Aggregating Risk Under CAT: The Fourth Circuit Requires the BIA to Meaningfully Address All Theories and Evidence of Torture — McDougall v. Bondi

Introduction

In Marlon McDougall v. Pamela Bondi, the Fourth Circuit granted a petition for review, vacated the decision of the Board of Immigration Appeals (BIA), and remanded for further proceedings after finding that the BIA abused its discretion in adjudicating a Convention Against Torture (CAT) claim. The petitioner, a Black Guyanese national and longtime U.S. lawful permanent resident diagnosed with schizophrenia and significant physical disabilities, sought CAT deferral based on a feared risk of torture in Guyana for multiple overlapping reasons: his mental illness, his physical disabilities, his criminal history and deportee status, and his race.

Both the Immigration Judge (IJ) and the BIA addressed only the mental-illness aspect of his claim and ignored substantial evidence supporting the other asserted bases. The Fourth Circuit’s published decision reinforces and sharpens key CAT adjudication requirements: adjudicators must consider the “full panoply” of risk-of-torture evidence, meaningfully address each asserted theory of harm, and aggregate the probabilities across all reasons and potential torturers. Boilerplate assertions that the agency “considered all evidence” will not suffice.

Summary of the Judgment

• The Fourth Circuit held that the BIA abused its discretion by disregarding important aspects of McDougall’s CAT claim—specifically, the record evidence and arguments that he faces a heightened risk of torture in Guyana on account of his race (as a Black man), his physical disabilities, and his status as a criminal deportee.
• Reaffirming its aggregation doctrine, the court emphasized that CAT relief is warranted if the cumulative probability of torture from all sources or for all reasons exceeds 50%.
• Because the BIA addressed only mental illness and ignored other nonfrivolous theories and supporting evidence, the court vacated and remanded without reaching McDougall’s separate argument that the BIA misapplied the regulatory definition of “torture” (including “specific intent” and acquiescence).
• The panel also noted that, under Supreme Court authority, the petition’s timeliness under 8 U.S.C. § 1252(b)(1) is nonjurisdictional; the government conceded timeliness and the case proceeded on the merits.

Analysis

Precedents Cited and Their Influence

• Garcia v. Garland, 73 F.4th 219 (4th Cir. 2023): Requires the agency to consider the “full panoply of the risk-of-torture evidence,” including multiple sources and reasons for torture. Garcia supplies the foundational duty that the BIA breached here by discussing only one theory (mental illness).
• Kouyate v. Garland, 122 F.4th 132 (4th Cir. 2024): Clarifies that an applicant is entitled to CAT protection if “the cumulative probability of torture by all entities, or for all reasons, exceeds 50%.” Kouyate underwrites the Fourth Circuit’s insistence on probability aggregation across all asserted risks in McDougall’s case.
• Portillo Flores v. Garland, 3 F.4th 615 (4th Cir. 2021) (en banc): Holds that the BIA abuses its discretion when it fails to provide a reasoned explanation, ignores important aspects of a claim, arbitrarily disregards relevant evidence, or commits legal error. Portillo Flores frames the core abuse-of-discretion analysis applied to the BIA’s opinion.
• Rodriguez-Arias v. Whitaker, 915 F.3d 968 (4th Cir. 2019): Reversal is warranted when the agency “arbitrarily ignore[s] relevant evidence” or fails to offer a “cogent, articulable basis” for discounting it. That rule maps directly onto the BIA’s handling of McDougall’s race, disability, and deportee-status evidence.
• Lopez-Sorto v. Garland, 103 F.4th 242 (4th Cir. 2024): Confirms the standards of review in CAT cases (substantial evidence for facts, de novo for legal questions) and recognizes the reviewability of whether the correct legal standard was applied to determine the likelihood of torture. It also notes that while the agency needn’t discuss every piece of evidence, it must demonstrate meaningful consideration of the record.
• Nasrallah v. Barr, 590 U.S. 573 (2020): Confirms court of appeals jurisdiction to review both legal and factual challenges to CAT decisions; factual findings are reviewed for substantial evidence. It clears the way for the Fourth Circuit’s review of the BIA’s factual inferences and its legal framework.
• Turkson v. Holder, 667 F.3d 523 (4th Cir. 2012): Distinguishes factual predictions (“what will likely happen”) from legal determinations (whether the likely treatment amounts to torture). It informs the panel’s division of labor when evaluating the BIA’s conclusions.
• Orellana v. Barr, 925 F.3d 145 (4th Cir. 2019) and Tassi v. Holder, 660 F.3d 710 (4th Cir. 2011): Require the agency to address “important aspects” and to provide “specific, cogent reasons” for discounting credible evidence—standards the court finds unmet here.
• Wambura v. Barr, 980 F.3d 365 (4th Cir. 2020): When the BIA issues its own decision (without expressly adopting the IJ), the court reviews only the BIA’s opinion. That rule matters because the BIA’s opinion, not the IJ’s, failed to engage the non-mental-illness theories.
• Ai Hua Chen v. Holder, 742 F.3d 171 (4th Cir. 2014): Boilerplate statements asserting that the agency considered the evidence are insufficient if the analysis shows only perfunctory engagement. The court applies this principle to reject the government’s reliance on an IJ boilerplate assertion of “aggregate” consideration.
• Riley v. Bondi, 145 S. Ct. 2190 (2025): Section 1252(b)(1)’s filing deadline is nonjurisdictional. The panel notes this in accepting the government’s concession that McDougall’s petition is timely.

