CAT Deferral Cannot Rest on Speculative Chains of Harm: Clear-Error Review in Tuang v. Blanche

1. Introduction

Tuang v. Blanche addresses the boundary between permissible appellate correction of clearly erroneous immigration findings and impermissible reconsideration of the facts. The underlying claim concerned protection under the Convention against Torture (CAT), but the central institutional question was whether the Board of Immigration Appeals (BIA) properly reviewed an immigration judge’s predictions about future harm.

Pau Than Tuang, a native of Burma, also known as Myanmar, belongs to the Chin ethnic minority and is Christian. He entered the United States as a refugee in 2012 and became a lawful permanent resident in 2017. In 2021, he pleaded guilty in Oklahoma to one count of forcible sodomy of a child and four counts of lewd molestation of a child. He received a fourteen-year sentence, ten years of which were suspended.

Following his release from prison, Tuang entered immigration custody and conceded removability. He nevertheless sought deferral of removal under CAT, arguing that his return would expose him to forced military recruitment and severe mistreatment connected to his ethnicity, religion, and absence from Burma’s household registry.

The immigration judge granted protection. The BIA reversed, concluding that the judge’s predictions depended on unsupported links in a hypothetical sequence of events. Tuang challenged that reversal on three grounds:

1. The BIA had effectively reviewed factual findings anew rather than for clear error.
2. The immigration judge’s decision was supported by the evidence, while the BIA’s contrary conclusions were erroneous.
3. The BIA had engaged in its own factfinding and therefore should have remanded the case.

The Fifth Circuit rejected all three arguments. Its essential conclusion was that identifying unsupported predictive inferences in an existing record does not necessarily constitute either improper reweighing or new factfinding.

2. Summary of the Opinion

The court accepted that Burma was dangerous, that serious human-rights violations were occurring, and that Chin people and Christians had experienced discrimination and persecution. Those conditions, however, did not themselves establish that Tuang personally was more likely than not to be tortured.

The immigration judge’s theory depended principally on Tuang being identified upon arrival, forcibly conscripted, and then subjected to dangerous assignments or severe mistreatment because of his minority identities. A further finding concerned disproportionate punishment associated with his absence from the household registry.

The BIA accepted the expected encounter with government authorities at the airport but rejected the remaining pivotal findings as clearly erroneous. The Fifth Circuit agreed that:

• The evidence did not adequately establish that Tuang would probably be conscripted.
• Reports of civilian abuses did not establish that Tuang, as a military recruit, would probably receive assignments or treatment amounting to torture.
• The predicted consequences of his household-registration status lacked sufficiently specific support.
• His serious childhood mistreatment did not establish the required probability of future torture.
• The BIA could reverse without remanding because it had identified deficiencies in existing findings rather than made new evidentiary findings.

Tuang had already been removed to Burma while judicial review was pending. Counsel reported that he had been detained there, but the record did not establish where or why. The court declined to speculate about the detention and denied the petition.

3. Governing Legal Framework

A. The applicant’s burden under CAT

An applicant must establish that removal would more likely than not result in torture inflicted by, or with the acquiescence of, government officials acting under color of law. The governing regulations identified in the opinion include 8 C.F.R. §§ 1208.16(c), 1208.17, and 1208.18(a).

This is an individualized probability inquiry. Evidence of civil conflict, persecution, or widespread brutality supplies important context, but it must support a sufficiently concrete prediction about the applicant’s own exposure to torture.

The opinion does not hold that torture must occur because of a protected characteristic such as ethnicity or religion. Those characteristics mattered because Tuang’s particular theory predicted that they would cause or intensify his mistreatment. The question was whether the evidence supported that theory.

B. Predictive findings remain factual findings

A prediction about what will happen after removal is a finding of fact. Under 8 C.F.R. § 1003.1(d)(3)(i), the BIA reviews such findings for clear error. Other matters, including legal questions, are reviewed de novo under § 1003.1(d)(3)(ii).

The distinction is decisive. The BIA cannot replace a permissible factual assessment merely because it would have drawn a different inference. Conversely, the immigration judge’s predictive role does not insulate an inference that the record cannot reasonably support.

C. Sequential predictions require evidentiary support

When an applicant’s theory depends on successive events, the necessary links cannot be supplied by speculation. Both the immigration judge and the BIA applied the principle that each essential step in the proposed chain must be more likely than not to occur.

