The United States currently detains more migrant families than any other nation in the world. Since 2001, parents and their children have been held in five different detention facilities in New Mexico, Texas, and Pennsylvania as they seek asylum in the United States. Yet, despite the sustained presence of family detention, little is known about how detained families fare in the immigration court process and what barriers they face in pursuing their asylum claims. This report presents the results of the first-ever empirical study of the adjudication of immigration court cases of families held in detention. 

Now is a particularly crucial time to inform policy with a data-driven analysis. The practice of detaining families has been sharply criticized by academics, practitioners, federal lawmakers, bar associations, immigrant rights advocates, medical experts, and the press. In 2016, an advisory committee of independent experts appointed by the Department of Homeland Security (DHS) to evaluate family detention found that “detention is generally neither appropriate nor necessary for families.”

Despite this growing consensus that family detention is a misguided policy, the administration remains interested in maintaining and possibly expanding family detention. President Donald Trump has made clear that he intends to use “all available resources” to grow “detention capabilities and capacities” at the border with Mexico. Furthermore, in the wake of public outrage denouncing the government’s 2018 decision to separate asylum-seeking families at the southern U.S. border, the administration sought to trade family separations for increased detention of families.

The Rise of Detaining Families

Our study covers 15 years of family detention in the United States, beginning in 2001 when the government opened the first detention facility to house exclusively families. Over the course of the period studied, the practice of detaining families changed but never disappeared. Between 2001 and 2016 there were five distinct brick-and-mortar family detention facilities in operation at different times in the United States. The names and locations of these facilities, and their dates of operation as family facilities, are displayed in Figure 1.

Figure 1. The Geography of Family Detention in the United States, 2001–2016

Source: Ingrid Eagly, Steven Shafer & Jana Whalley, Detaining Families: A Study of Asylum Adjudication in Family Detention, 106 Calif. L. Rev. 785 (2018), https://doi.org/10.15779/Z38WH2DF26.

Two detention centers used to detain families during our study period—the T. Don Hutto Detention Center (“Hutto”) in Taylor, Texas, and the Artesia Family Residential Center (“Artesia”) in Artesia, New Mexico—were shut down after subjecting families to harsh conditions, including unreasonably cold rooms, substandard food, and inadequate medical care. The other three family detention facilities in the United States—the Berks Family Residential Center (“Berks”) in Leesport, Pennsylvania; the Karnes County Residential Center (“Karnes”) in Karnes City, Texas; and the South Texas Family Residential Center (“Dilley”) in Dilley, Texas—remained in operation at the end of our study period in 2016.

Over the 15 years of our study, the number of detention beds reserved exclusively for families surged (Figure 2). Family detention capacity shot up temporarily between 2006 and 2009, the years that Hutto operated as a family facility. Family detention space again increased in 2014 when DHS repurposed a federal law enforcement training center in Artesia, New Mexico, as a temporary detention camp to hold families. The most dramatic increase began in 2015 with the opening of Dilley and Karnes. By 2016, family detention centers in the United States had the capacity to hold over 3,500 children and parents each day.

Figure 2. Monthly Average Bed Capacity in U.S. Family Detention Facilities, 2001–2016

Source: Ingrid Eagly, Steven Shafer & Jana Whalley, Detaining Families: A Study of Asylum Adjudication in Family Detention, 106 Calif. L. Rev. 785 (2018), https://doi.org/10.15779/Z38WH2DF26.

The rise in family detention has far outpaced the parallel growth in detention for individuals who are not part of family units, including asylum seekers. In 2001, the year family detention began at Berks, overall federal detention capacity was set at 20,000 beds. In 2016, the final year of our study, Congress required federal authorities to keep at least 34,000 detention beds available each day. This represents an increase in general detention capacity of 70 percent. In contrast, as seen in Figure 2, family detention capacity increased during the same time period by an astronomical 3,400 percent.

This tremendous growth in detention capacity imposes significant fiscal costs on the United States. According to federal officials, the average daily cost of detention is approximately $126 per person. For family members in detention, the daily average cost is even higher: approximately $161 per person, or $644 for a family unit of four. In its fiscal year 2017 budget, the federal government dedicated $1.748 billion to run detention facilities.

Moreover, detention exacerbates the extreme hardship and suffering many children and their families experience in their home countries and throughout the journey to the United States, where they seek protection. Families have endured adverse and punitive conditions inside family detention. The facilities in this study are all locked and guarded. Detained families work for as little as $1.00 a day and must abide by strict carceral rules. The family detention facilities in our study have also been the sites of severe medical neglect and psychological trauma and have been found to violate basic standards for detaining children. Several of these facilities are managed by private prison companies, including the GEO Group and CoreCivic—corporations that have been widely criticized for operating facilities with substandard conditions and poor accountability.

