UNPUBLISHED
UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT
No. 24-1105
JUAN JOSE PAREDES ALFARO, Petitioner,
v.
PAMELA JO BONDI, Attorney General, Respondent.
On Petition for Review of an Order of the Board of Immigration Appeals.
Argued: January 28, 2025 Decided: April 30, 2025
Before KING and RICHARDSON, Circuit Judges, and KEENAN, Senior Circuit Judge.
Petition for review denied by unpublished per curiam opinion.
ARGUED: Benjamin James Osorio, MURRAY OSORIO PLLC, Fairfax, Virginia, for Petitioner. Lynda Do, UNITED STATES DEPARTMENT OF JUSTICE, Washington, D.C., for Respondent. ON BRIEF: Simon Sandoval-Moshenberg, John W. Goodman, MURRAY OSORIO PLLC, Fairfax, Virginia, for Petitioner. Brian M. Boynton, Principal Deputy Assistant Attorney General, Stephen J. Flynn, Assistant Director, Office of Immigration Litigation, Civil Division, UNITED STATES DEPARTMENT OF JUSTICE, Washington, D.C., for Respondent.
Unpublished opinions are not binding precedent in this circuit.
2
PER CURIAM:
Petitioner Juan Jose Paredes Alfaro, a native and citizen of El Salvador, seeks review of a final order of removal of the Board of Immigration Appeals (the “BIA”), reversing an Immigration Judge’s (the “IJ”) determination that he was eligible for deferral of removal under the Convention Against Torture (the “CAT”). Paredes argues that the BIA impermissibly engaged in a de novo review of the IJ’s decision and, moreover, used the wrong legal standard in assessing the likelihood that he would be tortured if he was removed to El Salvador. We disagree. Accordingly, we deny Paredes’ petition for review.
I.
A.
Paredes was born in Cojutepeque, El Salvador, in December 1990. From an early age, he experienced significant mental health challenges, including learning disabilities, memory impairments, plus auditory and visual hallucinations. Those conditions persisted into adulthood. In March 2011, when he was approximately 20 years old, Paredes entered the United States on a tourist visa as a nonimmigrant visitor. He was authorized to remain in the United States until September 5, 2011. After his visa expired, however, Paredes remained in the country without legal authorization.
After arriving in the United States, Paredes first settled in Boston, Massachusetts, where, within his first year, he obtained nine tattoos. Those tattoos included the names of his loved ones, an Egyptian god, thirteen stars, and a dragon. After his time in Boston, Paredes moved to California to live with members of his family. He later returned briefly
3
to Massachusetts, before he ultimately relocated to Virginia in 2021. While living in Virginia, Paredes was arrested for robbery, grand larceny, credit card larceny, possession of marijuana, and public intoxication.
Shortly thereafter, on July 28, 2021, the Department of Homeland Security initiated removal proceedings against Paredes by issuing a Notice to Appear and charging him with removability under 8 U.S.C. § 1227(a)(1)(B) for overstaying the period permitted by his visa. Paredes initially appeared pro se at his removal proceedings before an IJ in Arlington, Virginia. On September 27, 2021, following concerns raised regarding his mental competency, the IJ convened a competency hearing. After reviewing Paredes’ medical evaluations and the testimony presented, the IJ found Paredes mentally incompetent to represent himself and, accordingly, appointed a qualified representative to assist him.
B.
On November 23, 2021, through counsel, Paredes admitted to the factual allegations contained in the Notice to Appear and conceded the charge of removability. As relief from removal, Paredes sought asylum, withholding of removal, and protection under the CAT — that is, deferral of removal. 1 He later withdrew his request for asylum.
In support of his request for relief, Paredes alleged that he feared torture by Salvadoran gangs, vigilante death squads, or government officials based on his tattoos,
1
Article 3 of the United Nations Convention Against Torture and Other Cruel, Inhuman or Degrading Treatment or Punishment provides that parties agree not to deport “a person to another State where there are substantial grounds for believing that he would be in danger of being subjected to torture.” See Turkson v. Holder, 667 F.3d 523, 525 (4th Cir. 2012) (internal quotation marks omitted).
4
criminal history, and mental health conditions. He asserted that his tattoos could be misinterpreted as gang affiliations, thereby exposing him to reprisal from rival gangs or extrajudicial punishment from law enforcement if he was returned to El Salvador.
Paredes also recounted that, while in custody, he had been approached by members of the MS-13 gang who questioned him about a dragon tattoo on his body. He explained that the tattoo is shaped like the number 8, which led MS-13 members to suspect that he may be affiliated with their rival gang, the “Barrio 18” or “18th Street” gang. 2 Paredes, however, maintains that he has never been a member of either the MS-13 or the 18th Street gangs.
For support, Paredes submitted a declaration and testimony from Dr. Sarah C.
