The Finality of Final Orders of Removal

I.  The Convention against Torture in the US Immigration System

This Part provides an overview of the key statutes and regulations governing CAT relief and federal jurisdiction to review CAT claims. Part I.A gives an overview of domestic removal proceedings, explaining how an individual comes before the Immigration Court and raises a claim for CAT relief. Part I.B explores the details of the CAT itself, including its international background and domestic implementation. Finally, Part I.C details the domestic statutes governing appellate court review of immigration proceedings.

A.    Removal Proceedings and Raising a Claim for CAT Protection

Before delving into the substantive law at issue, it is necessary to understand the system for immigration enforcement and removal. Ultimately, it is against this backdrop that the particular issue confronted by this Comment arises.

1.   The steps of a removal action.

When the US government decides to expel an immigrant, it initiates removal proceedings by issuing a Notice to Appear (NTA).22 The NTA operates as a subpoena, requiring the immigrant to appear in court at a certain day and time.23 IJs, who “determine removability and adjudicate applications for relief from removal,”24 oversee the hearings.25 IJ decisions are subject to review by the BIA.26 It is only after a BIA decision is issued that an immigrant may appeal to a federal circuit court.27

Any noncitizen is subject to removal if he or she falls within any of the specific categories laid out in 8 USC § 1227.28 Most relevant to this Comment, noncitizens convicted of certain criminal offenses are removable.29 Criminal offenses that can subject a noncitizen to removal proceedings include crimes of moral turpitude,30 aggravated felonies,31 many controlled-substance violations,32 firearm violations,33 crimes of domestic violence,34 and human trafficking.35 A noncitizen with criminal convictions may still seek relief from removal by applying for forms of discretionary or mandatory relief.36

For CAT claims, the relevant statutes and regulations place the burden of proof on the applicant at all times.37 Many CAT cases turn on factual details and evidentiary burdens, rather than legal issues, which is why questions of fact are so critical in removal proceedings. Consider, for example, the case of Kevin Wanjiru. Wanjiru sought CAT relief, fearing that a criminal gang in Kenya (the Mungiki) would torture and kill him upon return.38 The BIA’s denial of relief turned primarily on factual deficiencies, namely that Wanjiru did not establish that the Mungiki would recognize him as a former member and seek him out for torture or murder.39 The Seventh Circuit, in contrast, reviewed the factual question and found that Wanjiru’s presentation of personal testimony about the likelihood of torture and of documentary evidence about the Mungiki’s general acts of violence toward former members was sufficient to meet the burden of proof.40 As demonstrated by this dispute between the BIA and the Seventh Circuit, CAT claims often turn on the interpretation of facts, not of law. Applicants like Wanjiru are often unable to marshal large amounts of evidence, as many fled their home countries quickly and are from countries that do not have reliable records (for either political or technological reasons).41 In the absence of other evidence, IJs must instead base their decisions on a determination of the applicant’s credibility.42 This system of scant records, dispositive credibility judgments, and lack of judicial review produces inconsistent results. One study, for example, “found that even among immigration judges in the same jurisdiction, hearing cases of asylum seekers from the same country, the disparity rate [in granting petitions] was considerable.”43 Among New York City IJs hearing asylum cases for Chinese applicants, for example, the rate of denial ranged from a high of 94.5 percent to a low of 6.9 percent.44 Judicial review over questions of fact will alleviate the arbitrary way in which questions of fact are decided.45

2.   CAT relief.

Immigration relief can largely be put into two categories: asylum and CAT protection. CAT relief is governed separately from asylum relief. In one sense, CAT relief is broader than asylum because it can apply to anyone, not just an individual who meets the requirements of 8 USC § 1158(b)(1)(B)(i). While asylum requires that the persecution be caused by statutorily specified motivating factors,46 the CAT does not require any particular reason for the torture.47 On the other hand, “torture” requires a higher level of harm than “persecution,” and CAT relief may thus be a form of relief open to fewer claimants.48

CAT relief is available in two very similar forms, withholding of removal49 and deferral of removal.50 Withholding of removal will be granted if the applicant can prove it is “more likely than not that he or she would be tortured if removed to the proposed country of removal.”51 Withholding of removal must be denied, however, if the applicant has been convicted of a serious crime.52 For such an applicant, then, deferral of removal is the only avenue for CAT relief. Deferral of removal is granted for the same reasons as withholding of removal, but it has no criminal conviction bar.53

Deferral of removal is distinguished from withholding of removal based on the level of protection received. In both withholding of removal and deferral of removal, the recipient can be “remov[ed] . . . to a third country other than the country to which removal has been withheld or deferred.”54 Either form of temporary relief can also be terminated if it is shown there is no longer a likelihood of torture.55 An individual with deferred removal, however, can be held in Immigration and Naturalization Service detention centers for the duration of the time he remains in the United States.56 Thus, these individuals will not be returned to a country where they will be tortured, but they may not be allowed to be free in the United States. The process for terminating deferral of removal is also simpler and faster than for terminating withholding of removal. Deferral of removal, therefore, offers a less secure form of protection; due to the criminal conviction bar on withholding of removal, however, it is the only form of relief available to most criminal noncitizens fearing torture.

B.    The Convention against Torture

Although it provides no enforceable rights on its own,58 the CAT serves as the basis for understanding this area of law. The United Nations General Assembly adopted the CAT on December 10, 1984, and it entered into force on June 26, 1987.59 Article 3’s so-called nonrefoulement provision provides that “[n]o State Party shall expel, return (refouler) or extradite a person to another State where there are substantial grounds for believing that he would be in danger of being subjected to torture.”60 It further specifies the means by which states are to determine whether an individual is likely to be subject to torture if returned to his or her home country. The CAT directs countries to “take into account all relevant considerations including, where applicable, the existence in the State concerned of a consistent pattern of gross, flagrant or mass violations of human rights.”61

In ratifying the CAT, the United States expressly made it a non-self-executing treaty.63 A treaty that is not self-executing requires Congress to pass legislation specifically implementing the treaty provisions into US law.64 The US Senate, while considering the ratification of the CAT, attached reservations, understandings, and declarations (RUDs) to the United States’ adoption of the CAT.65 The United States attached an understanding requiring individuals seeking CAT protection to demonstrate a “more likely than not” certainty of torture.66 This imposes a higher burden of proof than the CAT itself, which requires only a “substantial” likelihood of torture, as something could be “substantial” but still be less than 51 percent likely.67 This higher burden of proof might point to a greater need for appellate review. Given the difficulty applicants often have in marshaling factual proof, this burden can be extraordinarily difficult to meet. As such, it becomes more critical to ensure that the facts that have been presented are properly considered and weighed.

C.    Jurisdiction and Appellate Review of CAT Claims

The current framework governing appellate jurisdiction over CAT claims in federal courts was not enacted in one fell swoop. Instead, it developed piecemeal over the span of a decade.

The Illegal Immigration Reform and Immigrant Responsibility Act of 199668 (IIRIRA) included a jurisdiction-stripping provision. Codified at 8 USC § 1252(a)(2)(C), the IIRIRA’s jurisdictional bar provides that “[n]otwithstanding any other provision of law . . . no court shall have jurisdiction to review any final order of removal against an alien who is removable by reason of having committed a criminal offense covered in [specified sections of Title 8].”69 This provision, however, was enacted before legislation implementing the CAT domestically was passed.70 While Congress did not explicitly overrule § 1252(a)(2)(C) in the CAT implementation, this timeline demonstrates that the legislative intent of the provision could not have been to strip review of CAT claims. Thus, it is not at all clear that the jurisdiction-stripping provision can be applied to CAT claims, especially given that the CAT is intended to help those that have no other options for relief and face serious harm if returned. Congress officially implemented the CAT into US law through the Foreign Affairs Reform and Restructuring Act of 199871 (FARRA). FARRA provided that “[i]t shall be the policy of the United States not to expel, extradite, or otherwise effect the involuntary return of any person to a country in which there are substantial grounds for believing the person would be in danger of being subjected to torture,”72 almost word for word the language from the original CAT.73 Further, FARRA stated that “nothing in this section shall be construed as providing any court jurisdiction to consider or review claims raised under the Convention . . . except as part of the review of a final order of removal.”74 Essentially, this provision means that CAT relief must be raised defensively—an applicant can seek CAT relief only as a last resort to prevent removal. This is different from asylum, which can be raised offensively before an individual is ever in removal proceedings.75 This provision, which focuses on when CAT relief can be claimed, suggests that Congress is particularly keyed into the importance of finality, an issue this Comment addresses further in Part III.

The REAL ID Act of 200576 explicitly returned review of questions of law to appellate courts in these cases. This statute, in 8 USC § 1252(a)(2)(D), provides that “[n]othing in subparagraph (B) or (C) . . . shall be construed as precluding review of constitutional claims or questions of law raised upon a petition for review filed with an appropriate court of appeals in accordance with this section.”77 As with IIRIRA, this provision is not specifically tied to CAT claims.

Implementing the complex removal system is a tangle of federal regulations. These regulations provide a variety of procedural rules and definitions of related terms. Importantly, these regulations specify what standards IJs are to use when evaluating claims of potential torture. 8 CFR § 1208.16(c)(2) specifies that the applicant faces the burden of proof of establishing a likelihood of torture. IJs are told to consider factors such as: “[e]vidence of past torture,” “[e]vidence that the applicant could relocate to a part of the country of removal where he or she is not likely to be tortured,” and “[e]vidence of gross, flagrant or mass violations of human rights,” among other unenumerated considerations.78 Additionally, 8 CFR § 1208.18 defines “torture” as “any act by which severe pain or suffering, whether physical or mental, is intentionally inflicted on a person” and as “an extreme form of cruel and inhuman treatment.”79

II.  Federal Appellate Jurisdiction to Review Questions of Fact in Deferral-of-Removal Claims

Ten circuit courts have addressed whether their jurisdiction extends to questions of fact for denials of deferral of removal. This Part lays out the reasoning that courts have used to address this question. Part II.A explores the reasoning of circuits that find review is limited to questions of law. Part II.B traces the logic of circuits that do allow for review of questions of fact. Because REAL ID was passed in 2005, this Comment limits its analysis to cases from that point forward.

A.    Review Is Limited to Questions of Law

This Section details the legal analysis of the courts—the First, Second, Third, Fourth, Fifth, Sixth, Eighth, and Eleventh Circuits—that restrict appellate review to questions of law. Most of the early decisions were perfunctory and operated on an assumption that there was no jurisdiction. The number of circuits to thoroughly engage in an analysis of the statutory language and the entire jurisdictional scheme is thus limited. Among these circuits, the emphasis is on a strict reading of the text of § 1252(a)(2)(C). This Section also explores a middle approach, which agrees that § 1252(a)(2)(C) restricts review to questions of law, but defines “question of law” in a broader manner to allow for more review.

1.   An assumption of the lack of jurisdiction.

The first set of opinions dealing with federal appellate court jurisdiction to review denials of deferral of removal operated on an assumption that there is no appellate review of questions of fact. There was little analysis of the statutory scheme. The courts addressed the question summarily in sections on jurisdiction, rather than as an issue on which the appeal would turn or as a contested issue. One of the earliest cases to address this question in the wake of REAL ID was Hamid v Gonzales.80 Here, the Seventh Circuit held that § 1252(a)(2)(C) “generally prohibits us from reviewing the removal orders of aggravated felons.”81 Akram Hamid sought review of an IJ decision refusing to allow expert testimony via telephone during the hearing and of an IJ decision denying him CAT relief.82 The court noted that § 1252(a)(2)(D) “gives us jurisdiction to review constitutional claims or questions of law—such as Hamid’s due-process claim—raised in a petition for review even if the petitioner is an aggravated felon.”83 Thus, the court reviewed Hamid’s claim that refusing to allow expert testimony via telephone violated his due process rights.84 The court refused to review the denial of CAT relief, however, because that question came “down to whether the IJ correctly considered, interpreted, and weighed the evidence presented,”85 which is a factual issue barred by § 1252(a)(2)(C). The First,86 Fifth,87 and Sixth88 Circuits have also found that they lack jurisdiction to review questions of fact, but have not thoroughly evaluated the question in any opinion.

In recent years, courts have begun to more fully analyze whether there is jurisdiction to review questions of fact. Therefore, decisions that merely assume a lack of jurisdiction are becoming outdated. The First, Fifth, and Sixth Circuits, however, still rely on decisions that assume a lack of jurisdiction. While the level of analysis in this circuit split has progressed beyond a simple assumption, it is still important to understand these cases—they not only explain the history and development of jurisprudence on this question, but help to account for why more circuits have found that jurisdiction is limited to questions of law than have extended jurisdiction to questions of fact. Among courts that have fully analyzed the issue, the split is more balanced.