The trauma suffered by asylum seekers is compounded when they are locked away in isolated detention facilities and denied access to adequate healthcare or supportive services. The fact that family detention has imposed such horrific conditions makes this study’s findings all the more important.

About the Data: Defining Immigration Proceedings

An immigration court “proceeding” consists of at least one, and often several, court hearings in front of the immigration judge. EOIR categorizes immigration court proceedings based on the type of decision made by the judge. There are nine proceeding types handled by the immigration courts: removal, credible fear review, reasonable fear review, claimed status review, asylum only, rescission, continued detention review, Nicaraguan Adjustment and Central American Relief Act (NACARA), and withholding only. These immigration proceedings are explained in the next section.

Understanding Immigration and Asylum Law

Since 1996, the term “removal proceeding” has referred to the immigration judge’s decision to exclude persons attempting to enter the United States, as well as to deport someone already in the United States. If the court finds that family members do not have the legal right to enter or remain in the country, they will be removed unless they apply for and are granted relief during the removal proceeding. One such form of relief, which is common in our family detention population, is asylum. If asylum is granted, the applicant will be allowed to remain and obtain work authorization. After a year, asylees may apply to become lawful permanent residents.

Under the immigration law in place since 1996, individuals arrested at the border may be placed into an administrative process known as “expedited removal.” Immigrants apprehended at a port of entry (including airports, sea ports, and within one hundred miles of a land border crossing) within two weeks of entry may be summarily expelled without ever being placed in formal removal proceedings in front of an immigration judge. The only way that family members placed in expedited removal can eventually see an immigration judge is by expressing a fear of returning to their country, which triggers a mandatory credible fear interview by an asylum officer. If the asylum officer finds that family members do have credible fear, they will be placed into “removal proceedings” before an immigration judge.

If, however, the asylum officer finds the family member does not have a credible fear, deportation will result unless the individual requests a “credible fear review proceeding” in front of an immigration judge. During this proceeding, the judge reviews the claim of credible fear de novo. If the judge reverses the decision and finds credible fear, then the individual is placed into a removal proceeding before an immigration judge. If the judge affirms the asylum officer’s decision, that decision is generally final.

Family members who were previously removed from the United States may face “reinstatement of removal,” an administrative procedure that reactivates the prior removal order without a right to judicial review. However, if they express a fear of returning to their countries, they must be given a “reasonable fear interview” with an asylum officer. If family members are found to have a reasonable fear, which is assessed under a higher standard than for credible fear, they will be placed directly into “withholding-only proceedings” before an immigration judge. Withholding-only proceedings are more limited than removal proceedings because the only defenses to deportation that family members may raise are (1) withholding of removal and (2) protection under the Convention Against Torture (CAT), both of which have more stringent requirements than asylum.

A family member granted withholding or CAT relief may remain in the United States, but is not granted a pathway to lawful permanent resident status (as is the case for those granted asylum). If, on the other hand, the asylum officer makes a negative determination, the family member can request a “reasonable fear review proceeding” in front of an immigration judge, in which the judge reviews the claim of reasonable fear de novo. The proceeding types just discussed are among those handled by the immigration courts.

About the Data: Identifying Family Detention Proceedings

By relying on EOIR court data, we were able to determine whether an EOIR court hearing occurred while a family member was detained at Artesia, Berks, Dilley, Hutto, or Karnes. We identified 18,378 immigration court proceedings that began between 2001 and 2016 and included at least one court hearing while the family member was held in a family detention facility. Throughout the report we refer to the 18,378 proceedings linked to family detention centers as family detention (family) proceedings. Because some families had more than one proceeding, these 18,378 proceedings correspond to 16,677 individual family members.

Importantly, our count of 16,677 family members does not reflect the total population of individuals who were held in family detention from 2001 to 2016. Because this study focuses on immigration courts, family members who were released or deported before reaching a court hearing are not part of our study.

Asylum Adjudication in Family Detention

Our analysis reveals that there were four proceeding types in family detention during our study period, all of which are associated with persecution claims: credible fear review, reasonable fear review, removal, and withholding only. Some families in our study experienced more than one type of proceeding.