Legal Reasoning

The court begins by delineating the CAT framework and standards of review. An applicant must demonstrate that it is “more likely than not” they will be subjected to “torture” if removed—where “torture” requires intentional infliction of severe pain or suffering for a prohibited purpose, by or with the consent or acquiescence of a public official acting in an official capacity. The agency’s factual predictions about what will happen on removal are reviewed for substantial evidence; legal questions (including application of the likelihood-of-torture standard) are reviewed de novo. Separately, the BIA’s case-management decisions are reviewed for abuse of discretion, including when it fails to consider important aspects of a claim or arbitrarily ignores relevant evidence.

Against that framework, the panel identifies a central error: the BIA confined its analysis to the risk flowing from McDougall’s mental illness and repeatedly framed its conclusions around a lack of police training or mental health resources in Guyana. But the record—and the appeal—contained developed theories and evidence that McDougall faced a heightened risk of torture for at least three additional reasons:

• his race, as a Black person;
• his physical disabilities, including wheelchair use; and
• his status as a criminal deportee with media attention in Guyana and specific threats.

The panel emphasizes that these were not stray allegations; the record included testimony, affidavits, and country-condition evidence pointing to police and public violence against individuals with these characteristics, poor conditions in prisons and mental health facilities, and patterns of nonintervention by authorities when the public inflicted violence.

Under Garcia and Kouyate, the BIA had to (1) isolate each asserted source or reason for torture, (2) analyze the risk posed by each, and (3) aggregate those risks to determine whether, in sum, the probability that McDougall would be tortured exceeded 50%. Instead, the BIA “disregarded important aspects” of the claim—a textbook abuse of discretion under Portillo Flores and Rodriguez-Arias. The court also rejects the government’s effort to salvage the BIA’s decision by pointing to generalized statements about the absence of governmental intent to torture; in context, those statements only addressed mental illness and did not engage with the race-, disability-, and deportee-based claims or their supporting evidence.

Because this analytical failure alone compels vacatur, the panel expressly declines to decide McDougall’s alternative argument that the BIA misapplied the regulatory definition of torture (e.g., by equating “lack of training” with the absence of “specific intent,” or by applying the wrong acquiescence standard). Nevertheless, the court instructs that on remand the agency must “apply the definition of torture precisely to each claim” and must “announce its decision in terms sufficient to assure a reviewing court that it has done so.”