2.   A focus on the plain text of the statute.

Of the circuits that have detailed their reasoning for denying jurisdiction to review questions of fact, the unanimous approach is to focus on an interpretation of the plain text. In Lovan v Holder,89 Chanh Lovan sought CAT relief to prevent his removal to Laos, claiming that he would likely be tortured because of his Christian faith.90 Both the IJ and the BIA held that Lovan failed to meet his burden of proof by failing to demonstrate that it was more likely than not that he would be tortured.91 On appeal to the Eighth Circuit, Lovan argued that § 1252(a)(2)(C) was not controlling because § 1252(a)(4) supersedes it.92 Section 1252(a)(4) provides that “a petition for review filed with an appropriate court of appeals in accordance with this section shall be the sole and exclusive means for judicial review of any cause or claim under [the CAT].”93 The court briefly dealt with this argument by noting that “Section 1252(a)(4) provides that CAT claims may only be raised in petitions for review under § 1252. It does not grant reviewing courts greater jurisdiction over CAT claims than over other claims.”94

The recent Second Circuit decision in Ortiz-Franco v Holder 95 provides the most thorough defense of the view that § 1252(a)(2)(C) precludes appellate review of questions of fact. As discussed in the Introduction, Ortiz-Franco was ineligible for withholding of removal due to criminal convictions.96 In seeking CAT protection, he claimed that MS-13, a violent Salvadoran gang, was likely to torture him upon removal to El Salvador in retaliation for Ortiz-Franco’s perceived cooperation with the US government in providing information about gang activities.97 The IJ denied relief, finding insufficient evidence to establish that anyone in MS-13 in El Salvador would have knowledge of Ortiz-Franco’s meetings with the US government.98 Like Lovan, Ortiz-Franco also challenged the jurisdictional bar by arguing that the structure of the statute, read in concert with other provisions of § 1252, demonstrated that the jurisdictional bar does not apply to CAT claims. Ortiz-Franco argued that “unless § 1252(a)(4) is read as an exception to the jurisdictional limitation, it would be surplusage because FARRA already makes clear that no court has jurisdiction to consider or review claims raised under the CAT except as part of the review of a final order of removal.”99 The Second Circuit reached the same decision as the Eighth Circuit did in Lovan, holding that § 1252(a)(4) merely “confirm[s]” that review is available only after a noncitizen has been judged removable and “clarif[ies]” that some form of review is available.100 Under this view, § 1252(a)(4) does not negate the jurisdiction-stripping provision of § 1252(a)(2)(C).

In Ortiz-Franco, the plaintiff raised additional statutory claims, beyond what Lovan addressed. Ortiz-Franco argued that because deferral of removal is a temporary remedy, the fact that § 1252(a)(2)(C) is limited to final orders of removal makes it inapplicable to his case.101 The Second Circuit expressed concern that if it “were to treat the adjudication of the deferral claim as some non-final determination . . . [it] would lack jurisdiction to review any denial of deferral, even one that did raise a constitutional claim or question of law.”102 The court additionally reasoned that a “denial of deferral means that a removal order may be carried out at once.”103 A denial of deferral of removal, thus, is a final order of removal.104 Further, Ortiz-Franco argued that § 1252(a)(2)(C) should not apply because the denial of deferral of removal did not turn on his criminal convictions.105 Rejecting this framing, the Second Circuit asserted that “denial of the application for deferral of removal under the CAT is not the reason the alien is removable. Ortiz-Franco is removable—and concededly so—because he entered the country illegally and then committed crimes that render an alien removable.”106 Here, the Second Circuit focused the question on the proximate cause of Ortiz-Franco’s removability. While Ortiz-Franco saw the proximate cause of his removability as the denial of the deferral of removal, the court saw his criminal convictions as the crux of his removability.

3.   The fact-and-law approach.

In Jean-Pierre v United States Attorney General,107 the Eleventh Circuit followed a middle path. The plaintiff, Jean Herold Jean-Pierre, was convicted of drug crimes, leading to his placement in removal proceedings.108 In seeking deferral of removal, Jean-Pierre argued that his status as an individual living with AIDS would cause him to be tortured.109 The court started with the same basic assumption as did the Hamid court: “[T]he REAL ID Act prevents us from reviewing factual determinations made by the IJ or BIA in cases involving aliens who have committed a listed criminal offense.”110 The question Jean-Pierre raised on appeal was whether the harm he was likely to face in Haiti reached the level of torture. The government responded, asserting that this was not a true “question of law” because Jean-Pierre was “really attempting to challenge a factual determination concerning the likelihood that he will be subjected to torture . . . [and] circumvent the unambiguous [congressional] limitations.”111

The Eleventh Circuit ultimately found it had jurisdiction to review mixed questions of law and fact, vacating and remanding the case to the BIA.112 The court addressed the dispute by drawing on analogies to noncitizens seeking habeas corpus relief, a mechanism allowed before REAL ID.113 In habeas claims, the Eleventh Circuit had held that jurisdiction applies not just to pure questions of law, but to mixed questions of law and fact as well.114 The court concluded that, “[w]hile the mechanism [of seeking relief from removal] changed, [ ] the scope of our review of the law did not.”115 Under the current statutory system, the court had “jurisdiction to review Jean Pierre’s claim in so far as he challenges the application of an undisputed fact pattern to a legal standard.”116 In coming to this conclusion, the court also cited legislative history specifically saying that courts may review the legal elements of mixed questions of fact and law.117 Noting that the government did not challenge Jean-Pierre’s factual claims, the court took issue with the BIA’s failure to address many of the specific facts Jean-Pierre raised.118 Given the undisputed nature of the facts, the court was unable to review how the BIA had dealt with the facts, ultimately remanding the decision for further review.119

In Pieschacon–Villegas v Attorney General of the United States,120 the Third Circuit noted that it also accepted a broader reading of the fact-and-law divide.121 The court observed that it would lack jurisdiction if the petitioner’s challenge were about a “disagreement with the BIA’s determination that he failed to sufficiently demonstrate that public officials in Colombia would likely acquiesce in his torture.”122 But, as the challenge was about the legal standard used, including whether the “BIA incorrectly stated that a number of specific circumstances cannot constitute acquiescence” or whether the BIA “misapplied the legal standard by ignoring evidence relevant to determining whether Pieschacon-Villegas will more likely than not be subjected to torture upon removal,” the court would retain jurisdiction over these questions.123 This allows plaintiffs to frame the question for review so that it invokes a legal question—whether the BIA correctly applied the legal standard—that is entirely fact dependent. For instance, the issue whether certain acts reach the level of “torture” is a legal question; it is also, however, entirely dependent on presenting the facts of what the alleged torture would entail. While this approach nominally conforms to the majority view among circuit courts and holds that § 1252(a)(2)(C) limits review to questions of law, it actually allows the court to review a greater number of appeals.

The flexibility of this approach, however, has not received universal support. The Fourth Circuit, in Saintha v Mukasey,124 took an alternate approach by keeping fact and law determinations more circumscribed. Mackentoch Saintha sought CAT protection, arguing that the Haitian government would torture him due to his family’s involvement with a rival political party.125 While the IJ granted Saintha’s deferral of removal, the BIA reversed because it found that Saintha had not sufficiently proved that the Haitian government would acquiesce in his torture.126 On appeal, Saintha argued that the BIA erred in finding that there was insufficient evidence to prove acquiescence.127 Specifically, Saintha argued that “the BIA fail[ed] to recognize the inherent difference between a government that cannot control torture by private actors and a government that ‘acquiesces’ by turning a blind eye toward such conduct.”128 To address concerns about the jurisdictional bar of § 1252(a)(2)(C), Saintha framed this question as one of mixed fact and law—whether the facts presented about the Haitian government’s behavior constituted acquiescence.129 The court did not accept this characterization, “declin[ing] to stretch reason to locate questions of law in what [is] properly analyzed as a factual determination”130 and rejecting “circuitous attack[s] on the BIA’s factual determination regarding acquiescence.”131 Unlike the court in Jean-Pierre, then, this circuit does not see the fact-and-law divide as a way to allow for more judicial review.

B.    Review Is Not Limited to Questions of Law

The Ninth and Seventh Circuits exert jurisdiction to review questions of fact arising from denials of deferral of removal under the CAT. Like the Second and Fourth Circuits, the Ninth Circuit analyzes § 1252(a)(2)(C) through a textual lens. However, the Ninth Circuit reaches the opposite result: it reads the text as allowing for judicial review of questions of fact. The Seventh Circuit, while generally agreeing with the Ninth Circuit’s reasoning, has emphasized its unique understanding of “finality.” Under this view, deferral of removal is not a “final order of removal” under § 1252(a)(2)(C), but is effectively final enough to allow for appellate jurisdiction.

1.   Decisions on the merits.

The Ninth Circuit was the first court of appeals to challenge the prevailing interpretation of § 1252(a)(2)(C). The court focused on interpreting the statutory language of “removable by reason of having committed a criminal offense.”132 In Morales v Gonzales,133 the court held that “when an IJ does not rely on an alien’s conviction in denying CAT relief and instead denies relief on the merits, none of the jurisdiction-stripping provisions . . . apply to divest this court of jurisdiction.”134 Thus, when an IJ denies CAT deferral of removal and the individual is removed, this removal is “by reason of” the denial of deferral of removal, not the criminal convictions. In Morales, the court held that because the IJ had concluded that Nancy Morales had not met her burden of proof for CAT relief, the “IJ’s denial of CAT relief was on the merits and . . . is reviewable by this court.”135 The Ninth Circuit’s approach bifurcates the merits of initial removability from the merits of CAT relief.

In Lemus-Galvan v Mukasey,136 the Ninth Circuit confirmed this approach. The IJ denied deferral of removal, finding that Gustavo Lemus-Galvan could return to a different part of Mexico, avoiding any possible violence from a drug cartel involved in a conflict over turf with Lemus-Galvan’s family.137 The Ninth Circuit explicitly held that “[t]he jurisdiction-stripping provision of 8 U.S.C. § 1252(a)(2)(C) does not deprive us of jurisdiction over denials of deferral of removal under the CAT, which are always decisions on the merits.”138 Because the denial of deferral of removal had nothing to do with the plaintiff’s criminal history, the Ninth Circuit held that § 1252(a)(2)(C) did not apply. This was a decision “on the merits,”139 as opposed to a technical decision based on criminal convictions.

2.   Finality.

The Seventh Circuit is the only other appellate court to hold that § 1252(a)(2)(C) does not bar review of questions of fact. This development marked a reversal from the Seventh Circuit’s jurisprudence in Hamid—there, the court held that CAT claims were subject to the same jurisdictional bars as any other immigration claim.140 Hamid, who had aggravated felony convictions, could receive appellate review only by presenting a legal question.141 This Section details the Seventh Circuit’s new jurisprudence post-Hamid, which focuses on the inherently nonfinal nature of deferral of removal.

The first step in this direction was in Issaq v Holder.142 Adnan Issaq was an Iraqi citizen of Assyrian Christian descent and feared torture by Muslim extremist groups if returned to Iraq.143 The court drew an analogy to another jurisdictional bar—the “persecutor bar.”144 Although the persecutor bar comes from a different part of the statute and is structured differently from the criminal conviction bar, the Supreme Court noted in Negusie v Holder 145 that “in light of the CAT’s requirement[,] . . . regulations implementing that convention provide ‘deferral of removal’ to aliens subject to the INA persecutor bar who would more likely than not be tortured if removed to their home country.”146 In applying Negusie to Issaq’s case, the Seventh Circuit explained that “[o]nce an alien succeeds in proving the factual prerequisites for relief under the CAT, we understand Negusie to hold that some kind of remedy (complete with judicial review) is available.”147 The court also emphasized the difference between challenging a final order of removal and challenging the denial of deferral of removal. Deferral of removal is “inherently non-final” so § 1252(a)(2)(C), “which speaks only of a final order,” does not apply.148 In Issaq, the petitioner had not explicitly raised the question of jurisdiction, so the court addressed the issue of finality in just one paragraph.149 The framing of deferral of removal as “inherently non-final,”150 though, serves as the basis for the court’s subsequent jurisprudence on this issue.