Out of the 16,677 family members in our study, approximately 11 percent had a removal proceeding after a credible fear proceeding, or a withholding-only proceeding after a reasonable fear proceeding. Figure 3 breaks down the family detention proceedings that occurred between 2001 and 2016 by the four proceeding types just discussed. Removal was the most common type of proceeding in our study, constituting 74 percent of the 18,378 family detention proceedings between 2001 and 2016.

In fact, removal was the only proceeding type associated with family detention between 2001 and 2006. Other proceeding types first emerged in the data in 2006: credible fear review, reasonable fear review, and withholding only. This shift in practice reflects the Bush administration’s decision to rely on the administrative procedures of expedited removal and reinstatement of removal to speed up deportations at the border. The number of family members in credible fear proceedings rose sharply between 2014 and 2016, reflecting the Obama administration’s expansion of expedited removal at the border.

Overall, as seen in Figure 3, credible fear review proceedings accounted for 19 percent (n = 3,547) of the family detention proceedings. Reasonable fear review proceedings (n = 469) and withholding-only proceedings (n = 694) were less frequent, accounting for 6 percent of the family detention proceedings in our study.

Figure 3. Family Detention Proceedings, by Proceeding Type (2001–2016)

Source: Authors’ analysis of Executive Office for Immigration Review data, 2001-2016.

Although removal remained the most frequent proceeding type in family detention (74 percent), it was even more prevalent in detained proceedings that did not involve families (“non-family detention proceedings,” as we discuss in the next section), where 98 percent were removal proceedings. This difference reflects the relative prevalence of expedited removal and reinstatement in the family detention context. Notably, this reliance on administrative procedures contributed to the overdetention of families. As we discuss later in the report, expedited removal and reinstatement allowed authorities to hold families in detention with no right to a bond hearing, unless and until they first successfully demonstrated a fear of returning to their home country.

Our study examined family detention proceedings and found serious barriers in access to justice. Namely, we highlight concerns with family detention regarding remote location, access to representation, and prolonged detention.

About the Data: Measuring Representation

The immigration court’s rules require the filing of the EOIR-28 form by all attorneys and certified representatives who appear in immigration court. We measured representation based on whether an attorney filed an EOIR-28 form with the immigration court at any point prior to the conclusion of the relevant court proceeding. We also counted as represented any respondent for whom an EOIR-28 form was filed after the conclusion of the relevant proceeding, so long as court records showed that an attorney appeared on the respondent’s behalf in at least one hearing during any proceeding.

Although we refer to individuals providing representation to immigrants in this study as “attorneys,” representation can also be provided by “accredited representatives,” non-attorneys working for nonprofit organizations who are certified to appear in immigration court. 

Access to Representation and a Fair Hearing Is Limited in Detention

A threshold issue of access to justice for detained families is the remote location of their imprisonment. All five family detention centers used from 2001 to 2016 were located in small or rural cities, far from the nearest immigration courts, nonprofit organizations, social services, and pro bono attorneys. 

Due to the remote location of detention facilities throughout the United States, immigration judges increasingly hear cases over a video connection without ever traveling to the detention center. Analysis of our family detention proceedings reveals that an overwhelming 93 percent of family detention hearings were handled via televideo. The heavy reliance on televideo for an entire group of litigants is concerning, especially given that televideo is associated with reduced engagement by respondents in their court proceedings, including being less likely to find an attorney and less likely to seek relief, when compared to similarly situated respondents who had their hearings in person.

Expedited Proceedings Further Limit Access to Representation

Having a lawyer is associated with better outcomes at every stage in the immigration court process. Attorney representation is particularly vital to ensuring a fair court process for parents and children who have endured violence in their countries and during their journeys. Asylum cases are particularly complex, making attorneys all the more critical for marshaling the necessary proof. A DHS advisory committee studying family detention recommended in 2016 that the government provide counsel to these families.

The devastating conditions of family detention have encouraged pro bono attorneys to travel long distances to detention centers and offer families legal representation. Across all five facilities studied, volunteer attorneys mobilized in significant ways to provide free legal assistance and expand access to counsel for families held in detention. Yet, despite these volunteer efforts, many family members did not have representation at their hearings.

About the Data: Family Detention Proceedings vs. Non-Family Detention Proceedings

To better understand what is unique about court adjudication of family detention cases, we compared a range of outcomes in family detention proceedings to outcomes in detention proceedings that did not involve families, what we call non-family detention proceedings. We define non-family detention proceedings as including all persons detained at some point in their EOIR case adjudication who were not associated with one of the five family detention centers. In total, we find there were 2,807,814 non-family detention proceedings during the 15-year study period.

