Colorado Navarro v. Bondi (4th Cir. 2025): Quantifying “More-Likely-Than-Not” Torture and Re-affirming the BIA’s Clear-Error Authority

1. Introduction

In Jonathan Alexander Colorado Navarro v. Pamela Jo Bondi, the United States Court of Appeals for the Fourth Circuit addressed yet another contest between immigration fact-finders: an Immigration Judge (IJ) who granted Convention Against Torture (CAT) protection to an admitted MS-13 member versus the Board of Immigration Appeals (BIA) that reversed the IJ for clear error. The petitioner, Jonathan Colorado Navarro (“Colorado”), sought review of the BIA’s decision after the Board rejected the IJ’s credibility and likelihood-of-torture findings. At stake was not only Colorado’s removal but also the proper allocation of fact-finding power and the evidentiary quantum required to satisfy the “more likely than not” test under the CAT.

The Fourth Circuit (Judges Richardson, King, and Keenan) denied the petition, laying down two clarifying principles that will resonate in future CAT and credibility litigation:

The BIA may, consistent with its limited “clear-error” review, overturn an IJ’s positive credibility finding when “major inconsistencies and material omissions” leave the record devoid of reliable testimony—a holding that shores up the precedential force of Garland v. Ming Dai and Herrera-Martinez v. Garland.
Statistical or country-conditions evidence that shows a less than 50 percent probability of severe harm cannot sustain the CAT’s “more-likely-than-not” standard; petitioners bear the burden to bridge the numerical gap with individualized evidence, expert testimony, or corroboration. Absent such proof, an IJ’s prediction of torture is subject to reversal for clear error.

2. Summary of the Judgment

The Fourth Circuit affirmed the BIA on every contested point:

Credibility. The BIA correctly found clear error in the IJ’s “generally credible” assessment because Colorado’s three evolving asylum applications and sworn testimony revealed undisputed falsehoods and omissions about his gang membership and triple homicide.
Likelihood of Torture. The IJ relied on generalized prison-condition statistics and the petitioner’s testimony; the BIA concluded that these materials did not meet the “more-likely-than-not” threshold and that contrary evidence of El Salvador’s prison reforms was ignored. Substantial evidence supported the Board’s reversal.
Procedural Exhaustion. Colorado’s argument that the BIA should have remanded for live expert testimony was unexhausted because he did not ask the BIA for such relief below; thus, the Fourth Circuit lacked jurisdiction to consider it.
Outcome. Petition for review denied; removal order reinstated.

3. Detailed Analysis

3.1 Precedents Cited and Their Influence

United States v. U.S. Gypsum Co., 333 U.S. 364 (1948) — provides the classic definition of “clear error,” adopted by both BIA and circuit courts.
Garland v. Ming Dai, 593 U.S. 357 (2021) — underscores that reviewing courts must accept the agency’s adverse credibility determinations unless any reasonable fact-finder would be compelled to reach the opposite conclusion. The Fourth Circuit leaned on Ming Dai to uphold the BIA’s rejection of Colorado’s testimony.
Herrera-Alcala v. Garland, 39 F.4th 233 (4th Cir. 2022) — restates the “compels” standard under § 1252(b)(4)(B) and guides the court’s refusal to disturb the BIA’s fact review.
Ibarra Chevez v. Garland, 31 F.4th 279 (4th Cir. 2022) — clarifies a court’s duty to verify that the BIA applied the proper standard and then defer under “substantial evidence.”
Portillo Flores v. Garland, 3 F.4th 615 (4th Cir. 2021) (en banc) — supplies the governing CAT framework and the “more-likely-than-not” definition.
Rodriguez-Arias v. Whitaker, 915 F.3d 968 (4th Cir. 2019) — warns against rote statistical analysis, reminding courts to consider individualized factors; the BIA referenced this to discount Colorado’s bare statistics.
Shaw v. Sessions, 898 F.3d 448 (4th Cir. 2018) — stands for exhaustion requirements; used to bar Colorado’s remand argument.

