Published by The Lawfare Institute
in Cooperation With
In October 2018, Ana Ruth Hernandez-Lara, a non-U.S. citizen who fled El Salvador in 2013 and entered the U.S. without permission, appeared before an immigration judge at a discretionary bond hearing and requested bond pending her removal proceeding. Following Board of Immigration Appeals (BIA) procedure, the judge required her to prove that she was neither a danger to the community nor a flight risk. Finding that she failed to meet her burden, the judge denied her request. Hernandez contested the immigration judge’s burden allocation and prevailed: the United States District Court for the District of New Hampshire granted her petition for a writ of habeas corpus and ordered the judge to provide a bond hearing where the government, not Hernandez, bore the burden to prove danger or flight risk by clear and convincing evidence. At her second hearing, the judge granted bond.
On Aug. 19, the U.S. Court of Appeals for the First Circuit primarily affirmed the district court’s decision. The court held that the Fifth Amendment’s Due Process clause requires the government to provide detained noncitizens awaiting removal proceedings a bond hearing where the government must prove the noncitizen is a danger by clear and convincing evidence, or flight risk by preponderance of evidence. If the government cannot meet its burden, it must offer bond or conditional parole. The court reversed the district judge only with respect to the proper standard for proving flight risk. Significantly, the decision expands the due process rights of noncitizens by upending long-standing BIA procedure, articulated in the interim decision Matter of Adeniji, which allocated the burden to the noncitizen at discretionary bond hearings—an allocation, Judge William Kayatta wrote for the majority, that places a noncitizen against “loaded dice.” Kayatta also asserted that the decision will ameliorate the “substantial societal costs” of unnecessary detention that 20 states reported in an amicus brief.
In dissent, Judge Sandra Lynch sought to resolve the matter on statutory grounds, asserted that the BIA’s burden allocation was arbitrary and capricious, and recommended enjoining the existing BIA procedure. Had the court taken Lynch’s approach, the BIA would likely have returned to its pre-Adeniji standard—outlined in its 1976 interim decision Matter of Patel, in which the government typically bore the burden of proof. But Lynch criticized the majority’s categorical requirement for the government to bear the burden at all discretionary bond hearings. Lynch asserted that the majority “constitutionalize[s] a rule … better left to the executive and the Congress,” called on the Supreme Court to step in, and asserted that the majority’s “error … will reverberate in thousands of immigration bond proceedings.”
Scope of the Decision and Probable Impact
The court’s decision applies to noncitizens detained under 8 U.S.C. § 1226(a). Section 1226(a) provides that “an alien may be arrested and detained pending a decision on whether the alien is to be removed from the United States.” Section 1226(a) also grants discretionary authority to the attorney general, who “may continue to detain the arrested alien … [and] may release the alien on … bond … or conditional parole.” (A separate section, § 1226(c), applies to “criminal aliens” and requires mandatory, rather than discretionary, continued detention.) Section 1226(a) is silent as to whether the government or noncitizen bears the burden of proof at the bond hearing. Since the 1999 interim decision in Adeniji, the BIA has interpreted that silence to place the burden on the noncitizen, allowing the immigration judge to use factors such as the noncitizen’s length of stay, employment history, family ties, and criminal record in assessing a request for bond. The Adeniji burden allocation has now shifted: Under the court’s decision, due process requires the government to bear the burden to justify continuing to detain a noncitizen held under § 1226(a).
It is unclear whether the decision will lead to a higher number of removal proceedings that noncitizens do not attend. Statistics show that the majority of noncitizens appear for removal proceedings, particularly in asylum cases where they stand a reasonable chance of obtaining relief. Some estimates show attendance rates as high as 83 percent, though official statistics provided by the Department of Justice’s Executive Office for Immigration Review suggest a significantly lower attendance rate of approximately 62 percent between fiscal years 2011 and 2020. But even if statistics showing a strong attendance rate are accurate, a sizable number of immigrants still do not appear at removal proceedings, especially as the number of pending immigration cases nears 1.4 million.