The EOIR data allow us to analyze whether family members who began their court cases in detention obtained counsel. During initial proceedings for the 16,677 family members in our study, representation rates diverged based on proceeding type. As seen in Figure 4, in initial credible fear or reasonable fear review proceedings (n = 3,943), only 23 percent of family members were represented. In contrast, in initial removal or withholding-only proceedings (n = 12,734), 34 percent of family members had representation. Across all proceeding types, only 32 percent of family members  had representation at their initial proceeding.

The lower rate of representation in credible and reasonable fear proceedings could be attributed to factors beyond a family member’s control. Court rules, for example, limit the role of attorneys in these proceedings. These restrictions may make some attorneys less inclined to involve themselves at this early stage, or to do so without filing an EOIR-28 form.  

The expedited nature of credible and reasonable fear proceedings may also result in a lower rate of attorney involvement. These proceedings can occur very quickly after a negative finding by the asylum officer and are generally resolved in a single hearing, as we later discuss (Table 1). The speed of the process means that families have less time to secure counsel. In contrast, removal or withholding-only proceedings take much longer to complete and generally involve more than one hearing. This gives family members more time to find counsel.

Figure 4. Representation in Initial Proceedings, by Proceeding Type

Source: Authors’ analysis of Executive Office for Immigration Review data, 2001-2016.

Representation Increases for Families Released from Detention

We next analyzed the percentage of family members that found counsel in their most recent removal or withholding-only proceeding. Because representation rates can vary based on detention status, we separately analyzed those family members who remained detained and those who were released at the most recent proceeding in their case history.

As Figure 5 shows, of those family members who remained detained (n = 1,614), 53 percent were represented by counsel at their most recent merits proceeding. This is more than double the representation rate of 20 percent for individuals held in non-family detention who were detained at their most recent proceeding. At the same time, 47 percent of family members who remained detained during our study never found counsel.

Family members who were released were far more likely to find counsel. Of released family members (n = 13,037), 76 percent were represented at their most merits proceeding. This rate is almost identical to the 77 percent representation rate for respondents released from non-family detention (Figure 5). This finding is particularly significant given that more individuals held in family detention (compared to those in non-family detention) were released from custody by their final removal or withholding-only proceeding: 89 percent of individuals in family detention were released, compared to only 30 percent of non-family members.

Figure 5. Representation at Most Recent Removal or Withholding-Only Proceeding, by Detention Status

Source: Authors’ analysis of Executive Office for Immigration Review data, 2001-2016.

 The fact that more family detainees, compared to non-family detainees, were able to find counsel is a testament to the robust pro bono and nonprofit networks that have mobilized to provide legal services to these families. Countless volunteers and active pro bono programs have worked tirelessly to provide families with free legal assistance. Attorneys practicing at private law firms, including solo practitioners, represented the lion’s share of individuals who found counsel in family detention (82 percent). Law school clinics (5 percent) and nonprofit organizations (13 percent) provided the remaining 18 percent of legal representation.

Families Have Been Detained for Prolonged Periods

A major issue in the detention debate is the exposure of families to long-term detention. Critics condemn the detention of asylum seekers and their children as harmful to the physical and mental health of these families. In addition, long-term detention raises due process concerns. In Zadvydas v. Davis, the Supreme Court found that a detention of six months was presumptively reasonable but warned that detention could not exceed a period “reasonably necessary to secure removal.” 

Table 1 presents our analysis of the detention times associated with completed EOIR court proceedings for family members who remained in detention throughout their court proceedings. For the 5,419 proceedings involving families that were completed in detention, the median period of detention associated with credible fear and reasonable fear proceedings was one day (with an average of three and four days, respectively). Given that these proceedings are expedited by design, the short period is unsurprising. 

Detained removal and withholding-only proceedings took considerably longer. The average duration for detained removal proceedings was 87 days (median of 29 days), whereas detained withholding-only proceedings had an average duration of 132 days (median of 78 days) (Table 1). 

Table 1. Duration of Proceedings for Family Members that Remained Detained, by Proceeding Type 

 Source: Authors’ analysis of Executive Office for Immigration Review data, 2001-2016.

We also analyzed the total time in detention across the multiple proceeding types of family members (e.g., credible fear review and removal). Among the 10,122 family members whose cases reached completion during our study period, 34 percent of them were detained throughout their entire case. We studied the duration of detention among these family members, finding 800 completed cases that involved detention of a month or more. We found that 397 family members were detained for more than three months and 115 were held for more than six months.   