In Wanjiru v Holder,151 the Seventh Circuit revisited and expanded on the Issaq opinion.152 Wanjiru, a native of Kenya, was placed in removal proceedings after a sexual misconduct conviction.153 Although he conceded removability to the IJ, he sought deferral of removal under the CAT.154 As a teen in Kenya, Wanjiru had been involved with a group known as the Mungiki.155 According to Wanjiru, the Mungiki “punish defectors by executing them.”156 Even if the Mungiki did not harm Wanjiru, he feared that the Kenyan police would recognize him as a Mungiki member and torture him.157 Although the government conceded jurisdiction at the appellate stage,158 the Seventh Circuit analyzed the availability of judicial review for deferrals of removal due to the circuit split.159 The court again focused on the importance of finality, noting that § 1252(a)(2)(C) explicitly governs “final orders of removal.”160 In the Seventh Circuit’s view, however, deferral of removal is not a final order; rather, a “deferral of removal is like an injunction . . . prevent[ing] the government from removing the person in question, but it can be revisited if circumstances change.”161 In so characterizing a deferral of removal, the Seventh Circuit preemptively addressed the Second Circuit’s concerns in Ortiz-Franco. In that case, the Second Circuit was concerned that if deferral of removal is not seen as a final order, appellate courts would lack jurisdiction to review any claims arising from deferral-of-removal actions.162 The Seventh Circuit walked a middle line—deferral of removal is “final enough to permit judicial review, but at the same time not [ ] the kind of ‘final’ order covered by § 1252(a)(2)(C).”163 Deferral of removal is “final enough to permit judicial review” because this decision can be “the dispositive ruling in a case.”164 Beyond the primary grounding of the decision in the interpretation of what constitutes a final order of removal, the court concluded with additional policy concerns, particularly that “[w]e should not lightly presume that Congress has shut off avenues of judicial review that ensure this country’s compliance with its obligations under an international treaty.”165

The question whether appellate courts can review questions of fact stemming from denials of deferral of removal has vexed courts for years now. Even among the courts that agree on an outcome, the reasons for said outcome vary. Eight circuit courts currently hold that they do not have jurisdiction to review questions of fact—the First, Fifth, and Sixth Circuits have not engaged in a thorough examination of the question; the Second, Fourth, and Eighth Circuits rely on a strict reading of the plain text; and the Third and Eleventh Circuits allow for review of mixed questions of fact and law but not pure questions of fact. On the other side of the split, the Seventh and Ninth Circuits also base their holdings on different grounds—although both find more latitude in the text than do the courts refusing to find jurisdiction. The Ninth Circuit focuses on the proximate cause of the removability, namely the denial of relief rather than the criminal convictions. The Seventh Circuit instead emphasizes the nonfinal nature of deferral of removal.

III.  Flexible Finality

This Part develops a framework of flexible finality, expanding on the Seventh Circuit’s holding that deferral of removal is final enough for appellate review, but not so final as to constitute a final order of removal. The Seventh Circuit’s concept of flexible finality is an important idea—yet the entire discussion in both Issaq and Wanjiru is comprised of just a few paragraphs. This Comment provides a much-needed analysis of the concept of flexible finality. Part III.A looks at interlocutory appeals and concludes that this approach, while illuminating, does not fit with the statutory restraints of CAT claims. Part III.B compiles and analyzes different areas of law to demonstrate that flexible finality is already the de facto law in certain subject areas. This conceptualization of flexible finality helps resolve this circuit split and has the potential to be expanded and incorporated in other areas of law. Part III.C applies the test for flexible finality developed in Part III.B to CAT deferral-of-removal claims. Part III.C also responds to counterarguments that this approach disrespects the statutory text and congressional intent.

A.    Treatment of Review as an Interlocutory Appeal

Following the Wanjiru court’s description of deferral of removal as an “injunction,” it is proper to consider whether appellate review of a denial of deferral of removal is best thought of as an interlocutory appeal. Federal courts of appeals have jurisdiction over appeals of interlocutory orders affecting the grant or dismissal of injunctions.166 If deferral of removal is thought of as an injunction, then it is valuable to consider how the rules of injunctions and interlocutory appeals might apply to deferral of removal.

1.   Deferral of removal as an injunction.

In general, appellate court jurisdiction is restricted to the review of final orders of federal district courts.167 In certain circumstances, however, appellate courts have jurisdiction to review nonfinal orders through interlocutory appeals.168 Appellate court jurisdiction over interlocutory appeals is governed by 28 USC § 1292. Importantly, appellate courts have jurisdiction over orders “granting, continuing, modifying, refusing or dissolving injunctions, or refusing to dissolve or modify injunctions.”169 The fact that injunctions are covered by interlocutory orders demons­trates that injunctions are not entirely “final.” Injunctive relief compels action (or inaction) from the opposing party.170 In the case of preliminary injunctions, the injunctive relief is granted before the trial court issues a final judgment—it is a temporary order compelling or preventing certain actions while the case works its way through the court system.171 Injunctions can also be the final remedy sought by a party.172

Similarly, deferral of removal prevents the US government from removing a noncitizen to his home country. Deferral of removal is a temporary remedy; 8 CFR § 1208.17(d)(1) states that “[a]t any time while deferral of removal is in effect,” the government can seek to demonstrate that the likelihood of torture or government acquiescence has abated.173 Further, the government can seek termination if the State Department receives assurances from the receiving country that there will be no torture.174 Similarly, in the injunction context, the Federal Rules of Civil Procedure (FRCP) provide that “the court may relieve a party . . . from a final judgment, order, or proceeding . . . [if] it is based on an earlier judgment that has been reversed or vacated; or applying it prospectively is no longer equitable.”175

2.   The collateral order doctrine.

The Supreme Court addressed the standards for interlocutory appeal in Lauro Lines SRL v Chasser.176 Only a limited set of orders can be reviewed under the “collateral order doctrine.”177 This doctrine creates three requirements that a plaintiff must meet to show that an interlocutory appeal is proper. The order plaintiffs seek review of must (1) “conclusively determine the disputed question,” (2) “resolve an important issue completely separate from the merits of the action,” and (3) “be effectively unreviewable on appeal from a final judgment.”178

The criterion that creates the most difficulty for treating the denial of deferral of removal as an interlocutory appeal is the second: the issue in the order must be “completely separate” from the merits of the action. To determine whether the issue is completely separate from the merits of the action, it is necessary to define what the “action” is. If the “action” is just the deferral of removal, then this prong fails. There is a complete overlap between the action and the order seeking an interlocutory appeal. If, however, the “action” is the entire removal proceeding, the outcome is different. The entire removal proceeding asks one overarching question: Will the individual be removed from the country? Deferral of removal asks a different, unique question: Is the individual likely to be tortured upon removal? Deferral of removal, therefore, is an independent issue from the question of removability, arguably satisfying the second prong of the test.

In contrast, the first prong (conclusive determination) and third prong (effectively unreviewable) of the collateral order doctrine are easily met by a denial of deferral of removal, regardless of how the “action” is defined. An order reviewing the denial of deferral of removal would “conclusively determine the disputed question.”179 When deferral of removal is denied, a noncitizen in removal proceedings has no other recourse and will be removed. This remains the same whether “action” is conceived of as the entire removal proceedings or just the deferral of removal. Further, the order would be “effectively unreviewable” on appeal, regardless of the characterization of “action.”180 From a practical perspective, after a noncitizen is removed after being denied deferral of removal, there is no way to get review of the claim while in another country.

3.   Appeal of CAT deferral of removal is not interlocutory.

While CAT deferral-of-removal claims operate as injunctions and can be framed to meet all the requirements of the collateral order doctrine, there is a serious conceptual issue with framing appeals of denials of CAT deferral of removal as interlocutory appeals. In the general case of an interlocutory appeal—for example, a challenge of the grant of a preliminary injunction—the appellate court addresses the interlocutory order so that the trial court can promptly continue with the merits of the original case.181 However, because of the statutory framework of the CAT, deferral-of-removal claims can be raised only once the petitioner is already in removal proceedings. Additionally, appellate courts can review CAT claims only “as part of the review of a final order of removal.”182 Therefore, the petitioner cannot seek review of the denial of deferral of removal until the original case (the removal proceeding as a whole) is over and a final order of removal (on the merits of removability) is in place. As such, by the time an appellate court would review the merits of the CAT claim, the IJs and BIA will have already adjudicated the merits of the removal of the individual (and the merits of the CAT claim of which review is sought). Thus, there is no “work” left to be sent back to a lower court. This is arguably a collateral issue, but it is not interlocutory as it is not an appeal in the middle of a larger case. An interlocutory appeal, by definition, must be an appeal that takes place in the midst of a larger, ongoing case;183 given the structure of the CAT, it is impossible to meet this baseline requirement. Treating judicial review of denial of deferral of removal as an interlocutory appeal, then, does not resolve the issue.

B.    The Spectrum of Finality

In general, the US court system requires a final judgment from the federal district courts to give the federal courts of appeals jurisdiction.184 In Wanjiru, however, the Seventh Circuit raised the possibility that there are different levels of finality. This Section establishes that flexible finality is a concept that, while not written about much, is already in effect in a number of areas of law. The following analogous areas of law demonstrate that finality is used in a flexible manner when: (1) the order at issue effectively decides the outcome of the case and (2) the enforcement of a strict finality rule would lead to inequitable results. These rules are applied to the CAT deferral-of-removal context in Part III.C.

1.   Consent decrees.

Consent decrees are “judgment[s] entered by a court based upon settlement terms negotiated and agreed to by the parties . . . [and] aimed at ending the litigation while securing an enforceable judgment.”185 Consent decrees can be advantageous, possibly achieving greater compliance than a traditional settlement might.186 They also have significant disadvantages, including that they “may become . . . nonresponsive to the problem[s they were] designed to address.”187 Given this problem, parties at times seek to modify consent decrees188 —an action that has a direct effect on the final judgment rule.

In Rufo v Inmates of Suffolk County Jail,189 the Supreme Court addressed the finality of consent decrees, holding that in certain circumstances they can be revised.190 Rufo dealt with an attempt to modify a consent decree that was in place to rectify conditions at a jail.191 Because a consent decree is a judgment entered by the courts, it acts like a final judgment. But, as noted above, consent decrees often fail to continue to provide proper remedies as conditions evolve. Therefore, the ability to modify consent decrees is essential to their future effectiveness. In Rufo, the Court returned to the question of when a consent decree can be modified. The Court emphasized that the precedent of United States v Swift & Co 192 was not controlling.193 Swift had allowed modification, but only when preserving the unaltered consent decree would cause a “grievous wrong.”194 Because Swift took place in the context of exceptional labor disputes, it did not represent “a hardening of the traditional flexible standard for modification of consent decrees.”195 Rather, Rufo held that modification is allowed when “a significant change in circumstances warrants revision of the decree.”196 Thus, a party must instead show “a significant change either in factual conditions or in law.”197

Allowing modification seems to cut against viewing consent decrees as final judgments. However, the court has jurisdiction to review consent decrees, demonstrating that a judgment does not need to be never changing to be “final” enough to review. Consent decrees, therefore, are final enough to be subject to appellate review, but not so final that they cannot be modified. Deferral of removal, just like a consent decree, is subject to modification based on changed circumstances. A final resolution is determined (the noncitizen will not be removed), but it is subject to change (if the country conditions change).

The status of consent decrees as final judgments was further complicated by the Prison Litigation Reform Act of 1995198 (PLRA), which Congress passed in 1996, four years after the Rufo decision. The PLRA limits the relief available to prisoners bringing civil rights suits and requires that courts ensure consent decrees follow such limitations.199 While prisoners often initiate the civil rights suits that result in consent decrees,200 modifications of the decrees under the PLRA are generally sought by state prison officials.201 The PLRA applied to previously entered consent decrees.202 District courts, however, initially disagreed over whether reopening consent decrees and changing them to conform with the PLRA violated separation-of-powers principles.203 For example, the Eastern District of Michigan held that consent decrees are final judgments for separation-of-powers purposes and cannot be vacated retroactively by the PLRA.204 In contrast, many courts now hold that consent decrees are not final for separation-of-powers purposes, without questioning the Supreme Court’s decision in Rufo. In one such case, Plyler v Moore,205 the court made a similar analogy to injunctions as the Seventh Circuit did in Wanjiru.206 The Plyler court held that because consent decrees “provide[ ] for prospective relief and [are] subject to the continuing supervisory jurisdiction of the district court,” they are “akin to [ ] final judgment[s] granting injunctive relief, and thus [are] subject to subsequent changes in the law.”207

Consent decrees are court-enforced agreements, but they are not “judgments” in the traditional sense of a trial and a verdict from a jury or a judge. Perhaps this accounts for their greater flexibility. But despite these differences, they are still enforced by a court and operate with the force of law. Further, denial-of-removal CAT claims are adjudicated by IJs, who are part of the executive branch, not the judicial branch; deferral-of-removal claims are not traditional judgments either. Thus, a consent decree is “final enough” to allow appellate review of a district court’s refusal to modify it, but not so final as to prevent modification (as in Rufo) or to have retroactive modification violate separation-of-powers principles (as in Plyler).