In their analyses of the challenges and obligations of the immigration system, the judges disputed the impact of shifting the burden. Kayatta emphasized that categorically shifting the burden to the government restores a fundamental liberty interest, improves “judicial and administrative efficiency,” and has the potential to reduce the significant social and economic costs caused by detention of noncitizens by preventing “ruptures in the fabric of communal life[.]” Conversely, Lynch asserted that requiring the government to bear the burden is unsupported by the Due Process clause and “imposes additional strains” on the immigration system. For Lynch, the government’s “superior knowledge” about the implications of burden allocation “reinforces [the] conclusion that the political branches should decide such questions in the first instance.”
One impact of the First Circuit’s decision is clear: Allocating the burden to the government will allow more detained noncitizens—many of whose journeys to the United States echo Hernandez’s story—to be released on bond pending removal proceedings.
In 2013, Hernandez fled El Salvador after receiving death threats from the 18th Street Gang. She settled in Portland, worked in a recycling plant and became engaged. On Sept. 20, 2018, an immigration officer detained Hernandez pursuant to § 1226(a). At her bond hearing, Hernandez argued that she was neither a danger to the community nor a flight risk, pointing to her family members in the U.S. and engagement. The government countered with an INTERPOL Red Notice that stated Hernandez was subject to an arrest warrant in El Salvador. (A Red Notice does not provide a basis for arrest in the U.S.) Still, applying BIA regulations from Adeniji, the immigration judge placed the burden on Hernandez and found she failed to meet it—“an apt demonstration,” Kayatta wrote, “of how the burden of proof can affect immigration bond hearings.” But after the district court granted Hernandez’s habeas corpus petition and required the immigration judge to hold a bond hearing at which the government bore the burden to justify Hernandez’s detention, Hernandez was released on bond. At the time of release, she had spent more than 10 months at the Strafford County Jail.
In a separate ongoing proceeding, Hernandez Lara v. Barr, the First Circuit vacated the BIA’s decision to affirm the immigration judge’s denial of asylum to Hernandez, and remanded to the BIA for further proceedings. In that case, the court found that the immigration judge denied Hernandez her right to be represented by the counsel of her choice. Hernandez’s asylum claim, as well as her claims for withholding of removal and relief under the Convention against Torture, are ongoing.
History of Board of Immigration Appeals (BIA) Bond Proceedings
Before providing summaries of the majority and dissent, it is worth outlining one area on which the judges agree: The BIA’s problematic decision to shift the burden to the detained noncitizen at discretionary bond hearings.
In 1952, Congress passed the Immigration and Nationality Act. Section 1226(a) provides that an “alien may be arrested and detained pending a decision on whether the alien is to be removed” and grants discretion to the attorney general to detain the noncitizen or release them on “bond or conditional parole.” Section 1226(a) does not specify whether the government or noncitizen bears the burden of proof for release. In its 1976 decision in Matter of Patel, the BIA stated that it presumed liberty in most cases: “An alien generally is not and should not be detained or required to post bond except on a finding that he is a threat to the national security.” But the BIA’s long-standing approach shifted in 1996, after Congress passed the Illegal Immigration Reform and Immigrant Responsibility Act (IIRIRA). IIRIRA provided a mandatory detention provision—currently § 1226(c)—for noncitizens convicted of certain crimes. However, Congress did not alter the discretionary authority provided in § 1226(a), except by increasing the minimum bond to $1,500. After IIRIRA’s passage, § 1226(a) remained largely unchanged.
To comply with IIRIRA, the Immigration and Naturalization Service (INS) implemented new regulations concerning detention. One provision states that immigration officers could release an apprehended noncitizen provided that the noncitizen “demonstrate[s] to the satisfaction of the officer that such release would not pose a danger … [and] is likely to appear for any future proceeding.” In other words, it places the burden on the detained noncitizen, but it makes no mention of the standard for immigration judges. A separate provision provides, in contrast, that the noncitizen could seek review of their detention from an immigration judge. In a procedural shift articulated in Adeniji, the BIA imputed the standard for immigration officers to immigration judges and placed the burden of proof on the noncitizen.