Importantly, these measurements do not include the amount of time families spent in custody prior to the initiation of their court proceeding with the immigration judge. ICE reports that the average length of detention for the credible fear interview screening process is 58 days. This measurement does not include the length of time that families may spend in Customs and Border Patrol (CBP) holding cells before being transferred to ICE custody to begin the credible or reasonable fear screening interview process. 

Experts on family detention have found that in some cases families can spend as long as six months in ICE detention before the EOIR court process begins. The federal government recently admitted as much, stating that “[i]n some instances, there is a significant period of time between issuance of the charging document by DHS, and filing the charging document with EOIR.” Moreover, our measurements do not include any additional detention time associated with appealing family detention cases. Notably, several families were detained at Berks for a year-and-a-half while lawyers sought habeas review of their credible fear process in federal court. 

Furthermore, families can continue to be held in detention after the EOIR proceedings are over until DHS deports them. By law, DHS has 90 days to remove individuals from the United States after the immigration court issues a final order of removal, and sometimes this removal period can take significantly longer. This additional period of post-order detention further prolongs the detention time for families.

Table 3. Outcomes in Pending and Completed Released Family Detention Cases, by Volume and Outcomes (2001–2016)

Source: Authors’ analysis of Executive Office for Immigration Review data, 2001-2016.
*Includes cases ended through termination, prosecutorial discretion, or administrative closure. **Includes voluntary departures. 

A major issue in the family detention debate is the unnecessary placement of families in detention to serve the government’s goal of deterring other families from seeking asylum. Critics argue that family detention now reaches beyond the purported civil organizing principles and instead functions as a punitive system of control. In 2014, federal officials publicly confirmed their use of family detention as a system of control when they stated that they hoped the practice of detaining mothers and their children would deter other families from making the trip to the United States. Although a deterrence strategy might be permissible in the criminal justice system, courts have made clear that it may not motivate the civil immigration system. The established law provides that in the absence of a showing of flight risk or danger to the community, civil detainees should be released.

The growing detention of families seeking asylum also raises serious issues given legal restrictions on the detention of bona fide asylum seekers. Article 31 of the 1951 Refugee Convention provides that countries may restrict the movement of refugees only when necessary—a standard that does not allow for detention to punish border crossers or to achieve general deterrence of asylum seekers. In addition, Article 9 of the International Covenant on Civil and Political Rights guarantees migrants freedom from arbitrary detention. The United Nations Committee on the Rights of the Child and the United Nations Committee on the Protection of the Rights of All Migrant Workers and Members of Their Families have issued joint guidance clarifying that holding children in detention, either alone or with their families, is never in the best interest of the child. Finally, United States law also severely restricts the detention of migrant children.

Despite these standards favoring the release of children and asylum seekers, the family members in our study have all been subjected to detention. In many cases, families have been subject to overdetention, which we define as their continued detention despite no finding that they pose a danger or flight risk. Central American families have increasingly been targeted for confinement in family detention centers. As a result, these families have turned to the immigration courts to reverse the decisions of DHS officials that prolong their detention.

Negative Credible and Reasonable Fear Decisions for Families Are Often Reversed

For families placed into expedited removal or reinstatement of removal, the only route to obtaining a court hearing is to express a fear of returning to their home country. If the asylum officer does not find that fear is established, the family’s only option is to request a review of the decision by an immigration judge. Families can be held in detention during the entire time of the credible fear or reasonable fear review in immigration court. 

Our study finds that immigration judges frequently overturned agency decisions on credible fear for family detainees, highlighting the existing concerns about summary removal processes. Moreover, the rate of reversal rose throughout our study period. This raises serious concerns about the possible exposure of family members to overdetention. Erroneous agency decisions at the credible fear or reasonable fear stage lead to dire consequences for detained families. 

The EOIR data allow us to analyze EOIR judge decisions in credible fear and reasonable fear review proceedings (n = 4,016). Overall, as Figure 8 highlights, immigration judges vacated 48 percent of negative credible fear findings. This rate is three times higher than the 16 percent reversal rate for detained cases not involving families. 

An even more striking reversal rate occurred in reasonable fear decisions. Across all five family detention centers, immigration judges overturned 58 percent of the negative reasonable fear findings of the asylum officers. In contrast, the reversal rate for non-family reasonable fear proceedings was only 15 percent. In other words, the reversal rate for reasonable fear decisions was almost four times higher in family decisions than in non-family decisions. The family detention reversal rate in credible and reasonable fear proceedings is particularly remarkable given that family and non-family detainees both obtained lawyers at an identical rate (23 percent) during such initial proceedings (see Figure 4).