That requirement should not be confused with a conclusion that establishing each link automatically proves the ultimate claim. The applicant retains the overall burden of showing that torture itself is more likely than not. Here, the court found the chain deficient at several necessary stages and did not undertake a numerical calculation of the combined risk.

4. Precedents Cited and Their Influence

A. Matter of Z-Z-O-, 26 I. & N. Dec. 586, 590 (B.I.A. 2015)

This administrative precedent classifies an immigration judge’s predictions about future events as factual findings subject to clear-error review. It supplied the starting point for Tuang’s challenge: the BIA could not treat predictions about conscription or mistreatment as freely reconsiderable legal questions.

The Fifth Circuit noted that the BIA expressly recognized this rule. The remaining issue was whether the Board actually honored it.

B. Alvarado de Rodriguez v. Holder, 585 F.3d 227, 234–35 (5th Cir. 2009)

This case supplied the principal Fifth Circuit limitation on BIA review. The Board may not overturn factual findings simply because it would have weighed the evidence or decided the facts differently. Where the evidence permits two views, the factfinder’s choice cannot be clearly erroneous.

The precedent therefore constrained the court’s analysis rather than merely supporting the result. The Fifth Circuit had to determine whether the immigration judge’s conclusions were permissible inferences, not whether the BIA’s interpretation was preferable.

C. Kabba v. Mukasey, 530 F.3d 1239, 1245 (10th Cir. 2008)

The court cited this decision for the proposition that stating the correct standard is not enough: a reviewing court must examine whether the BIA actually applied it. Although a decision from another circuit, it reinforced the need to scrutinize the substance of the Board’s reasoning.

Accordingly, the BIA’s reference to clear-error review did not end the inquiry. Its treatment of each disputed prediction remained subject to judicial examination.

D. Parada-Orellana v. Garland, 21 F.4th 887, 893 (5th Cir. 2022)

This authority establishes that whether the BIA applied the correct standard of review is a legal question reviewed de novo. It explains why the Fifth Circuit independently assessed the Board’s methodology while preserving the distinction between legal review and original factfinding.

E. Anderson v. City of Bessemer City, 470 U.S. 564, 573–74 (1985)

This Supreme Court decision supplied two complementary principles. A choice between two permissible views of the evidence is not clearly erroneous. Nevertheless, a finding may be clearly erroneous even when some evidence supports it, if consideration of the entire record produces a definite and firm conviction that a mistake occurred.

The second principle was particularly important to the result. Tuang identified evidence relevant to his fears, but the court concluded that relevance alone did not make the immigration judge’s ultimate predictions permissible.

F. United States v. U.S. Gypsum Co., 333 U.S. 364, 395 (1948)

Quoted through Anderson v. City of Bessemer City, this case supplied the classic “definite and firm conviction” formulation of clear error. It provided the governing definition rather than a separate immigration-specific rule.

G. Torosyan v. Blanche, No. 25-60450, 2026 WL 2244205, at *5 (5th Cir. Aug. 4, 2026)

This decision supplied the substantive CAT standard: the applicant must show that torture by, or with the acquiescence of, government officials acting under color of law is more likely than not. It framed the ultimate question against which the predictive findings were measured.

H. L.N. v. Garland, 109 F.4th 389, 394 (5th Cir. 2024)

The court’s statement of the CAT burden came from this decision through Torosyan v. Blanche. Its role was to support the governing probability and official-involvement requirements, not to provide a separate analysis of Burma or military recruitment.

I. Matter of J- F-F-, 23 I. & N. Dec. 912, 917–18 (A.G. 2006)

This was the central authority concerning hypothetical chains of future events. The immigration judge cited it when assessing the proposed sequence, and the BIA relied on its principle that a CAT claim cannot be established by stringing together suppositions where necessary steps are not themselves more likely than not.

The disagreement concerned application, not the governing rule. The immigration judge considered the sequence sufficiently supported; the BIA and Fifth Circuit found several links speculative.

J. Santos-Zacaria v. Garland, 126 F.4th 363, 368 (5th Cir. 2025)

This precedent informed the remand issue by describing factfinding as including development of a record, gathering new information, and choosing between disputed facts. The Fifth Circuit concluded that the BIA had done none of those things. It had examined whether the existing record supported the immigration judge’s findings.

5. Analysis of the Court’s Legal Reasoning

A. The airport encounter: an accepted starting point

The immigration judge found that government authorities would encounter Tuang as a deportee upon arrival. Evidence indicated that deportee lists were provided to Burmese officials. This finding was undisputed.