Impact and Forward-Looking Significance

• Reinforced Aggregation Obligation: This decision cements, within the Fourth Circuit, the obligation to aggregate the risk across all asserted reasons and perpetrators when adjudicating CAT claims. Practitioners should present—and adjudicators must address—each independent theory of torture risk and then perform a cumulative probability analysis.
• Meaningful Engagement Required: The BIA and IJs cannot insulate decisions with boilerplate statements that “all evidence” was considered. Opinions must reflect engagement with the specific theories and record evidence that matter to the outcome.
• Intersectional and Multi-Factor Claims: The case underscores that CAT risk can be intersectional. A petitioner’s race, disability status, mental illness, and deportee identity may interact to generate a higher risk of torture than any single factor would suggest. Adjudicators must address that interplay.
• Record-Building and Opinion-Writing: For DHS and the BIA, the opinion is a drafting directive: identify each asserted risk factor, discuss the supporting and contrary evidence, explain credibility and weight determinations, and articulate how the cumulative probability was derived. For respondents, it is a roadmap for building robust records.
• Acquiescence and Public Violence: Where the feared torturer is not a public official, the agency must still evaluate whether officials would consent or acquiesce, including by turning a blind eye. Evidence that authorities routinely fail to intervene or punish public violence becomes central.
• Procedural Guardrails: The decision continues the Fourth Circuit’s insistence on reasoned, transparent agency decision-making, fostering more reviewable, legally grounded CAT adjudications.

Complex Concepts Simplified

• CAT Deferral of Removal: A form of protection under the Convention Against Torture that prevents removal to a country where it is more likely than not the person would be tortured. Deferral is typically available even to those with certain criminal convictions that may bar other relief.
• “Torture” Under 8 C.F.R. § 1208.18: Intentional infliction of severe physical or mental pain or suffering, for a prohibited purpose (such as punishment, coercion, intimidation, or discrimination), by or with the consent or acquiescence of an official acting under color of law.
• Specific Intent: The torturer must intend to inflict severe pain or suffering. While negligence or poor training alone does not establish intent, the agency must still assess whether, on each theory, the record supports purposeful infliction for a prohibited purpose (e.g., punishment of a deportee, discriminatory targeting, or coercion).
• Consent or Acquiescence: Even if private actors are the immediate perpetrators, CAT is satisfied if public officials are aware (or willfully blind) and breach their legal duty to prevent or stop the torture.
• Aggregate Probability: The adjudicator must consider all potential torturers and all reasons together. Protection is required if the combined probability of torture from these overlapping risks exceeds 50%.
• Standards of Review:
  • Substantial Evidence: Factual findings stand unless any reasonable adjudicator would be compelled to conclude otherwise.
  • De Novo: Legal conclusions (such as the correct standard for “likelihood of torture”) are reviewed afresh.
  • Abuse of Discretion: The BIA must supply a reasoned explanation and cannot ignore important aspects of a claim or relevant evidence.
• Color of Law: Actions taken by officials in their official capacity—even if unlawful—are treated as taken under color of law for CAT purposes.

Practical Implications on Remand

• The BIA must explicitly analyze McDougall’s asserted risks tied to:
  • Race (as a Black man in Guyana),
  • Physical disabilities (including wheelchair use), and
  • Criminal deportee status (including notoriety and specific threats),
  in addition to mental illness.
• For each theory, the BIA must apply the regulatory elements of torture: severe pain/suffering, specific intent, prohibited purpose (including discrimination), and official involvement or acquiescence.
• It must then aggregate the resulting probabilities to determine whether the cumulative likelihood exceeds 50%.
• The decision must provide a cogent, articulable rationale that demonstrates engagement with the record and facilitates judicial review.

Conclusion

McDougall v. Bondi solidifies a critical practice point in CAT adjudication within the Fourth Circuit: the BIA must confront every nonfrivolous theory of torture and the supporting evidence, and must aggregate the risk across those theories and potential torturers. Failure to do so is an abuse of discretion warranting vacatur. The opinion advances transparent, reasoned decision-making and acknowledges the intersectionality of risk factors—race, disability, mental illness, and deportee status—in real-world assessments of torture risk. On remand and in future cases, the agency must apply the regulatory definition of torture precisely to each asserted basis and provide an explanation robust enough to assure meaningful appellate review.