Bishop, an expert on Salvadoran culture and society. In her testimony, Dr. Bishop explained that tattoos are often affiliated with gang membership in El Salvador and thus deportees with tattoos are frequently assumed to be members of gangs. Even if a tattooed deportee is not a member of a gang, she explained, this assumption often “leads to harassment, torture, and sometimes death.” See J.A. 1537. 3 Dr. Bishop, however,
The two dominant gangs in El Salvador are the Mara Salvatrucha 13 gang (“MS-
2
13”) and the 18th Street, or “Barrio 18,” gang. See Gov. Br. 8 n.8. Both originated in Los Angeles, California — MS-13 among Salvadoran immigrants, and 18th Street even earlier among Mexican youth — and later expanded their influence southward as members were deported following criminal convictions in the United States. See J.A. 2194-95 (United Nations report describing formation of MS-13 and 18th Street gangs). In El Salvador, these gangs have engaged in turf wars, extortion, drug trafficking, and brutal acts of violence, often exercising de facto control over neighborhoods and communities.
3
Our citations herein to “J.A. ___” refer to the Joint Appendix filed herein by the parties.
5
acknowledged that there were Salvadoran populations where tattoos were culturally acceptable — like the “big expat surfing culture” that exists along El Salvador’s coast. Id. at 1159; see also id. at 1151 (discussing “surfers along the beach that have tattoos” who utilize tattoo removal clinics).
Dr. Bishop also maintained that Paredes faced a risk of torture by Salvadoran officials in light of “Decree 717,” a measure passed by the Salvadoran government as a means of identifying, tracking, and detaining deportees who were suspected of having ties to gangs. See J.A. 1524; see also Ibarra Chevez v. Garland, 31 F.4th 279, 285 (4th Cir. 2022) (discussing Decree 717). She thus posited that Salvadoran law enforcement would process Paredes upon his return, record his tattoos and his criminal history, and potentially detain him if his tattoos aroused suspicion of gang affiliation. In her declaration, however, Dr. Bishop explained that this sequence of events would only take place if a deportee “is not met by family and is not detained by police upon arrival.” See J.A. 1539.
Dr. Bishop testified that Salvadoran President Nayib Bukele had authorized law enforcement and military personnel to “use lethal force against individuals suspected of being gang members”’ even if there was no proof of an individual’s affiliation with a gang. See J.A. 1148. She then explained that this policy has “emboldened police” to routinely beat and abduct individuals. Id. However, Dr. Bishop later clarified that, to her knowledge, individuals merely suspected of gang membership but not proven to be affiliated with a gang are within the scope of President Bukele’s directives. Id. at 1154-56.
Dr. Bishop also opined that Paredes faced a high probability of torture from extrajudicial groups that are made up of present and former police officers, or vigilante
6
death squads that “operate with relative impunity.” See J.A. 1152. She further explained that gang members would likely attempt to extort Paredes, recruit him as a member, or harm Paredes if he is perceived to be a member of a rival gang. Id. at 1161-62. To support her opinion, Dr. Bishop cited a Human Rights Watch Report stating that “tattoos were the most common factor” among deportees who were abused or murdered upon their return to El Salvador, although she acknowledged in her declaration that “most of [those] cases were gang related.” Id. at 1539, 1149.
On March 23, 2022, the IJ issued a decision granting Paredes deferral of removal under the CAT, after determining that Paredes would be more likely than not to face torture upon repatriation to El Salvador. In rendering this ruling, the IJ relied on Dr. Bishop’s testimony in which she stated, among other things, that if gangs did not view Paredes as a rival member, they would likely attempt to recruit him; that authorities would automatically assume he is affiliated with a gang due to his criminal record and tattoos; that at least fifteen extrajudicial groups routinely target individuals suspected of gang involvement; and that vigilante death squads can readily obtain information about suspected gang members from law enforcement.
DHS appealed the IJ’s decision to the BIA. On September 22, 2022, a majority of a BIA panel vacated the IJ’s ruling, concluding that the record required further factual development, and remanded Paredes’ case for additional proceedings to reconsider the
7
grant of CAT protection. See J.A. 920-24. 4 In so doing, the BIA ruled that the IJ made several clearly erroneous factual determinations and relied upon Dr. Bishop’s testimony without considering contrary evidence.
On remand, the IJ conducted supplemental hearings and again evaluated Paredes’
eligibility for CAT protection. Paredes submitted additional evidence to support his applications, including a copy of a travel advisory, issued in 2022 by the U.S. Department of State’s Bureau of Diplomatic Security, plus a panoply of news articles indicating that the Salvadoran legislature had enacted a “State of Emergency” in an effort to crack down on gang violence. Those articles detailed that El Salvador’s State of Emergency suspended certain constitutional rights associated with arrest, detention, and access to counsel.
On March 3, 2023, the IJ issued a decision denying withholding of removal, on the grounds that Paredes’ conviction for strangulation constituted a particularly serious crime. In that same decision, the IJ again granted Paredes deferral of removal as relief under the CAT. On April 3, 2023, the DHS appealed the IJ’s decision granting CAT protection, contending that the IJ’s ruling rested on an impermissibly speculative risk assessment, in that the IJ relied on select portions of Dr. Bishop’s testimony and generalized country conditions evidence. On that same day, Paredes filed a cross-appeal contending that the IJ
4
A three-member panel of the Board considered DHS’s appeal. One member dissented, stating — without a separate written opinion — that she would have dismissed the appeal. See J.A. 924. The panel ultimately remanded Paredes’s case, in part, to allow the IJ to reconsider the denial of withholding of removal in light of new evidence concerning Paredes’s mental illness. However, Paredes did not contest the IJ’s renewed decision to deny withholding of removal in his subsequent appeal to the Board, nor does he raise any challenge to that ruling before this Court.