2.   Bankruptcy.

When companies go bankrupt, the question of how to pay off creditors looms large. Generally, the bankruptcy system “impos[es] [ ] repayment priorities.”208 The nature of bankruptcy is that not all creditors will receive the full sum of what they are owed, making the order of priorities very important. In disputes over priority order, federal district courts have jurisdiction to hear appeals from bankruptcy hearings.209 Bankruptcy appeals are available from “final judgments, orders, and decrees,”210 and from certain “interlocutory orders and decrees.”211

Bankruptcy, however, presents a particularly difficult context in which to apply the final judgment rule, as bankruptcy cases are often long-term and involve the rights of many separate parties.212 Matter of Kilgus213 explains the unique nature of bankruptcy appeals, stating that “[t]here is a twilight zone of orders in bankruptcy law in addition to those that end the case and those that would be treated as ‘collateral orders’ immediately appealable in ordinary civil litigation.”214 The court explained that this is necessary because of the inherently different nature of bankruptcy relative to normal civil litigation. These “twilight zone” orders exist because of the presence of “many subsidiary matters that litter the road to the distribution of assets in bankruptcy. A court cannot wait until the end of the case to allow the appeal, because final disposition in bankruptcy . . . depends on the prior, authoritative disposition of subsidiary disputes.”215 Here, the court expressed the idea that there is a level of finality between “pure” final judgment and interlocutory appeals.

In re Kyung Sook Kim216 is an example of the unique approach to finality in the bankruptcy courts. There, the court clarified that an “order denying a motion to remove a trustee”217 is as much a final order as is an “order removing a trustee.”218 While the denial of the motion to remove the trustee does not conclude the bankruptcy action and there is no final judgment in the traditional sense, the court still considered this a final, appealable order because the order could not be effectively reviewed until the bankruptcy’s conclusion.219 Similarly, In re Lehman Brothers Holdings Inc220 held that the denial of a motion to intervene is considered final because “in the bankruptcy context, the standard for finality is more flexible than in other civil litigation[;] . . . finality is viewed functionally, focusing on pragmatic considerations rather than on technicalities.”221 Many other federal courts employ a more flexible concept of finality in bankruptcy as well.222

The bankruptcy cases demonstrate the principle that certain claims can be unreviewable if courts wait until traditional final judgment has been issued on the proceedings as a whole. In bankruptcy, the money is divided and gone by the time review comes. This means that an earlier review of the claim will have a significant impact on the final division of money. A strict finality rule also could lead to very inequitable results, if one party seeking recovery of money in bankruptcy is unable to get what it is due. With a CAT deferral-of-removal claim, the claim is effectively unreviewable because denial triggers the petitioner’s removal to another country. Additionally, as in Lehman Brothers and Kim, the denial of deferral of removal is just as final as the grant of deferral of removal. In some sense, the denial of deferral of removal is a stronger case for finality, since the grant of deferral of removal leads to temporary relief, whereas the denial leads to permanent removal. Certainly the task of divvying up monetary assets is different than adjudicating removability. The equitable concerns on which this flexibility in bankruptcy rests, however, are arguably stronger with deferral of removal given the life-and-death stakes.

3.   Patents.

The United States Patent and Trademark Office (PTO) adjudicates the initial determination of whether to grant a patent or not.223 The initial decision of the primary patent examiner may be appealed to the Patent Trial and Appeal Board.224 Then, an applicant “who is dissatisfied with the final decision in an appeal to the Patent Trial and Appeal Board . . . may appeal the Board’s decision to the United States Court of Appeals for the Federal Circuit.”225 Additionally, “any person at any time [may] petition the PTO to reexamine a patent to determine whether prior art raises a substantial new question of patentability that had not been considered during the original patent prosecution.”226 Thus, a dissatisfied party has two options: seek review in the federal court system or request reexamination from the PTO.227

The recent saga of Fresenius USA, Inc v Baxter International, Inc228 is instructive. The Fresenius cases dealt with the simultaneous occurrence of the two systems discussed above: invalidity claims in the courts and reexamination claims in administrative agencies. In 2003, Fresenius sued in federal district court claiming that certain Baxter-held patents were invalid.229 While a jury initially found in Fresenius’s favor, the district court entered judgment as a matter of law in Baxter’s favor and another jury awarded damages to Baxter in November 2007.230 After appeals, the Federal Circuit held in 2009 that one of the challenged patents was not invalid, affirming in part the district court’s judgment against Fresenius.231 Finally, in March 2012, the district court awarded Baxter damages for Fresenius’s infringement of the valid patent and entered final judgment.232 At the same time, a parallel suit was taking place in the PTO reexamination process. In 2005, Fresenius sought reexamination of the same patent challenged in federal court.233 Relying on new prior art not raised in the initial PTO examination or in the first district court trial, the PTO found on reexamination in December 2007 that Baxter’s patent was invalid.234 In May 2012, the Federal Circuit upheld the PTO’s reexamination.235

The issue before the court, then, was whether the district court’s 2007 final judgment was final enough to prevent being disturbed by the subsequent PTO determinations. Baxter argued that res judicata should not allow the reopening of a final judgment.236 The Federal Circuit disagreed, finding that there was no reopening of a final judgment because no final judgment was ever issued.237 Given that the parties appealed the Fresenius I decision, the Federal Circuit then had to explain how the parties’ appeal from Fresenius I was allowed if the 2007 judgment was not final. The court held that although the district court “entered a judgment final for purposes of appeal, and that judgment might have been given preclusive effect in another infringement case between these parties, it was not sufficiently final to preclude application of the intervening final judgment in [the PTO reexamination case].”238 Additionally, the court addressed concerns about separation-of-powers violations. The Federal Circuit noted that “the power to issue a final judgment . . . resides in the judicial branch.”239 This, according to the court, was not an issue—because no final judgment had ever been given from the district court, the subsequent administrative rulings did not disturb the judiciary’s ability to grant final judgment.240

A key takeaway is that through the reexamination process, an administrative agency can challenge the finality of a federal court decision.241 As the Federal Circuit noted, a judgment can be final for the purposes of appeal, but not final enough to prevent a subsequent administrative action from overturning the court’s initial holding. Critically, this establishes that the finality allowing for judicial review does not necessarily mean that the order itself is final for other purposes. The policy of flexible finality likely turns on equitable considerations. Courts and administrative agencies are willing to bend the strict finality rules because of inequitable outcomes: the court will not order enforcement of a patent infringement holding once the patent has been deemed invalid. If the inequity of enforcing these judgments is a sufficient reason to make final judgment flexible, then certainly the possibility of removal and torture is sufficient to allow review of the nonfinal deferral-of-removal order.

Examining consent decrees, bankruptcy “twilight” orders, and the impact of patent reexamination reveals that courts are already using the concept of flexible finality in many areas, albeit without calling it by name. This Comment argues that flexible finality should be an option in all areas of law. The analogies from Part III.B help create a test for when it might be appropriate to apply finality flexibly.242 That test, created using the lessons of the analogies in this Section, is two-part. First, the order of which review is sought must effectively resolve the case. This prong will help limit frivolous arguments for applying flexible finality. If the order does not effectively resolve the case, then there is no reason why a petitioner cannot wait for the traditional final judgment to review claims.

Second, an application of a pure final judgment rule must be inequitable. If the petitioner is able to seek adequate review and remedy after traditional final judgment is reached, then the burden on the court system outweighs the burden on the plaintiff. The first prong is a mechanism of screening to ensure flexible finality is limited to intermediate orders with a sufficiently high impact on the resolution of the case. The second prong invites a consideration of the costs to the government and the petitioner, allowing the administrative benefits of pure finality to be overridden only in contexts with extremely high costs to the petitioner. The next Section applies this test to CAT deferral-of-removal claims and responds to counterarguments.

C.    Balancing the Equities to Allow for Review of Fact in CAT Claims

Given this Comment’s assertion that finality is a spectrum, it must next be established that CAT deferrals of removal are the kind of orders that are properly reviewed under this flexible understanding of finality. This Comment does not purport to throw out the final judgment rule, but rather to allow for leniency in cases in which the equities point towards allowing review. Part III.B established the test for applying flexible finality: an order must effectively conclude the case, and the enforcement of a strict finality rule must be inequitable. The first prong is easily satisfied in the CAT context: CAT deferral-of-removal claims, if enforced, are effectively final because a denial of deferral results in removal. Part III.C.1 addresses the second prong—the equities—and concludes that deferral of removal necessitates a relaxed conception of finality. The rest of Part III.C responds to counterarguments: Part III.C.2 addresses concerns that words can have only one meaning, and Part III.C.3 refutes claims that allowing judicial review of questions of fact for deferral-of-removal cases subverts legislative intent.

1.   The equities support review of fact in CAT claims.

Federal appellate courts may face many of the same difficulties IJs do when adjudicating CAT claims, including the problem that reliable corroborating evidence is often unavailable. Under this conception of finality, however, appellate courts would not be reviewing questions of fact de novo. Instead, the review would assess whether IJ decisions are supported by “substantial evidence.”243 While “[i]t is true that immigration judges are often in the best position to evaluate credibility[,] . . . a second look at the facts, by a detached group of individuals, can lead to fewer wrongful denials.”244 Importantly, there is a real need for a check on abuses in immigration courts. In recent years, there has been a spate of criticism against IJs and the quality of their decisions.245 To cope with the seeming arbitrariness coming from some IJs, applicants have turned to the federal courts as the “first line of defense against mistaken or biased immigration judge decisions.”246 Judge Posner stated that “the adjudication of these cases at the administrative level has fallen below the minimum standards of legal justice.”247 Judge Julio Fuentes noted that the tone of certain IJs was “more appropriate to a court television show.”248 Judge Marsha Berzon complained that a certain IJ’s written decision was “literally incomprehensible.”249 In another instance, Berzon focused specifically on how the low quality of IJs has a direct impact on factual determinations, writing that “[t]he overall tone of Immigration Judge Brian Simpson’s opinion in this case is such that I can have no confidence in his factual findings. His opinion is belittling and patronizing as well as inaccurate, even as to less material details.”250 Given these anecdotal complaints, as well as empirical evidence of disparities among IJs in decisionmaking,251 CAT applicants have a strong argument that the equities support jurisdiction for federal appellate review.252

Additionally, there is evidence that the Supreme Court is skeptical of Congress’s jurisdiction stripping.253 Professor Richard Fallon notes that, in the context of the detention of suspected terrorists, the Court “did not question Congress’s employment of non-Article III tribunals” but that it did “require[ ] an oversight role for the Article III judiciary.”254 Although that decision was very limited,255 it expresses a general idea that when some fundamental rights are at issue, federal courts must have jurisdiction to review agency or tribunal decisions. Fallon acknowledges that Boumediene v Bush256 “rested wholly on the Suspension Clause” and issues relating to habeas corpus rights.257 However, Fallon argues that despite not being directly applicable to jurisdictional strips beyond the habeas context, “the decision was very much about jurisdiction-stripping”258 and “that it is the necessary and proper function of the Judicial Branch to say [authoritatively] what the law is.”259 While it is not clear that the Court will expand this approach beyond the habeas context, the questioning of the propriety of jurisdiction strips is relevant here. First, CAT deferral-of-removal claims certainly raise issues related to US treaty obligations. And second, noncitizens are not without due process rights, though the extent of those rights is unclear.260

The primary interest against granting review is that an expanded view of finality, whether in the CAT deferral-of-removal context or elsewhere, cuts against the efficiency of the federal court system.261 But administrability concerns do not necessarily “surpass[ ] the worth of any substantive right.”262 Factual review of CAT claims would not necessitate a full trial or even oral arguments, but merely a check that the petitioner’s factual claims were considered and weighed thoughtfully. It is likely that this could be done without expending a great deal of appellate resources.