Figure 8. Rate of Reversal of USCIS Negative Credible Fear and Reasonable Fear Decisions

Source: Ingrid Eagly, Steven Shafer & Jana Whalley, Detaining Families: A Study of Asylum Adjudication in Family Detention, 106 Calif. L. Rev. 785 (2018), https://doi.org/10.15779/Z38WH2DF26.

 Figure 9 tracks these reversals of credible and reasonable fear denials over the 15-year period of family detention. It reveals that the reversal rate for family detention cases has risen dramatically since 2007 when credible fear and reasonable fear proceedings first appeared in the family detention context. By 2016, 58 percent of appealed denials were reversed by immigration judges. Figure 9 compares this rate of reversal for family detainees to that of non-family detainees, who had a much lower and relatively stable reversal rate.

Figure 9. Rate of Reversal of USCIS Negative Credible Fear and Reasonable Fear Decisions, 2001–2016

Source: Authors’ analysis of Executive Office for Immigration Review data, 2001-2016.

 The credible fear and reasonable fear processes are intended to prevent the United States from erroneously returning bona fide asylum seekers to their home countries. However, these screening tools are part of summary removal processes designed to rapidly push individuals through—or skip entirely—due process protections. Legal service providers have documented situations of adverse decisions resulting from poor translation services, time or resource constraints, and interference from the very officers conducting interviews.

Family members who were eligible for release during our study period remained unnecessarily detained while they pursued review before immigration judges. Although the average court time for credible fear and reasonable fear proceedings was only three to four days, erroneous agency decisions denying credible or reasonable fear are also associated with other delays, such as the time it takes for the family to file for review with the EOIR and obtain notice of the hearing, as well as any delays in release following the judge’s reversal.

DHS’s Decisions to Detain Families Are Systematically Challenged and Reversed

All family members going through the credible and reasonable fear screening process are held in detention without release. Those who are not found to have a credible or reasonable fear are ordered removed. However, those found to have a credible or reasonable fear of persecution are placed into removal or withholding-only proceedings. At this point, ICE has the discretion to grant their release on parole. However, if immigration authorities deny parole, or require posting of an unaffordable bond amount, the family will remain detained unless an immigration judge orders their release at a bond hearing (also known as a custody hearing).

To identify family members who were eligible for such a hearing—but were not released by ICE—we analyzed the prevalence of bond hearings among family members in removal proceedings during our study period (completed and pending cases). Our analysis reveals that immigration courts have been called on to intervene in the detention of migrant families. Families held in the five U.S. family detention centers have been systematically more likely to request bond hearings—and to win release on bond—than other detained migrants. These and other findings suggest that agency officials subjected families to detention despite the fact that these families were ultimately found to present a low security risk and to be legally eligible for release.

Immigration Courts Regularly Ordered the Release of Detained Families

As reported in Figure 10, we find that 59 percent of the 13,668 detained family members placed in removal proceedings had at least one bond hearing, versus only 25 percent of the 2,755,862 non-family detainees in removal proceedings. In other words, when compared to the rest of the detained population, family members were more likely to call on immigration judges to secure their release from detention. Figure 10 also shows that these disparities between family and non-family bond hearing rates persist when we controlled for representation by counsel. These disparate rates for family and non-family bond hearings are an important indicator of the overdetention of families.

Figure 10. Prevalence of Bond Hearings in Removal Proceedings, by Family or Non-Family Status and Representation Status

Source: Authors’ analysis of Executive Office for Immigration Review data, 2001-2016.

The data also allow us to analyze the decisions made by immigration judges at bond hearings. Figure 11 presents the proportion of detained family members with bond hearings who had a successful outcome—meaning that the judge ordered release on a cash bond or on recognizance. Overall, 57 percent of detained family members with bond hearings had a successful outcome: 19 percent were released on their own recognizance and an additional 38 percent were granted a money bond. This success rate is higher and more favorable than in non-family detention cases with bond hearings, where only 1 percent were released on recognizance and 44 percent received a cash bond. The rates of successful outcomes also varied by detention center, with Karnes and Dilley both enjoying an overall success rate above 80 percent.

Figure 11. Successful Outcomes in Bond Hearings, by Decision Type

Source: Authors’ analysis of Executive Office for Immigration Review data, 2001-2016.

