.

Piecing Together the US Immigrant Detention Puzzle Table 3. Length of Detention by Status, September 22, 2012 Pre-removal order detainees Post-removal order detainees Post-removal order detainees (not detained all the time) Total 18,470 6,298 8,371 Average detention length < 90 days 91-180 days in detention 181-365 days in detention More than one year (1) (2) (3) (4) 85 days 64 days 143 days Number of Detainees by Detention Length 13,281 (72%) 2,932 (16%) 1,534 (8%) 723 (4%) 5,018 (80%) 727 (12%) 409 (6%) 144 (2%) 3,842 (61%) 976 (15%) 811 (13%) 669 (11%) 48 days 7,231 (86%) 742 (9%) 302 (4%) 96 (1%) Note: Category (1) refers to length of detention for pre-removal detainees, (2) refers to length of detention following receipt of a removal order, and (3) includes time in detention both prior to and after receiving a removal order. Category (4) covers detainees who received final orders before the current custody period and refers to length of detention for the most recent period of custody. The data from the DHS-ICE database for September 22, 2012 did not indicate whether or not a final order had been entered for 2,053 detainees. In addition, the database was missing a final-order date for five post-removal order detainees. Source: CMS analysis of DHS-ICE dataset for September 22, 2012. Of the 18,470 detainees with pending removal cases on September 22, 2012, 13,281 (72 percent) had been detained for less than 90 days; 2,932 (16 percent) had been detained for between 90 days and six months; 1,534 (8 percent) had been detained for between six months and one year; and 723 (4 percent) had been detained for one year or more. By way of comparison, of the 18,690 detainees with pending removal cases on January 25, 2009, 13,842 (74 percent) had been detained for less than 90 days; 2,486 (13 percent) had been detained for 90 days to six months; 1,792 (10 percent) from six months to one year; and 570 (3 percent) for one year or more (Kerwin and Lin 2009, 16). The average total length of detention for post-removal order detainees on September 22, 2012 (both prior to and after receiving a removal order) was 143 days, and 1,480 who had been continuously detained after receiving a removal order had been detained for more than six months. The average total length of detention for post-removal order detainees on January 25, 2009 was 114 days and 1,792 post-removal order detainees had been detained for more than six months. 343

Journal on Migration and Human Security Seventy-one percent of the 144 detainees in custody on September 22, 2012 for more than one year after receiving a removal order came from the following nine countries: Mexico (29 cases), El Salvador (26 cases), Jamaica (15 cases), Guatemala (nine cases), India (six cases), Nigeria (six cases), Philippines (five cases), Nicaragua (three cases), and Rwanda (three cases). In comparison, more than one-half of the 316 detainees in custody for more than one year post-removal order on January 25, 2009 came from the following 10 countries: Mexico (28 cases), China (28 cases), El Salvador (19 cases), India (19 cases), Jamaica (17 cases), the Philippines (13 cases), Haiti (13 cases), Dominican Republic (12 cases), Russia (nine cases), and Kenya (eight cases). On September 22, 2012, four detention facilities held detainees for an average of one year or more. On January 25, 2009, 10 detention facilities held detainees for an average of one year or more. These 10 facilities tended to hold small number of detainees, with many being soft detention or social service centers. The 10 persons in detention for the longest periods as of September 22, 2012 had been in custody from between nearly six years to more than 17 years (Appendix B). However, some of these figures may be unreliable. In five cases, for example, the final order date came years after the detainee had been in custody, well beyond the typical removal adjudication period. 27 The dataset indicated that two long-term detainees had not received a final order of removal. In another two cases, the period in custody was shorter than the period from the final order to September 22, 2012, indicating that the detainee had been out of custody at some point after receiving the final order. In addition, Maria Sacchetti (2014) reported on a very long-term Cuban detainee who had been mistakenly excluded from the September 22, 2012 dataset. Length of Detention All Detainees The 2009 MPI analysis concluded that DHS-ICE must, in effect, operate two detention systems within one, one for short-term detainees and the other for long-term detainees (Kerwin and Lin 2009, 20). CMS s analysis of the September 22, 2012 dataset found the same dichotomy. On the one hand, 26,200 detainees (74 percent) had been held for 90 days or less on September 22, 2012, with 16,427 of them (47 percent of all detainees) held for 30 days or less, and 5,252 (15 percent) held for a week or less. 28 On the other hand, 4,179 (12 percent) had been detained for more than six months, with substantial populations detained for more than one year (1,497 or 4 percent), 314 persons (slightly less than 1 percent) for more than two years, and 112 (0.3 percent) for more than three years. 29 344 27 One possible explanation might be that the database in some of these cases included prison time served, as well as time in immigrant detention. 28 The average (mean) length of detention for September 22, 2012 was 84 days, including 85 days for pre-removal order detainees and 89 days for post-removal order detainees. The average length of detention on January 25, 2009 was 81 days for pre-removal order detainees and 114 days for post-removal order detainees. 29 By way of comparison, 4,154 persons had been detained for more than six months as of January 25, 2009 (Kerwin and Lin 2009, 19).