2.   Terms with dual purposes.

While it may seem odd for a denial of deferral of removal to be both final enough for appellate review and not final enough to be a final order of removal, CAT claims are not the only instance of courts allowing statutory terms to serve dual purposes. Words—even the same statutory language—do not always mean the same thing in differing contexts.

The Supreme Court’s evaluation of taxes in National Federation of Independent Business v Sebelius263 is a useful comparison. In NFIB, the Court held that the Affordable Care Act’s “penalty” was not a tax for purposes of the Anti-Injunction Act, but was a tax for purposes of Congress’s taxing power.264 The Anti-Injunction Act is a jurisdictional bar, preventing litigants from challenging the government’s ability to collect taxes.265 The Affordable Care Act labeled the fine for not joining a health plan a “penalty,” and the Court reasoned that because “[t]here is no immediate reason to think that a statute applying to ‘any tax’ would apply to a ‘penalty,’” the Anti-Injunction Act did not apply to the fine.266 Having established that there was no bar to hearing the case, the Court then turned to the merits of whether the penalty was properly levied under Congress’s taxing power. Despite Congress’s labeling of the fine as a “penalty,” because “the only consequence [of not buying health insurance] is that he must make an additional payment to the IRS when he pays his taxes,” the “penalty” is properly considered a “tax” that Congress has the authority to levy.267 Addressing the apparent conflict between these two holdings, the Court noted that Congress has the ability to decide whether the Anti-Injunction Act applies to a statute, but that choice does not impact the constitutional analysis of the taxing power.268 Therefore, it is entirely possible that “final” means something different in both the final judgment context and the final-order-of-removal immigration context.

3.   The problem of legislative intent.

The text of § 1252(a)(2) clearly reveals a legislative intent to strip courts of jurisdiction to hear some claims from criminal noncitizens. There are problems with simply stopping the analysis there, though, particularly given the lack of legislative history about jurisdiction stripping as applied to CAT claims. As described in Part I, the jurisdiction-stripping provision was passed before the CAT was implemented domestically and Congress did not choose to explicitly extend the jurisdiction strip to CAT claims upon such implementation. In addition, the canon of immigration lenity, the canon of Murray v Schooner Charming Betsy,269 and an emphasis on pragmatism support the use of flexible finality to allow for jurisdiction to review questions of fact.

The canon of immigration lenity is based on a similar statutory construction canon from criminal law, which directs that “ambiguities in penal statutes be construed in favor of the defendant.”270 The Supreme Court blessed the application of this canon in the immigration context in Fong Haw Tan v Phelan.271 The Court held that “since the stakes [of deportation] are considerable for the individual, we will not assume that Congress meant to trench on his freedom beyond that which is required by the narrowest of several possible meanings of the words used.”272 Some scholarship has discussed the difficulty of reconciling administrative deference under Chevron U.S.A. Inc v Natural Resources Defense Council, Inc273 with the canon of immigration lenity.274 Chevron, however, applies to administrative rulemaking and “legal interpretations made by the BIA in adjudications.”275 The noncitizens’ claims at issue in this Comment, though, are factual claims. Thus, concerns about reconciling Chevron and the immigration lenity canon are less relevant here, because the two are not in direct conflict. Even if the BIA were to issue a ruling on the jurisdiction-stripping provision of § 1252(a)(2)(C), it is not at all clear that such a decision would be entitled to Chevron deference. Agencies are due deference in large part because of their ability to “bring [their] expertise to bear upon the matter.”276 The BIA may have expertise in adjudicating CAT deferral-of-removal claims, but that does not also mean that it has expertise interpreting jurisdiction-stripping provisions that solely impact the federal circuit courts’ ability to hear certain cases.

Appellate court review of factual claims in denial of deferral-of-removal cases is an appropriate place to utilize the canon of immigration lenity. Congress passed § 1252(a)(2)(C) before the CAT was implemented domestically. Thus, it is not entirely clear how Congress meant the CAT to interact with any jurisdiction-stripping provision. Moreover, there are clear ambiguities in the statute, as applied to CAT deferral-of-removal claims. Temporary relief is granted only through withholding of removal and deferral of removal. Only deferral of removal triggers the possibility of the jurisdictional strip—the jurisdictional strip applies to applicants who have criminal convictions, and deferral of removal is a remedy created to aid noncitizen applicants who have criminal convictions. As illuminated by the Seventh and Ninth Circuits, there are real ambiguities about what a “final order of removal” means when dealing with temporary relief, and what it means to be removed “because of” criminal convictions.277 Applying the canon of immigration lenity to the circuit split at issue here, the ambiguities in the statute should be decided in favor of the noncitizen to grant appellate review over questions of fact.

The Charming Betsy canon additionally supports judicial review of questions of fact.278 The Court noted that “an act of Congress ought never to be construed to violate the law of nations if any other possible construction remains.”279 Thus, like the canon of immigration lenity, the Charming Betsy canon instructs that ambiguous statutes be interpreted to not conflict with international obligations, including treaties.280 Given all the reasons stated above, the jurisdictional strip of § 1252(a)(2)(C) is ambiguous with regards to its application to CAT claims. As such, Charming Betsy instructs courts to take care to interpret the statute at issue, § 1252(a)(2)(C), in line with treaty obligations, the CAT itself.281 In reconciling domestic statutes with international treaty obligations, courts employ a last-in-time rule, meaning that the later law (domestic or international) controls.282 Charming Betsy, thus, comes into play when attempting to reconcile “a non-self-executing treaty and a later-in-time ambiguous statute.”283 That is precisely the case here: the CAT, a non-self-executing treaty, conflicts with § 1252(a)(2)(C), a later-in-time ambiguous statute. The CAT directs states to not remove individuals likely to be tortured. A denial of deferral of removal—without appellate review to check for erroneous IJ fact-finding—can lead to the exact result the CAT is meant to protect against. The Charming Betsy canon adds more weight on the side of interpreting ambiguities in § 1252(a)(2)(C) in favor of the CAT petitioner.

Flexible finality also relies on a theoretical underpinning of pragmatism rather than formalism. In Wanjiru, the Seventh Circuit utilized a pragmatic interpretation of finality.284 To start, the court recognized that deferral of removal is not “final” because it is a temporary relief; thus, it is not a “final order of removal.”285 The court also held that even though “deferral,” by its name, sounds nonfinal, the resolution of this question does in fact effectively end the litigation and determine whether the claimant will be removed or not.286 Accepting a spectrum of finality requires accepting a pragmatic approach and accepting that finality may mean different things in different contexts. One view of pragmatism defines it as “seek[ing] to refocus modern judging away from abstract, high-sounding ‘distractions’ such as ‘fidelity’ and ‘judicial restraint,’ and toward the concrete social consequences of adjudication.”287 Certainly this is not an uncontroversial legal philosophy.288 But, given that courts and scholars have long grappled with defining finality and applying it evenly across different areas of law, it is perhaps naive to cling to formalism in this context. Rigid formality in the CAT context would eliminate any judicial review for individuals who are claiming a high likelihood of torture.289

Pragmatism and balancing tests can be criticized on some of the same lines as formalism, namely that they are arbitrary and unpredictable.290 According to Professor Karl Llewellyn, however, pragmatism does not have to be equivalent to “ad hoc decisionmaking.”291 It can be better conceived of as “adopting a ‘solving rule’ to deal with the ‘type of situation’ before the court.”292 This is similar to the issues with finality discussed above. CAT deferral of removal is a unique “type of situation”—it affects the basic human right to not be tortured, it implicates the United States’ international treaty obligations, and the current system relies on IJs who are sometimes derelict in their duties. This is the type of situation that needs a “solving rule” that is not constrained by an overly rigid reading of the statutory terms.

Conclusion

The issue of whether federal courts of appeals have jurisdict­ion to review questions of fact arising from denials of deferral of removal is an entrenched dispute involving ten courts of appeals. It implicates the fundamental human right not to be tortured and the United States’ treaty and statutory obligations.

In Wanjiru, the Seventh Circuit held that an order can be “final enough” to receive appellate review, but not so “final” as to constitute a “final order of removal.”293 This was just one sentence in the court’s decision, but unpacking its meaning and origin is crucial to resolving this circuit split and to providing a new way of conceptualizing finality in other contexts. This Comment establishes that flexible finality is already the rule in many areas of law. Although courts do not call it by this name, they have effectively created a separate scheme of flexible finality. Using the lessons of those areas of law, this Comment creates a test for when flexible finality should be an option. Consent decrees, bankruptcy “twilight” orders, and patent reexamination provisions are distinct from each other and from CAT deferrals of removal. But they reveal two key features of the situations calling for flexible finality: the order at issue must effectively resolve the claim, and a balancing of the equities must demonstrate that the plaintiff’s interests outweigh the efficiency benefits of a strict final judgment rule.

Both of these requirements are met in the denial of deferral of removal under the CAT. On the equities, CAT applicants’ interests outweigh the interests of efficiency. The immigration courts are notoriously unpredictable and unreliable. The fidelity of the United States to its treaty obligations and the determination of whether or not an individual may face torture are at issue in CAT deferral-of-removal cases. Additionally, a denial of deferral of removal effectively ends the CAT applicant’s case and results in removal. While there are counterarguments about congressional intent to strip courts of jurisdiction, this Comment shows that those can be resolved. The interpretation this Comment argues for is a valid interpretation of the text of § 1252(a)(2)(C). Given that there are ambiguities in the statute, the canon of immigration lenity instructs that those ambiguities be resolved in favor of the noncitizen.