Government Regularly Refused to Set Reasonable Bond for Eligible Families

We further examined the decisions immigration judges issued at bond hearings in family and non-family detention, focusing on bond amounts. Our findings raise serious questions about whether DHS subjects family detainees to overdetention.

Under immigration law, bond amounts reflect the judicial assessment of danger and flight risk. As shown in Table 4, we find that immigration judges systematically set family detention bonds at a lower amount than for non-family detention cases. This suggests that, as a group, immigration judges perceive family detainees as presenting less of a flight risk or danger than non-family detainees.

Over the 15 years of our study, the average bond amount set by immigration judges in family detention was $3,226 and the median was $2,000 (Table 4). Notably, these bond amounts were much lower than in the non-family detention context, where the average bond amount over the same period was $11,829 and the median was $5,000. A $0 bond amount reflects the judge’s decision to release an individual without requiring a money bond.  

Table 4. Bond Amounts Issued by Immigration Judges, Overall and by Family Detention Facility

Immigration judges also had an important role lowering unreasonably high bond amounts for families previously set by DHS officials. Lawyers representing detained families reported that DHS refused to set bonds—or set prohibitively expensive bonds—to deter others from coming to the United States during the 2014 increase of Central American families seeking protection in the United States.

Using the EOIR data, we investigated these “no bond” and “high bond” policies, focusing on bond decisions at Artesia, Karnes, and Dilley. As seen on the right side of Table 5, within these three detention centers there were a total of 3,118 court bond hearings in which the court granted bond or release on recognizance. Of those, only 524 (17 percent) had a bond set by DHS at the time of the immigration court bond hearing. In other words, at the time of the bond hearing before the immigration court, DHS still argued in favor of detention with “no bond” in 83 percent of these cases. The judicial reversals of DHS’s “no bond” decisions reveal that immigration judges play an important role in tempering agency decisions that would otherwise subject families to overdetention.

Table 5 also reveals ICE’s general pattern of setting prohibitively high bonds. In the handful of cases in which ICE did set a bond, the average amount was high: $7,500 in Karnes; $6,180 in Dilley; and $5,600 in Artesia. After the immigration judge reviewed the case, these average bond amounts decreased significantly—between 32 and 70 percent lower (in Karnes and Artesia, respectively). In Dilley, where ICE more commonly chose to set bond amounts, the average bond amount dropped by 40 percent.

Table 5. DHS vs. Judge Bond Decisions in Removal Proceedings with Bond Hearings, by Detention Facility

Overdetention Disproportionately Affected Central American Families

The EOIR data raise important questions as to whether Central Americans have been disproportionately subjected to family detention. We find that 80 percent of individuals in family detention proceedings over the 15-year study period were Central Americans from El Salvador (34 percent), Honduras (27 percent), and Guatemala (19 percent). The remaining 21 percent came from one of 28 different countries, with individuals from Mexico (6 percent) and China (2 percent) representing the largest shares.

Importantly, as displayed in Figure 12, we find that the national origin of family detainees has radically shifted over time. In the first five years of family detention, family detainees came primarily from the group of 28 different countries, with very few from the Central American Northern Triangle countries of El Salvador, Honduras, and Guatemala. Over time, the proportion of family detainees from the Northern Triangle has skyrocketed, reaching a high of 94 percent of family detainees in 2016. At the same time, the proportion of detained families from other countries has plummeted.

Our research reveals that the increase of Central Americans in family detention cases has far outpaced their presence in non-family detention cases. Figure 12 compares the percent of family detention proceedings associated with the Northern Triangle to non-family detention proceedings over the same time period. Although the proportion of Northern Triangle nationalities in non-family detention proceedings doubled during the relevant time period, it increased by thirteen times in family detention.

Figure 12. Individuals from the Northern Triangle Held in Family and Non-Family Detention, 2001–2016

Source: Authors’ analysis of Executive Office for Immigration Review data, 2001-2016.

These striking patterns raise questions about why Central American families are so heavily represented in family detention. To be sure, the increase in Central American migration is in part due to the extreme levels of violence in El Salvador, Honduras, and Guatemala. These three countries have some of the highest murder rates in the world. Central American women and children have been especially vulnerable to gang violence, domestic abuse, and sexual abuse. However, Central American families have been detained to the near exclusion of families of other nationalities and at levels that are disproportionate to their presence in non-family detention during the same time period. These facts suggest that Central American families have been subjected to overdetention.