Its significance was limited: probable contact with officials did not establish probable conscription, much less probable torture. The subsequent links required their own support.

B. Forced conscription: evidence of occurrence did not establish individual probability

The immigration judge inferred that Tuang would be conscripted from reports about Burmese men deported by Thailand. The BIA and court found that inference too strong for the evidence.

Several considerations drove the analysis:

• National recruitment figures: The military reportedly planned to recruit approximately 60,000 people annually from an eligible population of about fourteen million. Even assuming a male-only pool of roughly seven million, those figures did not establish that Tuang probably would be selected.
• Large-scale deportation reporting: One article described more than 144,000 deportations from Thailand and the forced recruitment of dozens of returnees. It did not establish that conscription was the probable experience of returnees generally.
• Smaller, more specific groups: Another report identified at least seventy-one recruits from two groups totaling more than 270 deported men. The record did not explain the outcomes of all remaining men.
• Bribery and release: Reports described payments securing release or avoiding service. The record did not establish whether Tuang or his family could make such payments.

The court did not require proof that deportees were never conscripted. It distinguished evidence that recruitment happened from evidence that it would probably happen to Tuang.

The figures nevertheless require careful interpretation. The larger deportee population was not necessarily a precisely comparable pool of military-eligible men, and the smaller reports did not resolve all unreported outcomes. The court’s conclusion rested substantially on those evidentiary gaps, rather than a comprehensive statistical model.

Likewise, the bribery discussion should not be expanded into a categorical requirement that CAT applicants prove inability to pay officials. In this opinion, payment evidence was treated as an unresolved feature of the specific recruitment theory.

C. Dangerous assignments: civilian abuse did not establish treatment of military recruits

The immigration judge predicted that Tuang would receive dangerous assignments because of his Chin identity, including possible use as a human shield or porter.

The BIA examined the reports’ factual settings. Much of the evidence concerned civilians, villagers, or ethnic minorities outside the military-conscription context. The court agreed that these accounts did not establish that Tuang would probably receive comparable treatment as a conscript.

The Board also distinguished documented incidents from predictions by researchers and activists. Reports expressing fears that new recruits might become human shields did not establish that the practice was occurring among military recruits.

Geography and the victims’ identities also affected the analysis. The cited 2023 report’s human-shield examples concerned other states. Another report did describe an incident in Chin State, but it occurred far from Tuang’s hometown and involved a Buddhist monk and civilians. That incident demonstrated brutality, but did not itself substantiate the asserted targeting of Christian conscripts.

These distinctions did not establish that Tuang’s hometown was safe or that Chin Christians were immune from abuse. They explained why the particular evidence did not support the particular prediction.

D. Severe ethnic mistreatment in military service

The immigration judge separately predicted severe mistreatment because Chin people were perceived as political dissidents and ethnic minorities experienced arbitrary arrest and detention.

The Fifth Circuit treated the BIA’s analysis of dangerous assignments as applicable here too. General evidence of ethnic abuses did not sufficiently connect Tuang’s anticipated military service to a probability of torture. The court thus rejected the inference without denying the underlying existence of discrimination or violence.

E. Household-registration consequences and religious identity

Tuang believed his parents had removed him and his brother from their household registry in approximately 2014 or 2015. The immigration judge observed that Burma had historically used registration systems to monitor suspected opponents and punish unauthorized departures or their families.

The missing elements were specificity and severity. The judge did not identify adequately supported findings about what punishment Tuang would probably receive, why it would be disproportionately severe because of his Chin identity, or how it would amount to torture. The prediction that Christianity would intensify the mistreatment was similarly unsupported.

F. Past persecution remained relevant but was not conclusive

Tuang’s childhood experiences were serious. He and other Chin children had been compelled to clear brush in areas containing active land mines. He and his brother were later forced to carry military supplies, denied water, confined overnight, and kicked after resisting recruitment. His brother was severely beaten.

The immigration judge expressly considered Tuang’s age and the frightening circumstances, but found that the harm did not rise to past torture. The judge nevertheless treated it as relevant to future risk and consistent with the persecution underlying his refugee admission.

The Fifth Circuit agreed that this history did not establish a present probability of torture. The distinction is important: recognizing past persecution does not automatically resolve the separate CAT inquiry. Past harm is evidence to be evaluated, not a substitute for the required prediction.