• 22See 8 USC § 1229(a); 8 CFR § 1003.13 (“Charging document means the written instrument which initiates a proceeding before an Immigration Judge. . . . For proceedings initiated after April 1, 1997, these documents include a Notice to Appear.”). See also US Department of Justice, Executive Office for Immigration Review, Office of the Chief Immigration Judge, Immigration Court Practice Manual § 4.2(a) at *59 (Feb 4, 2016), archived at http://perma.cc/BMN3-HM86 (“Removal proceedings begin when the Department of Homeland Security files a Notice to Appear . . . with the Immigration Court after it is served on the alien.”).
• 23See 8 USC § 1229(a)(1)(G) (noting that an NTA must inform the noncitizen of the time and place of the removal proceedings and the consequences of failure to appear at the removal hearing); 8 USC § 1229a(b)(5) (requiring IJs to conduct in absentia hearings when an individual fails to appear after receiving proper notice).
• 24Office of the Chief Immigration Judge, Immigration Court Practice Manual § 1.3(b) at *5 (cited in note 22).
• 258 USC § 1229a(a)(1).
• 26Office of the Chief Immigration Judge, Immigration Court Practice Manual § 1.5(g) at *10 (cited in note 22).
• 27See id (noting that the reviewing court is determined by the nature of the appeal).
• 28Some of the grounds for removal, other than criminal offenses, include marriage fraud, falsification of documents to receive an immigration benefit, foreign policy considerations, participation in acts of persecution perpetrated by the Nazi government, and unlawful voting in US elections. 8 USC § 1227(a)(1).
• 298 USC § 1227(a)(2).
• 308 USC § 1227(a)(2)(A)(i).
• 318 USC § 1227(a)(2)(A)(iii).
• 328 USC § 1227(a)(2)(B).
• 338 USC § 1227(a)(2)(C).
• 348 USC § 1227(a)(2)(E).
• 35See 8 USC § 1227(a)(2)(F) (referencing 8 USC § 1182(a)(2)(H)).
• 368 USC § 1229a(c)(4).
• 378 USC § 1229a(c)(4)(A).
• 38Wanjiru v Holder, 705 F3d 258, 260 (7th Cir 2013).
• 39Id at 262.
• 40Id at 265–67.
• 41For a discussion of the importance of medical testimony in proving CAT claims, see, for example, Katherine J. Eder, Comment, The Importance of Medical Testimony in Removal Hearings for Torture Victims, 7 DePaul J Health Care L 281, 304 (2004) (emphasizing the importance of medical testimony given that “[m]edical and psychological evidence may corroborate the applicant’s claim where other evidence is unavailable”).
• 42See 8 CFR § 208.16(c)(2).
• 43Veena Reddy, Note, Judicial Review of Final Orders of Removal in the Wake of the REAL ID Act, 69 Ohio St L J 557, 559–60 (2008), citing Seeking Asylum in the U.S.? Choose Your Judge Carefully (Wash Post, Aug 1, 2006), archived at http://perma.cc/79LF-VKZV.
• 44Immigration Judges (TRAC Reports, July 31, 2006), archived at http://perma.cc/BM9A-5FXS.
• 45But see Reddy, Note, 69 Ohio St L J at 575 (cited in note 43), quoting Xiao Ji Chen v United States Department of Justice, 434 F3d 144, 156 (2d Cir 2006) (“[A]ppellate courts have generally circumscribed their scope of review . . . because these determinations ‘require intensive factual inquiries that appellate courts are ill-suited to conduct.’”).
• 46An applicant seeking asylum “must establish that race, religion, nationality, membership in a particular social group, or political opinion was or will be at least one central reason for persecuting the applicant.” 8 USC § 1158(b)(1)(B)(i).
• 47See Agha v Holder, 743 F3d 609, 615 (8th Cir 2014).
• 48Id.
• 49Note that withholding of removal under asylum is distinct from withholding of removal under the CAT. Compare 8 CFR § 1208.16(b) with 8 CFR § 1208.16(c).
• 508 CFR § 1208.17.
• 518 CFR § 1208.16(c)(2).
• 528 CFR § 1208.16(d)(2)–(3). See also Bobbie Marie Guerra, Comment, A Tortured Construction: The Illegal Immigration Reform and Immigrant Responsibility Act’s Express Bar Denying Criminal Aliens Withholding of Deportation Defies the Principles of International Law, 28 St Mary’s L J 941, 984 (1997) (recommending that the statute be amended to provide for case-by-case review of eligibility for withholding rather than an outright bar for noncitizens with criminal convictions).
• 53See 8 CFR § 1208.17.
• 548 CFR § 1208.16(f). See also Elizabeth Eschbach, Note, The Refugee Convention and the Convention against Torture: Failures of China and the United States, 13 Wash U Global Stud L Rev 353, 368 (2014) (noting that there is no international consensus on what counts as “effective protection” by a third country and that there is no funding mechanism to support the third countries).
• 558 CFR § 1208.24(b)(1).
• 56See 8 CFR § 1208.17(c) (noting that deferral does not “alter the authority of the Service to detain an alien”); Wanjiru, 705 F3d at 267 (explaining that because those who receive deferral of removal can be kept in US custody until it is safe to be returned, applicants are essentially saying they would “rather live in a U.S. jail than risk return” to a dangerous country).
• 58Ortiz-Franco v Holder, 782 F3d 81, 87 (2d Cir 2015), quoting Pierre v Gonzales, 502 F3d 109, 114 (2d Cir 2007) (noting that “[t]he CAT, however, ‘is not self-executing; by its own force, it confers no judicially enforceable right on individuals’”).
• 59CAT, 1465 UNTS at 85, 113.
• 60CAT, 1465 UNTS at 114.
• 61CAT, 1465 UNTS at 114.  
• 63Convention against Torture and Other Cruel, Inhuman or Degrading Treatment or Punishment, S Rep No 101-30, 101st Cong, 2d Sess 12 (1990) (noting that Articles 1 through 16 of the CAT are not self-executing, although they are “consonant with U.S. law”). See also Andrea Montavon-McKillip, CAT among Pigeons: The Convention against Torture, a Precarious Intersection between International Human Rights Law and U.S. Immigration Law, 44 Ariz L Rev 247, 251 & n 29 (2002) (explaining that US immigration authorities refused to enforce Article 3 of the CAT until domestic implementing legislation was passed and citing In re H-M-V-, 22 I&N Dec 256 (BIA 1998), as an example of this policy in action).
• 64Medellín v Texas, 552 US 491, 505 n 2 (2008) (noting ambiguities in the meaning of these terms, but defining a “self-executing” treaty as a treaty with “automatic domestic effect as federal law upon ratification” and a “non-self-executing” treaty as a treaty that “does not by itself give rise to domestically enforceable federal law . . . [but requires] implementing legislation passed by Congress”).
• 65RUDs are commonly used in ratifying multilateral human rights treaties, allowing individual countries to express any dissent or disagreement about provisions of the treaty without jeopardizing the overall adoption of the treaty. See Eric Neumayer, Qualified Ratification: Explaining Reservations to International Human Rights Treaties, 36 J Legal Stud 397, 397–98, 420–21 (2007) (noting that RUDs are common in human rights treaties, and finding that liberal democracies use RUDs more frequently than other countries, providing evidence to suggest that those countries do not think RUDs are “damaging to the international human rights regime”).
• 66S Rep 101-30 at 10 (cited in note 63). This RUD is explicitly based on prior case law regarding withholding of removal in the asylum context. See id, citing INS v Stevic, 467 US 407, 424 (1984) (noting that the clear-probability-of-persecution standard means the government cannot deport any individual who is “more likely than not” going to face persecution).  
• 67See Montavon-McKillip, 44 Ariz L Rev at 260 (cited in note 63) (noting that the “more likely than not” standard “is arguably higher than that generally accepted by the international community”).
• 68Pub L No 104-208, 110 Stat 3009-546, codified in various sections of Title 8.
• 698 USC § 1252(a)(2)(C). The specified criminal provisions are 8 USC § 1182(a)(2) (including, among other crimes, crimes of moral turpitude and violations of controlled-substance laws), 8 USC § 1227(a)(2)(A)(iii), (B), (C), or (D) (including aggravated felonies, controlled-substance violations, firearms offenses, and other miscellaneous crimes), and 8 USC § 1227(a)(2)(A)(ii) (including multiple criminal convictions for crimes of moral turpitude covered under 8 USC § 1227(a)(2)(A)(i) regardless of their date of commission).
• 70The CAT was opened for signatures in December 1984, and the United States signed the treaty on April 18, 1988. See Garcia, The U.N. Convention against Torture at *1, 3 (cited in note 62). On October 21, 1994, the United States ratified the CAT. Id. The CAT was implemented domestically on October 21, 1998.
• 71Pub L No 105-277, 112 Stat 2681-761, codified at 8 USC § 1231.
• 728 USC § 1231 (note).
• 73See note 60 and accompanying text.
• 748 USC § 1231 (note).
• 75See Department of Justice, Asylum and Withholding of Removal Relief at *3 (cited in note 57) (describing the difference between “affirmative” and “defensive” asylum claims).
• 76Pub L No 109-13, 119 Stat 302, codified at 8 USC § 1101 et seq.
• 778 USC § 1252(a)(2)(D). REAL ID also created 8 USC § 1252(a)(4), which provides that appellate review is the only means for reviewing CAT claims. This was done to eliminate the option of habeas corpus as an alternate means of relief. See David M. McConnell, Judicial Review under the Immigration and Nationality Act: Habeas Corpus and the Coming of REAL ID (1996-2005), 51 NY L Sch L Rev 75, 111 (2006) (arguing that in REAL ID, Congress sought to “reassert the supremacy of the INA’s judicial review scheme” as opposed to habeas corpus).
• 788 CFR § 1208.16(c)(3).  
• 798 CFR § 1208.18(a)(1)–(2). See also Irene Scharf, Un-torturing the Definition of Torture and Employing the Rule of Immigration Lenity, 66 Rutgers L Rev 1, 12–14 (2013) (describing criticism of the requirement of specific intent in the definition of torture and arguing for a rule of immigration lenity to “ameliorate the harsh effects of these statutes”); Jansen Averett, Note, Addressing an Alien’s Fears of Torture under the Convention against Torture, 36 NC J Intl L & Comm Reg 471, 496–97 (2011) (arguing that courts should “more seriously consider the threat of torture and human rights violations in light of the current political climate”); In re J-E-, 23 I&N Dec 291, 298–301 (BIA 2002) (denying CAT relief to an applicant who had not proved the supposed torture would “be specifically intended to inflict severe physical or mental pain or suffering”).
• 80417 F3d 642 (7th Cir 2005).
• 81Id at 645.
• 82Id.
• 83Id (quotation marks omitted). The Seventh Circuit has since abandoned this view, and now holds that appellate courts do have jurisdiction to review questions of fact. See Part II.B.2.
• 84Hamid, 417 F3d at 645–47.
• 85Id at 647.
• 86See, for example, Gourdet v Holder, 587 F3d 1, 5 (1st Cir 2009) (explaining that the court “may not review the administrative fact findings of the IJ or the BIA as to the sufficiency of the alien’s evidence and the likelihood that the alien will be tortured if returned to the country in question”).
• 87See, for example, Escudero-Arciniega v Holder, 702 F3d 781, 783–85 (5th Cir 2012) (holding that the court can review whether certain criminal offenses count as aggravated felonies, but cannot review claims dealing with the burden of proof, as those are factual issues and therefore are barred).
• 88See, for example, Tran v Gonzales, 447 F3d 937, 943 (6th Cir 2006) (“Pursuant to § 1252(a)(2)(C) and (D), our review of Tran’s CAT claim is limited to questions of law or constitutional issues.”).
• 89574 F3d 990 (8th Cir 2009).
• 90Id at 992, 997.
• 91Id at 998.
• 92Id.
• 938 USC § 1252(a)(4).
• 94Lovan, 574 F3d at 998.
• 95 782 F3d 81 (2d Cir 2015).
• 96Id at 83–84.
• 97Id at 84–85.
• 98Id at 85.
• 99Ortiz-Franco, 782 F3d at 88–89 (quotation marks omitted).
• 100Id at 89.
• 101Id. This argument is explored again in the discussion of Wanjiru v Holder, 705 F3d 258, in Part II.B.2.
• 102Ortiz-Franco, 782 F3d at 89.
• 103Id.
• 104The court here noted that a denial of deferral of removal is a final order of removal. But that is a one-way street—if deferral of removal is granted, then there is no final order of removal. Thus, while a denial of deferral of removal can be final, deferral of removal itself is not a final order.
• 105Ortiz-Franco, 782 F3d at 90.
• 106Id.
• 107500 F3d 1315 (11th Cir 2007).
• 108Id at 1317.
• 109Id at 1317–19.
• 110Id at 1320.
• 111Jean-Pierre, 500 F3d at 1321.
• 112Id at 1322, 1326–27.
• 113See id at 1321.
• 114Id, citing Cadet v Bulger, 377 F3d 1173, 1184 (11th Cir 2004).
• 115Jean-Pierre, 500 F3d at 1321. See also Alexandre v United States Attorney General, 452 F3d 1204, 1206 (11th Cir 2006) (holding that the current scope of review for removal orders is the same as in habeas cases).