Relying on data obtained through public records requests, this report reveals an expanding system of detention facilities that imprisons families seeking asylum, sometimes for prolonged periods. We also document some of the serious challenges that families face in pursuing their asylum claims inside family detention. The merger of detention and adjudication is understudied, yet more important than ever given the Trump administration’s explicit enforcement plans to expand expedited removal, heighten standards for asylum claims, and maintain all migrants in detention throughout the adjudication process. 

Our study provides empirical support for a different set of policy decisions. First, we recommend that authorities place families directly into proceedings before an immigration judge rather than first subjecting them to an administrative process. Federal immigration authorities have long had the discretion to place families directly into removal proceedings rather than rely on the expedited removal process. Doing so would allow immigration judges to be involved in ensuring due process from the outset of the case. As our report highlights, the expedited removal process limits the ability of individuals to fully pursue their claims with the procedural and substantive protections available in regular removal proceedings before an immigration judge. 

Eliminating summary removal processes, such as expedited removal, for asylum seekers and vulnerable populations would also reduce overreliance on detention, since families in removal proceedings can generally be released immediately pending a hearing. Alternatively, if it is not feasible to immediately release families following their apprehension, families in removal proceedings should be released as soon as is practicable into a less restrictive custody setting or community-based alternative to detention. We establish a high validity rate for the asylum claims of family members in our study, as evidenced by successful case outcomes including judicial decisions allowing released family detainees to remain in the United States. This should reassure policymakers that moving away from reliance on expedited removal and detention is a sensible policy choice. In addition, our evidence reveals that family members released from detention are unlikely to abscond. Indeed, released family members in completed and pending cases who filed claims for asylum achieved a 96 percent appearance rate at their future hearings. 

Second, we recommend funding for court-appointed counsel. At a minimum, the government should provide attorneys for immigrants who are especially vulnerable, including detained families and individuals, and generally should not move forward with a case until counsel may be obtained. Despite intensive pro bono efforts, we find that access to counsel remains a pressing issue for families in detention. The facilities detaining parents and children are located in remote areas, far away from city centers. A shocking 68 percent of parents and children in our study remained unrepresented in their initial family detention proceeding. Those who continued beyond this stage were more likely to obtain counsel, but still, nearly half of those who remained detained went without lawyers. These findings are troubling given that representation in immigration court is strongly correlated with better outcomes, and success in these cases can mean the difference between life and death. Short of funding for appointed counsel, addressing the dearth of counsel would require expanding nonprofit and law school clinic resources, which currently are unable to handle the high volume of family detention cases. It would also require recruiting more pro bono volunteers and providing them with the support and training necessary to take on this work. 

Third, our study highlights the need for increased training and monitoring of immigration judges. Although EOIR has taken steps in the past to identify those judges with unusually high or low grant rates, a more comprehensive and independent review is necessary. We find that judges in different jurisdictions treat the cases of released family detainees quite differently. In addition, there is surprising jurisdictional variation in the rates of legal representation and applications for relief. These sorts of patterns are likely to intensify as the Trump administration hires new immigration judges that lack judicial experience, increases judicial caseloads, and cuts funding to train existing immigration judges. Future research should examine indicators of judicial decision-making beyond grant rate, including steps judges take to facilitate the filing of applications for relief, to notify respondents of their hearings, and to permit sufficient time to find competent counsel. Procedures for hearing bond claims should also be included among priority issues for training and standardization. 

Fourth and finally, our report underscores the vital need to maintain independence of the immigration courts. In particular, our study documents how immigration officials have delayed the timely release of families from detention and erroneously denied their claims of persecution at the asylum office’s screening stage. We find, however, that immigration courts have played an important due process role in reviewing these decisions and charting a different course. Over the 15 years of our study, immigration judges reversed negative credible fear findings 48 percent of the time and reversed negative reasonable fear findings 58 percent of the time. In the current era of increased immigration enforcement, the independence of immigration courts must be protected in order to preserve their vital function in reviewing the administrative decisions of ICE and CBP authorities and asylum officers. It is also crucial that immigration judges are given sufficient time to decide their cases, without the imposition of numerical quotas for case completion. 

In conclusion, although the United States has detained families in prison-like facilities since 2001, this study is the first to empirically analyze how these families fared in the immigration court process. We identify multiple barriers that families experience in pursuing asylum and highlight the underappreciated role that immigration courts have played in securing their release from custody and reviewing the merits of their claims. These and other findings are meaningful to current policy debates regarding the role of immigration courts in maintaining due process in the asylum process and the appropriate use of detention to manage the migration of families fleeing violence in their home countries.