Piecing Together the US Immigrant Detention Puzzle Table 4. Length of Detention by Removal Case Category Field, September 22, 2012 Case Category [1A] Voluntary Departure - Un-Expired and Un-Extended Departure Period [2A] Deportable - Under Adjudication by IJ [2B] Deportable - Under Adjudication by BIA [3] Deportable - 0-90 days 91-180 days 181 days- 1 year 1-2 years 2-3 years Over 3 years Total 5 0 0 0 0 0 5 3,011 1,045 497 202 25 15 4,795 43 141 202 97 17 11 511 987 494 434 311 80 48 2,354 [5B] Removable - ICE Fugitive 41 1 0 0 0 0 42 [5C] Relief Granted - Withholding of Deportation / Removal [5D] of Deportation / Removal - Deferred Action Granted [5E] Relief Granted - Extended Voluntary Departure [5F] Unable to Obtain Travel Document [8A] Excludable Hearing Not Commenced [8B] Excludable Under Adjudication by IJ [8C] Excludable Issued [8D] Excludable Under Adjudication by BIA 2 0 3 0 0 0 5 1 2 2 0 0 0 5 1 1 0 0 0 0 2 24 7 1 1 0 1 34 2,825 137 19 4 0 0 2,985 5,602 1,089 462 154 15 4 7,326 2,047 469 333 172 36 24 3,081 44 106 145 72 9 4 380 [8E] Inadmissible - ICE Fugitive 103 0 0 0 0 0 103 [8F] Expedited Removal 3,402 75 10 3 2 0 3,492 [8G] Expedited Removal - Credible Fear Referral [8H] Expedited Removal - Status Claim Referral [8I] Inadmissible - ICE Fugitive - Expedited Removal 2,620 508 141 53 7 3 3,332 3 4 2 1 1 0 11 6 1 0 0 0 0 7 [9] VR Under Safeguards 54 9 0 0 0 0 63 [10] Visa Waiver Deportation / Removal [11] Deportation / Removal 93 15 9 8 0 0 125 536 63 39 12 0 0 650 [12] Judicial Deportation / Removal 5 0 1 0 0 0 6 [13] Section 250 Removal 1 0 0 0 0 0 1 [14] Crewmen Stowaways S-Visa Holders 235(c) Cases 6 5 3 2 0 0 16 [16] Reinstated 4,738 646 379 91 10 2 5,866 Total 26,200 4,818 2,682 1,183 202 112 35,197 Source: CMS analysis of DHS-ICE dataset for September 22, 2012. 345

Journal on Migration and Human Security Table 4 illustrates length of detention by removal case category field. It reveals that long-term detainees can be found in large numbers among deportable non-citizens, with the highest numbers and percentages in the Deportable- field. Substantial numbers of long-term detainees can also be found in the Excludable/ Inadmissible fields, which cover detainees with pending cases and those ordered removed. Smaller, but not insignificant numbers of long-term detainees were in expedited removal on September 22, 2012, but had been referred for credible fear interviews; that is, they were asylum seekers. Detainees without a criminal conviction had been detained on average for 52 days on September 22, 2012, and those with criminal convictions had been detained on average for 104 days. Non-criminal detainees had been detained an average of 65 days on January 25, 2009, compared to 121 days for detainees with criminal convictions. Criminal Convictions The ENFORCE database classifies detainees with criminal convictions (marked simply yes or no ) according to the most serious criminal charge based on NCIC s criminal offense classifications. NCIC classifies individual crimes into broad categories like dangerous drugs and traffic offenses, and lists specific offenses under these categories. Thus, the immigration category encompasses individual offenses like illegal entry, false citizenship, smuggling aliens, possession of fraudulent immigration documents, trafficking of fraudulent immigration documents, and illegal re-entry (DHS-ICE 2013b). Figure 4. Detainees with Criminal Convictions, January 25, 2009 and September 22, 2012 346 Sources: Kerwin and Lin 2009, 21; CMS analysis of DHS-ICE dataset for September 22, 2012.