• 116Jean-Pierre, 500 F3d at 1322. See also generally Rebecca Sharpless, Fitting the Formula for Judicial Review: The Law-Fact Distinction in Immigration Law, 5 Intercultural Hum Rts L Rev 57 (2010). Professor Rebecca Sharpless argues that the jurisdiction savings clause, § 1252(a)(2)(D), “presupposes a simple image—one that clearly delineates fact from law,” but that this is not how most questions present themselves in the real world. Sharpless, 5 Intercultural Hum Rts L Rev at 66 (cited in note 116). Sharpless points to the question “whether certain types of mistreatment (the established facts) rise to the level of ‘torture’ within the meaning of Article 3 in the Convention Against Torture (the rule)” as an example of a mixed fact-and-law question. Id at 69. Sharpless encourages litigants to frame more questions as mixed (and she thinks most truly are) as a way to “maximize review[ ]” but still allow federal court to “abide by their constitutional Article III mandate.” Id at 88.
• 117Jean-Pierre, 500 F3d at 1322, citing Conference Report, HR Rep No 109-72, 109th Cong, 1st Sess 175 (2005), reprinted in 2005 USCCAN 240, 300.
• 118See Jean-Pierre, 500 F3d at 1325–26 (noting that the BIA did not address Jean-Pierre’s claims that Haitian guards would strike him on the head and ears and confine him in a small crawl space).
• 119Id at 1327.
• 120671 F3d 303 (3d Cir 2011).
• 121Id at 309 n 6, citing Kaplun v Attorney General of the United States, 602 F3d 260, 271 (3d Cir 2010). While the court never specifically calls this a “mixed question,” its framing of the question is: “[D]o the facts found by the IJ . . . meet the legal requirements for relief under the CAT?” Pieschacon–Villegas, 671 F3d at 309 n 6, quoting Kaplun, 602 F3d at 271. This type of question has been called “the basic formula for a mixed question.” Sharpless, 5 Intercultural Hum Rts L Rev at 74 (cited in note 116) (describing the “basic formula” as, “Do established facts A–B satisfy rule X?”) (emphasis omitted).
• 122Pieschacon–Villegas, 671 F3d at 309.
• 123Id at 310. The court noted that the BIA was allowed to hold that a plaintiff’s facts did not meet the burden of proof. It could not, however, “ignore this evidence altogether.” Id.
• 124516 F3d 243 (4th Cir 2008).
• 125Id at 246.
• 126Id at 247.
• 127Id at 247–48.
• 128Saintha, 516 F3d at 250 (quotation marks omitted and brackets in original).
• 129Id at 249–50.
• 130Id at 251.
• 131Id at 250. But see Sharpless, 5 Intercultural Hum Rts L Rev at 87–88 (cited in note 116) (arguing that there is almost nothing that is a pure question of fact).
• 1328 USC § 1252(a)(2)(C) (emphasis added).
• 133478 F3d 972 (9th Cir 2007).
• 134Id at 980 (citations omitted).
• 135Id at 981.
• 136518 F3d 1081 (9th Cir 2008).
• 137Id at 1083.
• 138Id (emphases added).
• 139Id.
• 140Hamid, 417 F3d at 647. See also text accompanying notes 80–85.
• 141Hamid, 417 F3d at 647.
• 142617 F3d 962 (7th Cir 2010).
• 143Id at 964–65.
• 144See id at 969–70, citing 8 USC § 1101(a)(42). In defining “refugee,” the statute specifically excludes “any person who ordered, incited, assisted, or otherwise participated in the persecution of any person on account of race, religion, nationality, membership in a particular social group, or political opinion.” 8 USC § 1101(a)(42).
• 145555 US 511 (2009).
• 146See id at 541 (Thomas dissenting). The persecutor bar prevents noncitizen persecutors from receiving refugee status, while § 1252(a)(2)(C) does not prevent noncitizen criminals from receiving CAT protection, it just conscribes their appellate review.
• 147Issaq, 617 F3d at 970.
• 148Id.
• 149Id. See also id at 970–71 (Ripple concurring) (noting that “this question is not squarely presented in the case before us and, therefore, need not be decided at this time,” and criticizing the majority for not addressing other circuit court decisions, either favorable or unfavorable).
• 150Id at 970.
• 151705 F3d 258 (7th Cir 2013).
• 152Id at 263–65. See also Lenjinac v Holder, 780 F3d 852, 855 (7th Cir 2015) (confirming the court’s holding in Wanjiru and reviewing questions of fact raised by the plaintiff).
• 153Wanjiru, 705 F3d at 261.  
• 154Id.
• 155Id at 260.
• 156Id.
• 157Wanjiru, 705 F3d at 261. Wanjiru believed this because his cousin was captured by Kenyan police and forced to reveal information about the Mungiki. His cousin initially fled Kenya but later decided to return; he was ultimately found murdered. Wanjiru believed that the Mungiki were responsible for the murder. Id.
• 158Id at 262–63. Wanjiru had a misdemeanor conviction for sexual misconduct. Id at 261. 8 USC § 1182(a)(2) and 8 USC § 1227(a)(2)(A)(ii) both cover crimes of moral turpitude, which sexual misconduct falls under. Wanjiru, 705 F3d at 262–63. But § 1182(a)(2) covers only noncitizens “seeking admission” to the United States, and § 1227(a)(2)(A)(ii) requires multiple convictions for crimes of moral turpitude. See Wanjiru, 705 F3d at 263. Because Wanjiru did not fall within these categories, he was not subject to the criminal conviction bar in § 1252(a)(2)(C). See Wanjiru, 705 F3d at 263.
• 159Wanjiru, 705 F3d at 263.
• 160Id at 264.
• 161Wanjiru, 705 F3d at 264.
• 162See Ortiz-Franco, 782 F3d at 89. See also text accompanying notes 101–04.
• 163Wanjiru, 705 F3d at 264. The court also pointed to its holding in Calma v Holder, 663 F3d 868 (7th Cir 2011), in which it decided that an appellate court could sometimes review an IJ’s denial of a continuance, even when the final BIA decision might be unreviewable. See Calma, 663 F3d at 876–77. Because this denial of a continuance can also be the “dispositive ruling,” it is “final enough” for the court to review. Wanjiru, 705 F3d at 264.
• 164Wanjiru, 705 F3d at 264.
• 165Id at 265. The court specifically stated that “canons of construction favor statutory interpretations that preserve judicial review.” Id. See also Kucana v Holder, 558 US 233, 251–52 (2010) (noting the “well-settled” doctrine “favoring interpretations of statutes [to] allow judicial review of administrative action”) (brackets in original). Both sides of the circuit split have called on this canon. The Second Circuit came to the opposite conclusion because of concerns about eliminating final judgment review for all CAT claims. See Ortiz-Franco, 782 F3d at 89.
• 166See 28 USC § 1292(a)(1). Clearly this statute is not directly applicable because it refers only to decisions of district courts, and deferral-of-removal decisions come from IJs and the BIA. That does not mean, however, that it is not a useful and helpful analogy.
• 16728 USC § 1291 (“The courts of appeals . . . shall have jurisdiction of appeals from all final decisions of the district courts of the United States.”).
• 168See Michael E. Solimine, Revitalizing Interlocutory Appeals in the Federal Courts, 58 Geo Wash L Rev 1165, 1169 (1990) (noting that the classic example of an interlocutory appeal is the issuance of an injunction because “[t]he effect of trial court rulings before a final judgment may be irreparable even if reversed after a later appeal”).
• 16928 USC § 1292(a)(1). Appellate courts also have jurisdiction to review interlocutory orders relating to receiverships and admiralty cases. 28 USC § 1292(a)(2)–(3). Finally, appellate courts may review certain questions of law when a district judge certifies that the question is one over which there is much disagreement and an early resolution could help prevent later appeals and lead to a quicker end to the lawsuit. 28 USC § 1292(b).
• 170See Timothy Stoltzfus Jost, From Swift to Stotts and Beyond: Modification of Injunctions in the Federal Courts, 64 Tex L Rev 1101, 1101 (1986).
• 171See John Leubsdorf, The Standard for Preliminary Injunctions, 91 Harv L Rev 525, 525 (1978).
• 172See Joshua P. Davis, Taking Uncertainty Seriously: Revising Injunction Doctrine, 34 Rutgers L J 363, 369 (2003).
• 1738 CFR § 1208.17(d)(1).
• 174See 8 CFR § 1208.17(f). See also 8 CFR § 1208.18(c) (requiring that the attorney general “determine, in consultation with the Secretary of State, whether the assurances are sufficiently reliable to allow the alien’s removal” under the CAT).
• 175FRCP 60. This rule is generally seen as a codification of United States v Swift & Co, 286 US 106, 114 (1932) (“A continuing decree of injunction directed to events to come is subject always to adaptation as events may shape the need.”). See also Jost, 64 Tex L Rev at 1105 (cited in note 170).
• 176490 US 495 (1989).
• 177Id at 498, quoting Midland Asphalt Corp v United States, 489 US 794, 798 (1989).
• 178Lauro Lines, 490 US at 498 (quotation marks omitted).
• 179Id.
• 180For a discussion of other rights that are “effectively unreviewable,” see id at 499–500 (describing the right not to be tried and claims affecting qualified immunity as rights which are effectively unreviewable on appeal).
• 181See, for example, Winter v Natural Resources Defense Council, Inc, 555 US 7, 12 (2008) (reviewing the grant of a preliminary injunction restricting the Navy’s use of sonar training, which the trial court had put in place while continuing to review substantive claims under the National Environmental Policy Act and other environmental statutes).
• 1828 USC § 1231 (note).
• 183Black’s Law Dictionary 938 (West 10th ed 2014) (defining “interlocutory” as “interim or temporary; not constituting a final resolution of the whole controversy”).
• 184See 28 USC § 1291 (“The courts of appeals . . . shall have jurisdiction of appeals from all final decisions of the district courts of the United States.”). There have been many critiques of the final judgment rule. See, for example, Robert J. Martineau, Defining Finality and Appealability by Court Rule: Right Problem, Wrong Solution, 54 U Pitt L Rev 717, 788 (1993) (exploring various reforms to the final judgment rule, including an expansion of the term “final,” but concluding against that option, as it would only increase litigation); Carleton M. Crick, The Final Judgment as a Basis for Appeal, 41 Yale L J 539, 564 (1932) (arguing that the best solution to the final judgment rule is to allow appeals to be taken at the discretion of the appellate courts).  
• 185Thomas A. Engelhardt, Bringing More Finality to Finality: Conditional Consent Judgments and Appellate Review, 89 St John’s L Rev 293, 301 (2015).
• 186See Bernard T. Shen, Comment, From Jail Cell to Cellular Communication: Should the Rufo Standard Be Applied to Antitrust and Commercial Consent Decrees?, 90 Nw U L Rev 1781, 1786 (1996).  
• 187Id at 1788.
• 188Id.
• 189502 US 367 (1992).
• 190Id at 391.
• 191Id at 371–72. The appellate court had found the jail conditions unconstitutional, and ordered the jail to close unless plans for a new facility were submitted. They were, and the trial court issued a formal consent decree in which the government pledged to erect a better shelter. Id at 374–75.
• 192286 US 106 (1932).
• 193Rufo, 502 US at 393.
• 194Swift, 286 US at 119.
• 195Rufo, 502 US at 379 (quotation marks omitted).
• 196Id at 383.
• 197Id at 384.
• 198Pub L No 104-134, 110 Stat 1321-66 (1996).
• 19918 USC § 3626(a)(1)(A), (c)(1).
• 200See Theodore K. Cheng, Invading an Article III Court’s Inherent Equitable Powers: Separation of Powers and the Immediate Termination Provisions of the Prison Litigation Reform Act, 56 Wash & Lee L Rev 969, 972 (1999) (noting that “[p]rison conditions lawsuits” usually seek relief for “due process violations, [ ] equal protection violations, or violations of the prohibition against cruel and unusual punishment”).
• 201See id at 979 (quoting Senator Robert Dole as supporting the PLRA as a way to “restrain liberal Federal judges who see violations of constitutional rights in every prisoner complaint and who have used these complaints to micromanage State and local prison systems”).
• 202See 18 USC § 3626(b)(1)(A)(iii), (b)(2) (creating rules for the termination of orders “issued on or before the date of enactment of the [PLRA]”).
• 203See Jennifer A. Puplava, Note, Peanut Butter and Politics: An Evaluation of the Separation-of-Powers Issues in Section 802 of the Prison Litigation Reform Act, 73 Ind L J 329, 333–39 (1997); Thomas Julian Butler, Commentary, The Prison Litigation Reform Act: A Separation of Powers Dilemma, 50 Ala L Rev 585, 586–88 (1997).
• 204Hadix v Johnson, 947 F Supp 1100, 1110–13 (ED Mich 1996), revd, 133 F3d 940 (6th Cir 1998).
• 205100 F3d 365 (4th Cir 1996). See also, for example, Benjamin v Jacobson, 172 F3d 144, 162 (2d Cir 1999) (en banc) (rejecting a separation-of-powers challenge to the PLRA because it involved modifying only injunctions, rather than money damages); Gavin v Branstad, 122 F3d 1081, 1087–89 (8th Cir 1997) (rejecting a separation-of-powers challenge to the PLRA based on the injunction analogy).