Both the majority and the dissent found problematic the BIA’s decision to shift the burden to the detained noncitizen but disagree on the reasons and the proper relief.
Summary of the Majority Opinion
Writing for the majority, Kayatta began discussion of the merits by squarely addressing “[w]hether the Due Process clause of the Fifth Amendment entitles a noncitizen detained pursuant to § 1226(a) to a bond hearing at which the government bears the burden of proving by clear and convincing evidence that the noncitizen is dangerous or a flight risk.” Kayatta first weighed the factors in Mathews v. Eldridge, then addressed relevant Supreme Court precedent, and finally discussed the proper extent of the burden.
In Mathews, the Supreme Court articulated a three-factor balancing test—considering private interest affected, risk of erroneous deprivation of the interest, and the government’s interest—to evaluate procedural due process. Under Kayatta’s analysis, all three weigh for Hernandez. First, Kayatta emphasized that freedom from imprisonment is at the heart of liberty and noted that Hernandez “suffered a substantial deprivation [of liberty]” when she was detained for more than 10 months. Kayatta rebuked the government’s argument that no liberty interest is deprived because the detained can choose at any point to return to their home country. “That argument,” Kayatta wrote, “is … like telling detainees that they can help themselves by jumping from the frying pan into the fire.” The majority’s analogy here contrasts with DHS v. Thuraissigiam, where the Supreme Court held that a noncitizen has no right to habeas corpus review after expedited removal proceedings (discussed on Lawfare here and here). In that case, after Thuraissigiam’s asylum claim was denied, Thuraissigiam sought a writ of habeas corpus for a new hearing. In the court’s view, habeas is limited to release from unlawful detention, but not to “achieve an entirely different end [of] additional administrative review.” Writing for the court, Justice Samuel Alito framed Thuraissigiam’s petition as seeking “the right to enter or remain in a country or to obtain administrative review potentially leading to that result.” In contrast, Kayatta framed Hernandez’s petition in terms of release from unlawful detention—not to enter or remain in the country.
On the second factor, Kayatta found a significant “risk of error in the current burden allocation.” He emphasized the lack of safeguards, including difficulty in obtaining counsel and gathering evidence, that could jeopardize noncitizens’ ability to meet their burden. On the third factor, Kayatta asserted that the government has a valid interest in ensuring appearance at removal proceedings. However, detainment of noncitizens who do not pose a flight risk fails to advance that interest. Kayatta weighed the benefit of detention with the financial costs of removing breadwinners and parents from their homes and communities. He concluded that the public interest in placing the burden on the detainee is “uncertain … and may well be negative.”
The majority then addressed and distinguished Supreme Court precedent in three cases that limited due process rights—Demore v. Kim, Carlson v. Landon and Reno v. Flores. The distinction with Demore deserves close attention.
In Demore, the Supreme Court held that the detention provision in § 1226(c), which “mandates detention during removal proceedings for a limited class of deportable aliens—including those convicted of an aggravated felony[,]” did not violate due process. In that case, Justice William Rehnquist emphasized the deference the court has historically granted to Congress regarding immigration and argued that “this Court has firmly … endorsed the proposition that Congress may make rules as to aliens that would be unacceptable if applied to citizens.” While acknowledging that the Fifth Amendment entitles noncitizens to due process during immigration proceedings, the majority firmly asserted that “detention during deportation proceeding [is] a constitutionally valid aspect of the deportation process.” Further, the court distinguished Demore from Zadvydas v. Davis, in which the Supreme Court held that the government could not detain a noncitizen for more than a reasonable time beyond the statutorily provided 90 days after that noncitizen was ordered removed. In Zadvydas, the detained individual could not be removed because multiple countries to which he was to be relocated did not accept him. Rehnquist distinguished Zadvydas from Demore, pointing out that Zadvydas involved an “indefinite” detention, whereas Demore involved a discrete period with pending proceedings. The court thus held that § 1226(c) did not violate a due process right.