Piecing Together the US Immigrant Detention Puzzle A total of 21,408 detainees (61 percent) on September 22, 2012 had previously been convicted of a crime, a substantially higher percentage than on January 25, 2009 when 13,310 (42 percent) had a criminal record (Figure 4). However, the percentage of detainees whose most serious criminal charges were traffic and immigration-related violations was far higher on September 22, 2012 than on January 25, 2009. Table 5 constitutes a listing of the most serious criminal charges by NCIC category of crime for the 21,408 immigrant detainees with convictions on September 22, 2012 and the 13,310 detainees with convictions on January 25, 2009. In order of frequency in the 2012 database, these charges include dangerous drugs (5,480 cases, representing 25.6 percent of detainees with criminal convictions), traffic offenses (4,865, 22.7 percent), assault (1,894, 8.8 percent), immigration-related violations (1,702, 8 percent), and larceny (968, 4.5 percent). In contrast, on January 25, 2009, 30.3 percent of the detainees with criminal records had been convicted of dangerous drug offenses, 13.1 percent of traffic offenses, 10 percent of assault -related offenses, 6.1 percent of immigration offenses, and 4.3 percent of larceny (Kerwin and Lin 2009, 21). Table 5. Most Common Criminal Charges by Offense Category, January 25, 2009 and September 22, 2012 Detainees by Conviction Status and Offense Category January 25, 2009 September 22, 2012 No Criminal Convictions 18,690 (58.4%) 13,789 (39.2%) With Criminal Convictions Offense Category Number of Detainees 13,310 (41.6%) 21,408 (60.8%) Percent w/ criminal convictions Rank (2012 baseline) Number of Detainees Percent w/ criminal convictions Dangerous Drugs* 4,033 30.3% 1 5,480 25.6% 1 Traffic Offenses 1,738 13.1% 2 4,865 22.7% 2 Assault 1,329 10.0% 3 1,894 8.8% 3 Immigration** 812 6.1% 4 1,702 8.0% 4 Larceny 573 4.3% 5 968 4.5% 5 Robbery 549 4.1% 6 654 3.1% 6 Sexual Assault 443 3.3% 8 522 2.4% 7 Fraudulent Activities 365 2.7% 9 487 2.3% 8 Burglary 443 3.3% 7 478 2.2% 9 Weapon Offenses 341 2.6% 10 430 2.0% 10 Public Peace 335 2.5% 11 417 1.9% 11 General Crimes*** - - - 399 1.9% 12 Sex Offenses (Not Involving Assault or Commercialized Sex) 292 2.2% 13 348 1.6% 13 Forgery 281 2.1% 14 323 1.5% 14 Rank 347