G. Why the BIA’s reasoning was not improper de novo review

The BIA did more than announce disagreement. It identified record-based reasons why the immigration judge’s conclusions exceeded the evidence: incomplete recruitment data, differences between civilians and conscripts, predictions rather than documented practices, and unspecified registration penalties.

The Fifth Circuit therefore characterized the Board’s work as identifying clear error, not selecting its preferred view between two permissible alternatives. This distinction is the opinion’s central administrative-law holding.

H. Why supporting evidence did not resolve Tuang’s second challenge

Tuang emphasized that evidence supported the immigration judge’s decision. But clear-error doctrine allows reversal even when a finding has some support, if the entire record leaves the reviewing body firmly convinced that a mistake occurred.

The court accordingly treated his second argument as substantially overlapping with the first. Its analysis of the unsupported predictive links resolved both.

I. Why no remand was required

The BIA neither developed a new record nor gathered additional information. The Fifth Circuit also concluded that it had not chosen between disputed historical facts. Instead, it assessed whether the immigration judge’s predictions could stand on the evidence already considered.

Once those findings were properly rejected, Tuang had not carried his burden. The Board did not need affirmative findings that Burma was safe or that torture was impossible. Failure of proof was sufficient, and the court found no need for further factfinding.

6. Potential Impact and Limits

A. More precise development of country-conditions evidence

The opinion illustrates the importance of connecting country reports to an applicant’s actual circumstances. Claims involving military service may require evidence distinguishing ordinary eligibility, recruitment of returnees, treatment of conscripts, treatment of civilians, and risks associated with particular identities or locations.

B. More explicit reasoning by immigration judges

Findings are more defensible when they explain why the evidence establishes probability rather than possibility. Where a claim depends on multiple events, identifying the support for each necessary step can expose gaps before appellate review.

C. A record-based route for BIA reversal

The decision confirms that meaningful clear-error review can include close examination of reports, their scope, and the inferences drawn from them. It does not authorize unrestricted reweighing: the prohibition against replacing one permissible view with another remains intact.

D. Limited implications for other CAT claims

The court did not hold that forced conscription can never lead to torture, that civilian abuses are always irrelevant to military-treatment claims, or that Chin Christians cannot qualify for CAT protection. A materially different record could support a different outcome.

Nor did the court deny protection because Tuang’s convictions made him undeserving. He did not assert that his criminal record would cause torture, and the opinion identified no evidence about the treatment of returnees with comparable convictions. The dispositive issue was failure to establish the required risk.

E. Post-removal detention remained unresolved

The report of Tuang’s detention could not establish, on this record, why he was held or what treatment he experienced. The court’s refusal to speculate was not a finding that the detention was harmless. It left those circumstances unresolved.

Finally, the opinion’s unpublished designation limits its precedential significance. Its principal value lies in illustrating how established CAT and clear-error principles may operate on a particular evidentiary record.

7. Complex Concepts Simplified

CAT deferral of removal

A form of protection preventing removal to a country where the applicant satisfies the CAT torture standard. It does not erase a criminal conviction or the underlying determination of removability.

More likely than not

The feared outcome must be probable, not merely possible. Ordinarily, this means a likelihood greater than fifty percent, though proof need not take the form of numerical statistics.

Predictive factual finding

A finding about what is expected to happen in the future, such as whether a returnee will be recruited or mistreated.

Clear-error review

A deferential form of review. A finding cannot be overturned merely because another conclusion seems preferable, but it can be reversed when the entire evidence produces a firm conviction that it is mistaken.

De novo review

Independent consideration without deference to the earlier decision on the issue being reviewed. The court used this approach to decide whether the BIA applied the proper review standard.

Hypothetical chain of events

A theory requiring several events to occur before torture results. Unsupported necessary steps cannot be filled in by assumption.

Persecution versus torture

Persecution and torture are distinct legal categories. Serious past persecution can matter greatly without establishing either past torture or a probability of future torture.

Factfinding versus evidentiary review

Making new findings is different from examining whether existing findings are adequately supported. The court held that the BIA performed the latter task.

8. Conclusion

Tuang v. Blanche illustrates two simultaneous requirements: CAT applicants must substantiate the necessary links between removal and probable torture, and the BIA must respect the immigration judge’s role as factfinder.

The Fifth Circuit concluded that the Board respected that boundary because it identified concrete deficiencies in the existing record rather than merely preferring different factual conclusions. The decision’s broader lesson is not that severe country conditions are irrelevant, but that they must support a sufficiently individualized and coherent prediction of torture. Its significance remains that of an unpublished application of established law.