3.2 Court’s Legal Reasoning

Nested Standards of Review.
The Fourth Circuit traced a three-step staircase:
1. IJ’s fact findings → reviewed by BIA for clear error.
2. BIA’s application of clear-error review → examined by the court de novo to ensure the correct standard was used.
3. BIA’s resulting determinations → affirmed unless substantial evidence compels a contrary result.
By articulating these layers in detail, the opinion bolsters clarity on what “substantial evidence of clear error” actually means.
Credibility Analysis.
The BIA catalogued Colorado’s shifting story: denial of gang membership, partial concession, eventual confession to triple murder, and conflicting explanations. The Circuit emphasized that deliberate omissions are as damaging as affirmative lies. Hence, the BIA was not merely “re-weighing” evidence but correcting an IJ misstep.
Quantifying the Likelihood of Torture.
The IJ cited deaths of “hundreds” during El Salvador’s state of exception, but the raw data—<200 deaths out of 52,000 arrests and 95,000 total detainees—translated to <0.1 % mortality, far below 50 %. The Fourth Circuit held that such statistics, without more, cannot satisfy the CAT threshold. This mathematical approach provides litigants with a clearer evidentiary yardstick.
Role of Expert Evidence.
Dr. Montgomery’s declaration (predicting a “100 %” torture risk) was discounted because she did not testify, was waived by the petitioner, and was never weighed by the IJ. The opinion tacitly signals to practitioners that written expert reports must be supported by live testimony or explicit IJ findings to carry appellate weight.

3.3 Impact on Future Litigation

Elevated Proof Expectations in CAT Cases.
Petitioners can no longer rely on generalized violence statistics; individualized, corroborated facts are needed to push the probability over 50 %.
Strategic Use of Experts.
Counsel must present experts at the IJ stage and obtain explicit findings. Otherwise, the BIA and courts may dismiss the reports as “less probative.”
BIA’s Oversight Strengthened.
The decision affirms that the BIA may overturn IJ credibility or CAT findings when the IJ ignores glaring inconsistencies or misreads country-conditions evidence.
Numerical Benchmarks.
Though the CAT standard remains qualitative, the opinion introduces a quantitative lens, making raw percentages a crucial element in future CAT showings.
Exhaustion Vigilance.
Arguments for remand or further fact-finding must be preserved before the BIA; otherwise, they are jurisdictionally barred.

4. Complex Concepts Simplified

Clear Error: A reviewing body (here, the BIA) may reverse a finding only if it has a “definite and firm conviction” that the initial fact-finder (the IJ) was wrong.
Substantial Evidence Review: The court asks whether any reasonable fact-finder could agree with the agency; if so, the finding stands—even if the judges might have decided differently.
More-Likely-Than-Not Standard (CAT): The applicant must show at least a 51 % probability of future torture by officials or with official acquiescence if removed.
Nested Review: A multi-tiered appellate structure where each level applies its own standard (IJ → BIA → Court), and the higher level largely defers to lower-level fact findings unless specific errors are proven.
Exhaustion: A litigant must raise each claim before the BIA; failure to do so deprives courts of jurisdiction to consider the claim later.

5. Conclusion

Colorado Navarro v. Bondi reinforces two pillars of immigration adjudication: (1) the BIA’s authority—and duty—to correct clearly erroneous IJ credibility and torture predictions, and (2) the petitioner’s burden to present evidence that mathematically and factually tips the scale above 50 %. By wedding doctrinal clarity to numerical rigor, the Fourth Circuit has set a precedent that will tighten CAT litigation, encourage meticulous evidentiary presentation, and remind practitioners that credibility remains the linchpin of human-rights-based relief. Going forward, asylum and CAT advocates must scrupulously prepare truthful, complete applications, secure live expert testimony, and compile individualized proof of likely torture; otherwise, the double-deference ladder from IJ to BIA to court will prove insurmountable.