The majority’s approach in Hernandez-Lara is difficult to square with the deferential approach in Demore. In Demore, Congress had evidence showing the need for mandatory detainment, including evidence that “permitting discretionary release of aliens pending their removal hearings would lead to large numbers of deportable criminal aliens skipping their hearings and remaining at large in the United States unlawfully.” One could find similar evidence justifying detention under § 1226(a): Even the highest estimates of attendance show significant numbers of in absentia removals, or those where the noncitizen fails to attend. The need to ensure attendance at removal proceedings that justified the mandatory detention of § 1226(c) could thus be used to justify discretionary detention under § 1226(a). Further, § 1226(c) applies to criminal aliens convicted of a wide range of crimes, including drug offenses, but it does not require a finding of “dangerousness.” The Demore court’s deference to mandatory detainment of a range of criminal aliens, even those not convicted of a per se dangerous crime, might suggest that the Due Process clause does not require the government to bear the burden to show dangerousness in discretionary proceedings. While Kayatta distinguished Demore by asserting that noncitizens not convicted of a crime are entitled to a significantly greater due process right than are noncitizens convicted of a crime, this logic does not entirely square with the deferential treatment of immigration policy the court took in Demore.
The majority then turned to Carlson. In Carlson, the Supreme Court held that the INS could detain noncitizens suspected of engaging in communist activities without bail, even without proving flight risk or danger. For the court, Carlson is distinguishable: The dangerousness associated with participation in communist activities and a “legislative scheme [designed] to eradicate the evils of Communist activity” justified detention. Finally, the court distinguished Reno v. Flores. In Reno, the Supreme Court determined that a regulation denying bail to noncitizen minors not accompanied by their parents and limiting their release to “a parent or legal guardian” comported with due process, because that regulation was rationally related to the government’s interest of promoting juveniles’ welfare. For the majority, Reno is distinguishable as it did not address the government’s burden of proof to detain but rather whether the government was obligated to seek the “best interests” of the minor (it was not).
Finally, the court articulated the extent of the burden of proof: clear and convincing for danger and preponderance of evidence for flight risk. As for danger, the court emphasized that there is a risk of prejudicial error, that the government has the best access to relevant information including “numerous databases” provided by federal and state authorities, and that in analogous contexts, including involuntary civil commitment, termination of parental rights and denaturalization, the government must justify its decision by clear and convincing evidence. Conversely, for flight risk, the government must only show a preponderance of evidence. The court offered two reasons. First, the detained noncitizen possesses significant knowledge of the “most relevant factors,” such as family ties and employment record. That the detained need only “marshal evidence readily available to her” provides sufficient protection for due process. Second, the court also found an analogous situation in pretrial bail under the Bail Reform Act, where proving flight risk requires a preponderance of evidence.
Dissent Scrutinizes the Majority’s Due Process Analysis
In the dissent, Lynch steered away from the due process question and insisted that creating a categorical requirement to shift the burden contravenes what the Due Process clause requires. Lynch focused instead on why the Adeniji decision violates the Administrative Procedure Act. She offered several reasons. First, the BIA justified its shift from the Patel standard (which placed the burden on the government) by pointing to the provision that provides bond standards for immigration officers, not immigration judges. The BIA’s decision to impute standards provided for immigration officers to immigration judges, who have more knowledge and access to arguments from both sides, is thus arbitrary. Second, 8 CFR §236.1(d)(1)—which does apply to immigration judges—provides that they may “ameliorate” the bond conditions applied by the arresting officer. For the dissent, the specification that immigration judges may ameliorate the bond conditions suggests they follow a standard distinct from immigration officers. Thus, the BIA’s decision to collapse the distinction between two standards contravenes the regulatory text. Also, Lynch discerned a more elementary flaw in INS’s adoption of 8 CFR §236.1: failure to implement congressional intent. When Congress enacted a mandatory detainment provision for convicted criminals in § 1226(c), it sought to recognize the “relative dangerousness and flight risk of different classes of detainees.” INS’s failure to implement different standards for different classes of detainees was arbitrary insofar as it overlooked a critical aspect of the problem, namely, relative dangerousness. Finally, the explanation INS provided for its decision contradicted relevant evidence: The INS focused on flight risk after final orders of removal rather than prior to such orders, at which stage flight risk is significantly lower.