Journal on Migration and Human Security Table 5. (Continued) Most Common Criminal Charges by Offense Category, January 25, 2009 and September 22, 2012 Offense Category Number of Detainees January 25, 2009 September 22, 2012 Percent w/ criminal convictions Rank (2012 baseline) Number of Detainees Percent w/ criminal convictions Family Offenses 320 2.4% 12 296 1.4% 15 Obstructing Judiciary, Congress, Legislature 144 1.1% 19 254 1.2% 16 Obstructing the Police 147 1.1% 18 245 1.1% 17 Invasion of Privacy 136 1.0% 20 232 1.1% 18 Homicide 156 1.2% 17 202 0.9% 19 Stolen Property 163 1.2% 15 167 0.8% 20 Stolen Vehicle 163 1.2% 16 157 0.7% 21 All other categories including uncategorized 547 4.1% 888 4.1% Rank * This category includes manufacturing, distribution, sale and possession of illegal drugs. ** This category includes illegal entry and re-entry, false claims to citizenship, smuggling aliens, and possession and trafficking of fraudulent immigration documents. *** This category includes crimes against a person, property crimes, morals/decency crimes, and public order crimes. Sources: CMS analysis of DHS-ICE dataset for September 22, 2012; Kerwin and Lin 2009, 20. Table 6 lists the most serious criminal charge by individual crime on September 22, 2012 and on January 25, 2009. The percentage and number of detainees whose most serious criminal charges were driving under the influence (DUI), traffic offenses, and illegal entry were far higher on September 22, 2012 than on January 25, 2009. In the 2012 dataset, the most serious charges (by frequency) were DUIs (3,017 or 14.1 percent of those with criminal convictions), traffic offense (1,680 or 7.8 percent), illegal entry (1,393 or 6.5 percent), marijuana possession (1,089 or 5.1 percent), and cocaine possession (878 or 4.1 percent). In comparison, the most common individual offenses of detainees on January 25, 2009 were DUIs (8.6 percent), dangerous drugs (7.9 percent), assault (6.2 percent), cocaine possession (5.9 percent), marijuana possession (4.3 percent), and traffic offense (4 percent) (Kerwin and Lin 2009, 20). Illegal entry was the ninth most common offense at 3.4 percent (ibid.). Appendices C and D constitute a complete listing of detainees with criminal convictions by NCIC crime category and by individual offense on September 22, 2012 and January 25, 2009. 348

Piecing Together the US Immigrant Detention Puzzle Table 6. 10 Most Common Criminal Charges by Individual Offense for 2009 and 2012 Most Serious Criminal Charge Driving Under Influence Liquor Number of detainees January 25, 2009 September 22, 2012 Of those w/ criminal convictions Rank (2012 baseline) Number of detainees Of those w/ criminal convictions 1,141 8.6% 1 3,017 14.1% 1 Traffic Offense 534 4.0% 6 1,680 7.8% 2 Illegal Entry (8 USC 1325 only) 459 3.4% 9 1,393 6.5% 3 Marijuana Possession 577 4.3% 5 1,089 5.1% 4 Cocaine Possession 787 5.9% 4 878 4.1% 5 Assault 821 6.2% 3 875 4.1% 6 Dangerous Drugs 1,045 7.9% 2 804 3.8% 7 Larceny 471 3.5% 7 739 3.5% 8 Cocaine Sell 410 3.1% 10 547 2.6% 9 Robbery 461 3.5% 8 487 2.3% 10 Sources: Kerwin and Lin 2009, 21; CMS analysis of DHS-ICE dataset for September 22, 2012. Rank The NCIC offense code listing identifies violent crimes (SSA 2015). On September 22, 2012, nearly 84 percent of the 21,408 detainees with a criminal conviction had been convicted of non-violent crimes, with assault (by category of crime) representing 54 percent of all violent crimes for which detainees had been convicted. Less than 10 percent of all detainees on September 22, 2012 had been convicted of a violent crime. Conclusion and Recommendations The US Department of Homeland Security combined 22 previously distinct federal agencies and departments and sought to unify the homeland security response of 100 federal organizations (Bush 2012). DHS also assumed the responsibilities of the merged agencies and departments that were not directly related to securing the homeland. 30 For any federal agency, much less one of DHS s breadth and scope, this transformation occurred at lightning speed. Just nine days after the 9/11 terrorist attacks, President Bush announced the creation of a new White House Office of Homeland Security under the leadership of Pennsylvania Governor Tom Ridge. In June 2002, following a series of federal law enforcement and intelligence agency restructurings, the president proposed the creation of DHS, which he envisioned as a single, unified homeland security structure which would improve protection against today s threats and be flexible enough to help meet the unknown threats of the future (Bush 2002). On November 25, 2002, the president signed into law the Homeland Security Act of 2002, which created DHS. The agency became a live concern in January 2003. 30 Homeland Security Act of 2002, Pub. L. No. 107-296, 116 Stat. 2135 (2002), 101(b)(1)(E). 349