• 206See Plyler, 100 F3d at 372.
• 207Id at 371–72.
• 208Elizabeth Warren, “Why Have a Federal Bankruptcy System?”, 77 Cornell L Rev 1093, 1094 (1992) (noting that generally government tax collectors and employees are paid before business creditors).
• 209See 28 USC § 158(a). The statute governing bankruptcy appeals uses language very similar to the statute governing general federal appellate jurisdiction. Compare 28 USC § 158(a) with 28 USC § 1291.
• 21028 USC § 158(a)(1).
• 21128 USC § 158(a)(2)–(3).
• 212Karen K. Brown, Impact of the Final Judgment Rule on Bankruptcy Appeals, 38 Judges’ J 13, 13 (Winter 1999).
• 213811 F2d 1112 (7th Cir 1987).
• 214Id at 1116.
• 215Id.
• 216433 Bankr 763 (D Hawaii 2010).
• 217Id at 773–74.
• 218Id at 773 (emphasis omitted).
• 219Id at 774–75.
• 220697 F3d 74 (2d Cir 2012).
• 221Id at 77.
• 222See Charles Alan Wright, et al, 16 Federal Practice and Procedure: Jurisdiction and Related Matters § 3926.2 (West 2016 Supp) (“Virtually all decisions agree that the concept of finality applied to appeals in bankruptcy is broader and more flexible than the concept applied in ordinary civil litigation.”). See also, for example, In re Atlas IT Export Corp, 761 F3d 177, 181–82 (1st Cir 2014) (noting that “because a bankruptcy case is quite often a conglomeration of separate cases that lives on for many years, we take a flexible approach to finality here, giving that requirement a ‘practical’ rather than a ‘technical’ construction,” and that “‘final’ includes an order that decides all—repeat, all—the issues of a discrete dispute within a larger case”) (quotation marks omitted); Matter of AFI Holding, Inc, 530 F3d 832, 836 (9th Cir 2008), quoting In re Lazar, 237 F3d 967, 985 (9th Cir 2001) (noting that the Ninth Circuit has “adopted a pragmatic approach to finality in bankruptcy cases”); Grovenburg v Homestead Insurance Co, 183 F3d 883, 885 (8th Cir 1999) (noting that “[i]n bankruptcy proceedings, we apply a more liberal standard of finality”).
• 22335 USC § 131.
• 22435 USC § 134.
• 22535 USC § 141(a).
• 226Flexiteek Americas, Inc v PlasTEAK, Inc, 2012 WL 5364263, *8 (SD Fla), citing 35 USC § 302.
• 227See Flexiteek, 2012 WL 5364263 at *9. The Leahy-Smith America Invents Act of 2011, Pub L No 112-29, 125 Stat 284, did create new means of challenging patents at the PTO. However, the reexamination process still exists and the Federal Circuit still retains jurisdiction over appeals, so this Act does not affect this analysis much.
• 228721 F3d 1330 (Fed Cir 2013) (“Fresenius II”).
• 229Id at 1332.
• 230Id at 1332–33.
• 231Id at 1333. See also Fresenius USA, Inc v Baxter International, Inc, 582 F3d 1288, 1304 (Fed Cir 2009) (“Fresenius I”).
• 232Fresenius II, 721 F3d at 1333–34.
• 233Id at 1334.
• 234Fresenius II, 721 F3d at 1334.
• 235In re Baxter International, Inc, 678 F3d 1357, 1366 (Fed Cir 2012).
• 236Fresenius II, 721 F3d at 1340.
• 237Id at 1341.
• 238Id.
• 239Id at 1345, citing Plaut v Spendthrift Farm, Inc, 514 US 211, 218–19 (1995).
• 240Fresenius II, 721 F3d at 1345–46. See also ePlus, Inc v Lawson Software, Inc, 789 F3d 1349, 1358–62 (Fed Cir 2015) (applying the holding of Fresenius II to vacate an injunction and civil contempt order which had been based on a finding of infringement of a patent that had been subsequently invalidated by PTO reexamination).
• 241See Shashank Upadhye and Adam Sussman, A Real Separation of Powers or Separation of Law: Can an Article I Administrative Agency Nullify an Article III Federal Court Judgment?, 25 Fordham Intel Prop Media & Enter L J 1, 5 (2014) (“[B]ecause a patent infringement verdict necessarily involves an issued and valid patent, once the infringer loses, can he get a ‘do-over’ by expunging or impeaching the patent back at the PTO?”).
• 242In the determination of whether flexible finality should apply to other areas of law, application of the test this Comment establishes is the first step. Beyond that, it will also be important to consider the language of the relevant statute, as well as its legislative history, to ensure that neither explicitly or implicitly prevents the application of flexible finality.
• 243See, for example, Wanjiru, 705 F3d at 265; Sosnovskaia v Gonzales, 421 F3d 589, 592 (7th Cir 2005) (stating that administrative findings must be “supported by reasonable, substantial, and probative evidence on the record [considered] as a whole”) (brackets in original); 8 USC § 1252(b)(4)(B) (“[T]he administrative findings of fact are conclusive unless any reasonable adjudicator would be compelled to conclude to the contrary.”).
• 244Evelyn H. Cruz, Double the Injustice, Twice the Harm: The Impact of the Board of Immigration Appeals’s Summary Affirmance Procedures, 16 Stan L & Pol Rev 481, 507 (2005).
• 245See, for example, Deborah E. Anker, Determining Asylum Claims in the United States: Summary Report of an Empirical Study of the Adjudication of Asylum Claims before the Immigration Court, 2 Intl J Refugee L 252, 257–58 (1990) (noting problems ranging from translation difficulties to a lack of cultural sensitivity).
• 246Lindsey R. Vaala, Note, Bias on the Bench: Raising the Bar for U.S. Immigration Judges to Ensure Equality for Asylum Seekers, 49 Wm & Mary L Rev 1011, 1024 (2007).
• 247Adam Liptak, Courts Criticize Judges’ Handling of Asylum Cases (NY Times, Dec 26, 2005), online at http://www.nytimes.com/2005/12/26/us/courts-criticize-judges
  -handling-of-asylum-cases.html (visited July 10, 2016) (Perma archive unavailable), quoting Benslimane v Gonzales, 430 F3d 828, 830 (7th Cir 2005).
• 248Liptak, Courts Criticize Judges’ Handling of Asylum Cases (cited in note 247), quoting Wang v Attorney General of the United States, 423 F3d 260, 269 (3d Cir 2005).
• 249Liptak, Courts Criticize Judges’ Handling of Asylum Cases (cited in note 247), quoting Recinos de Leon v Gonzales, 400 F3d 1185, 1187 (9th Cir 2005).
• 250Zehatye v Gonzales, 453 F3d 1182, 1195 (9th Cir 2006) (Berzon dissenting).
• 251See Rachel L. Swarns, Study Finds Disparities in Judges’ Asylum Rulings (NY Times, July 31, 2006), online at http://www.nytimes.com/2006/07/31/us/31asylum.html (visited July 10, 2016) (Perma archive unavailable) (noting that one Miami IJ denied asylum 96.7 percent of the time if the applicant had a lawyer, while a New York IJ denied asylum only 9.8 percent of the time in such cases).
• 252See Richard H. Fallon Jr, Of Legislative Courts, Administrative Agencies, and Article III, 101 Harv L Rev 915, 987–88 (1988) (noting that “fairness and integrity values could, in some cases, make [factual] review mandatory” and that “a crucially relevant factor involves the administrative safeguards available to ensure the fairness of an agency’s adjudication”).
• 253For a debate on the constitutionality of jurisdiction stripping, contrast Akhil Reed Amar, A Neo-federalist View of Article III: Separating the Two Tiers of Federal Jurisdiction, 65 BU L Rev 205, 206–07 (1985), with John Harrison, The Power of Congress to Limit the Jurisdiction of Federal Courts and the Text of Article III, 64 U Chi L Rev 203, 252 (1997) (arguing that the Constitution “does not imply Amar’s mandatory version of the principle of coextensiveness”).
• 254Richard H. Fallon Jr, Jurisdiction-Stripping Reconsidered, 96 Va L Rev 1043, 1053 (2010), citing Boumediene v Bush, 553 US 723, 732 (2008).
• 255Fallon, 96 Va L Rev at 1053 (cited in note 254).
• 256553 US 723 (2008).
• 257Fallon, 96 Va L Rev at 1053 (cited in note 254).
• 258Id.
• 259Id at 1061, quoting Boumediene, 553 US at 765, quoting Marbury v Madison, 5 US (1 Cranch) 137, 177 (1803) (brackets in original, emphasis added, and quotation marks omitted).
• 260See Fallon, 96 Va L Rev at 1135 (cited in note 254) (noting that “[e]ven when initial adjudication by a non-Article III tribunal is permissible, the Constitution may mandate the availability of either appellate review or some other mode of access to an Article III court” and pointing to the Due Process Clause as one such source of the constitutional mandate). See also Landon v Plasencia, 459 US 21, 32 (1982) (noting that while a noncitizen “seeking initial admission to the United States . . . has no constitutional rights regarding his application,” once a noncitizen has been admitted, “his constitutional status changes accordingly” and, at least in some cases, such a noncitizen “can invoke the Due Process Clause”).
• 261See Martineau, 54 U Pitt L Rev at 771 (cited in note 184).
• 262Nancy Levit, The Caseload Conundrum, Constitutional Restraint and the Manipulation of Jurisdiction, 64 Notre Dame L Rev 321, 363 (1989).
• 263132 S Ct 2566 (2012) (“NFIB”).
• 264Id at 2584.
• 265See id at 2582. See also 26 USC § 7421(a).
• 266NFIB, 132 S Ct at 2583.
• 267Id at 2593–94.
• 268See id at 2594–95.
• 2696 US (2 Cranch) 64 (1804).
• 270Scharf, 66 Rutgers L Rev at 25 (cited in note 79). See also United States v Gradwell, 243 US 476, 485 (1917) (“[B]efore a man can be punished as a criminal under the federal law his case must be plainly and unmistakably within the provisions of some statute.”) (quotation marks omitted).
• 271333 US 6 (1948).
• 272Id at 10.
• 273467 US 837, 866 (1984) (holding that courts must defer to an agency’s interpretation of a statute when the statute is ambiguous and the interpretation proposed is reasonable).
• 274Compare Brian G. Slocum, The Immigration Rule of Lenity and Chevron Deference, 17 Georgetown Immig L J 515, 577 (2003) (arguing that lenity should be “one factor among several that courts use to determine the reasonableness of an agency’s interpretation of a statute”), with David S. Rubenstein, Putting the Immigration Rule of Lenity in Its Proper Place: A Tool of Last Resort after Chevron, 59 Admin L Rev 479, 504 (2007) (arguing that lenity should be considered only “after the court determines both that the statute is ambiguous . . . and that the agency’s interpretation is unreasonable”).
• 275Slocum, 17 Georgetown Immig L J at 530 (cited in note 274), citing INS v Aguirre-Aguirre, 526 US 415, 424–25 (1999).
• 276INS v Orlando Ventura, 537 US 12, 17 (2002).
• 277See Part II.B.
• 278Charming Betsy, 6 US (2 Cranch) at 118. The Seventh Circuit, in fact, noted this canon in support of its Wanjiru decision. Wanjiru, 705 F3d at 265.
• 279Charming Betsy, 6 US (2 Cranch) at 118.
• 280See Restatement (Third) of the Foreign Relations Law of the United States § 114 (1987); Rebecca Crootof, Note, Judicious Influence: Non-self-executing Treaties and the Charming Betsy Canon, 120 Yale L J 1784, 1793 (2011) (noting that “[a]s currently formulated, the canon . . . may be used only when a statute is ambiguous”).
• 281See, for example, Crootof, Note, 120 Yale L J at 1819 (cited in note 280) (arguing that applying Charming Betsy to non-self-executing treaties “results in relatively costless compliance with international law”).  
• 282Chae Chan Ping v United States, 130 US 581, 600 (1889) (holding that “the last expression of the sovereign will must control”).
• 283Crootof, Note, 120 Yale L J at 1801 (cited in note 280).
• 284See Part II.B.2.
• 285See text accompanying note 160.
• 286See text accompanying note 161.
• 287Edward Cantu, Posner’s Pragmatism and the Turn toward Fidelity, 16 Lewis & Clark L Rev 69, 70 (2012).
• 288See Jonathan T. Molot, Ambivalence about Formalism, 93 Va L Rev 1, 2–3 (2007) (noting that “scholars focus on the relationship between Congress as principal and the judiciary as agent . . . [and] have gravitated toward formalism in coping with this [principal-agent] problem”).
• 289See James G. Wilson, The Morality of Formalism, 33 UCLA L Rev 431, 436 (1985) (arguing that bright-line, formal rules can be “rigid, potentially arbitrary, and frequently ignore relevant aspects of reality”).
• 290See Karl N. Llewellyn, Remarks on the Theory of Appellate Decision and the Rules or Canons about How Statutes Are to Be Construed, 3 Vand L Rev 395, 401–06 (1950) (noting that each formalist canon of construction has an opposing canon).
• 291Daniel A. Farber, The Inevitability of Practical Reason: Statutes, Formalism, and the Rule of Law, 45 Vand L Rev 533, 535 (1992), citing Llewellyn, 3 Vand L Rev at 399 (cited in note 290) (arguing that even though Llewellyn attacked the formalistic interpretative canons, he still thought that judges have a duty “to make sense, under and within the law,” rather than just “to announce what [they] believe[ ] to be the just outcome on the facts of a particular case”).
• 292Farber, 45 Vand L Rev at 535 (cited in note 291), quoting Llewellyn, 3 Vand L Rev at 398 (cited in note 290).
• 293Wanjiru, 705 F3d at 264.