In summary, Lynch would vacate the district court’s judgment, enjoin the current BIA procedures, and remand to the district court, which would likely apply Patel-era procedures (not a categorical rule). While Patel places the burden on the government, it does not state a standard of proof. Under the dissent’s view, the injunctive relief would allow the BIA (or Congress) to develop a burden allocation in accordance with the dissent’s enjoinment.
Notably, neither the majority nor the dissent focused on the seemingly broad discretion that the law appears to grant to the attorney general. On a plain text reading, the section simply says that the attorney general “may continue to detain the alien” and “may release the alien on bond … or conditional parole.” As various courts have pointed out, permissive verbs such as “may” suggest discretion. Shifting the burden back to the attorney general to justify continued detention arguably contravenes the plain text of the statute, because such a shift limits the discretion provided in the section.
Four Fundamental Areas of Dispute
Four areas in dispute between the majority and the dissent—constitutional avoidance, separation of powers, circuit splits and the policy impact of the court’s opinion on the administration of U.S. immigration policy—warrant further explanation, as they are relevant to ongoing controversies over due process rights for noncitizens.
First, on the issue of constitutional avoidance, the majority and the dissent differed on whether the court should grant the relief sought on due process grounds. For the majority, avoiding the constitutional claim would “neglect a central aspect of the relief sought,” because it would enjoin the current BIA procedures without establishing the extent of the burden of proof necessary to protect the due process right. Conversely, for the dissent, avoiding the constitutional question, “even where the litigants lead with their constitutional claims,” is necessary whenever the court can resolve the matter on alternative grounds.
Second, the opinions presented differing views on separation of powers. The majority emphasized the importance of constitutional limitation even in the realm of immigration and cited to Zadvydas v. Davis. Further, the majority emphasized that the judiciary plays a particularly important constitutional role in protecting the rights of the individual. The court asserted that not only does the Due Process clause “apply to all persons,” but protecting the right of due process is particularly important for “discrete and insular minorities” who cannot vote, as the Supreme Court articulated in a footnote in Carolene Products. In contrast, the dissent emphasized the broad power Congress and the executive branch hold in matters pertaining to immigration. For example, Lynch pointed to instances in which the Supreme Court has allowed detainment of noncitizens on a categorical basis but required individualized determinations of citizens, contrasting the categorical detainment of noncitizens permitted in Carlson v. Landon with the individualized showing required for citizen criminal defendants in United States v. Salerno. Lynch argued that, because the Supreme Court has permitted categorical detainment of noncitizens, the current BIA procedure providing individualized discretionary hearings all the more so does not violate a constitutional right.
Third, there is disagreement on circuit splits. The majority asserted that its decision does not conflict with the Eighth Circuit’s decision in Ali v. Brott nor the Third Circuit’s decision in Borbot v. Warden Hudson County Correctional Facility or in German Santos v. Warden Pike County Correctional Facility. As to Ali, the majority conceded that the Ali decision “contains dicta that portends a difficult result from that reached here” but asserted that the Ali court did not reach the due process constitutional question. The majority then addressed Borbot. In Borbot, the Third Circuit held that a detainee denied release on bond pending removal was not entitled to a second bond hearing because he had been granted “meaningful process.” Under the majority’s view, the Third Circuit described the discretionary bond hearing provided pursuant to § 1226(a) as a “meaningful process” only to contrast it with § 1226(c) (providing for automatic detention for detainees who have committed certain criminal offences) but did not decide the constitutional adequacy of the bond hearing under § 1226(a). The majority then turned to German Santos, the decision that was based on the Addington v. Texas line of cases. In German Santos, the Third Circuit held that, when the government has detained a noncitizen under § 1226(c) for an unreasonable amount of time, it bears the burden of proving by clear and convincing evidence that a noncitizen is a danger or flight risk. Under the majority’s reading of German Santos, the fact that the Third Circuit required a clear and convincing standard after unreasonable detention suggests that the Third Circuit sought a heightened standard for the “meaningful process” it described in Borbot for persons detained under § 1226(a).
The dissent presented a strikingly different view of the implications of the Third Circuit’s decisions in Borbot and German Santos, concluding that the majority’s opinion creates a circuit split on “a question of national importance.” Lynch asserted that Borbot held that the Due Process clause does not require the government to bear the burden of proof in bond proceedings and thereby conflicts directly with the majority’s opinion. With respect to German Santos, Lynch asserted that, because the Third Circuit required a bond hearing for persons detained mandatorily under § 1226(c) who have been held for an unreasonable period, the Third Circuit implicitly recognized that the “burden may lie at first with the noncitizen.” Given that Lynch sees the circuits in direct conflict, she urged the Supreme Court’s review of the majority’s holding.
Fourth, the disagreements extended to policy implications. The majority argued that placing the burden of proof on the government will not significantly increase the government’s administrative responsibilities because the government will retain its current incentive to gather criminal records and relevant information on detainees. Further, in the majority’s view, the burden allocation aligns with studies showing that detainees overwhelmingly attend their hearings. A study by Eagly, Shafer and Whalley of asylum adjudication for families released from detention found that between 2001 and 2016, 96 percent of former family detainees seeking asylum attended their asylum hearings. But the Eagly data is arguably no longer relevant, as the majority of cases the study’s authors examined precede the spike in family apprehensions. Adding to the dispute over statistics, the dissent cited a separate study: a 2014 Government Accountability Office report, limited to data from 2013, that found a 77 percent appearance rate for all non-detained noncitizens at final removal hearings and a 95 percent appearance rate for noncitizens subject to enhanced monitoring.
Challenges in Determining Attendance Rates at Removal Proceedings
Different methods for determining in absentia Executive Office for Immigration Review removals likely contributed to the different statistics cited by the court. The Executive Office for Immigration Review method divides the number of in absentia removal orders by the total number of initial case completions, or the first decision made by an immigration judge. In contrast, the alternative method proposed by Eagly and Shafer “calculates the in absentia rate as a proportion of all pending and completed cases.” In doing so, their method seeks to account for pending and backlogged cases and yields a significantly higher attendance rate.
Notably, different classes of noncitizens have differing incentives to appear. Lawful permanent residents, or noncitizens “granted authorization to live and work in the United States on a permanent basis,” can lose their status if they fail to attend proceedings. Lawful permanent residents thus arguably have more to lose from nonattendance than a noncitizen in Hernandez’s position seeking asylum or other relief, who does not have a status to lose—only one to gain.
Aftermath of the First Circuit Decision
Should the Supreme Court take up the decision, it would not be its first time addressing § 1226(a). In Jennings v. Rodriguez, the court held that § 1226(a) does not require the government to provide immigrants detained under the section with bond hearings every six months at which the government justifies continued detainment. The court stated that “nothing” in the statutory text supports a requirement for periodic bond hearings. In dissent, Justice Stephen Breyer wrote, “I would find it alarming, to believe that Congress wrote these statutory words … to put thousands of individuals at risk of lengthy confinement … without hope of bail.”
On the one hand, shifting the burden of proof to the government per Hernandez-Lara could increase the number of no-shows by adding to false negatives. In other words, noncitizens not viewed as flight risks might fail to appear for a hearing, leading to an increase in nonattendance rates. But the opposite could also be true. Conditions imposed under 8 U.S.C. § 1226, such as ankle monitoring, might promote high attendance rates. Further, greater access to counsel at bond hearings statistically yields higher attendance rates. The American Immigration Council, using data from fiscal years 2008 to 2018, found that between 96 percent and 92 percent of noncitizens with counsel did attend hearings. While the data does not account for more recent migrant surges, it suggests that any immigration policy seeking to balance the necessity of removal proceeding attendance with the rights of noncitizens should centralize access to counsel.
Ultimately, as immigration cases continue to surge and President Biden seeks to craft an immigration policy, the standard of proof at bond hearings will remain a consequential issue.