8
erred in its determination that his application for relief under the INA was statutorily barred by his conviction of a particularly serious crime. 5 C.
On January 22, 2024, the BIA sustained the DHS appeal, dismissed Paredes’ cross-
appeal, and reversed the IJ’s grant of CAT protection. See Juan Jose Paredes Alfaro, A220- 477-000 (BIA Jan. 22, 2024) (the “BIA Decision”). 6 The BIA, after reviewing the totality of the record evidence, explained that the IJ’s finding that Paredes would likely be tortured was predicated, at least in part, on assumptions regarding the interpretations of his tattoos by unknown actors, conjectures about the behavior of gangs and death squads, and speculation regarding the likelihood of institutionalization.
Notably, the BIA did not disturb the IJ’s credibility determination as to Dr. Bishop’s testimony, nor did it dispute that Paredes genuinely feared harm. See, e.g., BIA Decision 4 (discussing Dr. Bishop’s testimony). Rather, the BIA identified defects in the IJ’s analysis of whether the record, as a whole, established a clear probability of torture under the CAT’s regulatory standards — those being that the IJ had clearly erred in relying on generalized or anecdotal evidence from Dr. Bishop to conclude that Paredes had
5
On appeal, the Board concluded that Paredes had waived his challenge to the finding that his conviction constituted a particularly serious crime. The BIA explained that, aside from a single, conclusory statement asserting that his strangulation conviction should not be considered particularly serious due to his mental health issues at the time of the offense, Paredes offered no meaningful argument in either his Notice of Appeal or his appellate brief. See BIA Decision at 2. Paredes does not dispute the BIA’s waiver determination in this appeal.
6
One member of the three-judge panel dissented without a written opinion.
9
demonstrated a particularized risk of harm based on his non-gang related tattoos; mischaracterized or disregarded record evidence that contradicted Dr. Bishop’s assessments; and failed to differentiate between torture and lesser forms of mistreatment.
More specifically, the BIA concluded that the IJ had unduly relied on the testimony of Dr. Bishop “to conclude that, in large part because of his non-gang tattoos, [Paredes] is likely to be tortured in El Salvador by the police or vigilante death squads.” See BIA Decision at 3. The Board explained that “[w]hile Dr. Bishop’s testimony may have accurately described the anecdotal evidence regarding some aspects of tattoos in El Salvador,” the IJ failed to address contrary evidence, or — in some instances — mischaracterized “evidence regarding the wide-spread nature of non-gang tattoos in El Salvador.” Id. at 4 (explaining that IJ “did not reference any part of Dr. Bishop’s testimony that engaged” with evidence “that contradicted her view”); see also id. at 3-4 (acknowledging that DHS submitted evidence of “international tattoo conventions,” plus “20 commercial tattoo salons in the capital city of San Salvador, with additional salons elsewhere in the country”).
The BIA next addressed the IJ’s findings on whether the Salvadoran government would acquiesce in Paredes’s potential torture. In reviewing the IJ’s determination that Paredes was likely to be tortured by or with the consent of Salvadoran officials, the BIA first noted that the record contained no credible evidence showing that government actors would be aware of, and choose not to prevent, any harm inflicted on Paredes by gangs or extrajudicial vigilante groups. See BIA Decision 5.
10
The BIA then pointed out that the IJ’s conclusion relied heavily on generalized concerns about corruption, human rights violations, and impunity in El Salvador—such as “evidence of nationwide corruption and anecdotal evidence of violence.” See BIA Decision 6. While it acknowledged that such issues exist, the BIA stressed that the sources cited by the IJ used vague language — such as stating that authorities “have been known to carry out extrajudicial killings” — and lacked the specific, individualized detail required to support a claim under the Convention Against Torture. Id. at 5. The BIA concluded that the IJ’s reliance on these broad and non-specific assertions did not constitute a “permissible view” of the evidence, and therefore found the IJ’s finding of governmental acquiescence to be clearly erroneous. Id. at 6. More specifically, the BIA pointed out that Paredes had not offered any evidence that he would be institutionalized or detained upon his return to El Salvador. The BIA noted an absence of evidence demonstrating that Salvadoran authorities routinely detain deportees with conditions similar to those affecting Paredes.
Finally, the BIA assessed the scope of torture underpinning the IJ’s determination that Paredes was likely to face torture upon his return to El Salvador. See BIA Decision 6. In doing so, the BIA determined that the IJ “did not correctly distinguish the likelihood of mistreatment from the likelihood of torture.” Id. The BIA emphasized that, contrary to the IJ’s conclusion, Paredes had not shown that he would be subjected to treatment rising to the level of torture. It further explained that the evidence the IJ relied upon — namely, a Human Rights Watch report describing generalized harm to deportees with tattoos and the lawful detention of suspected gang members — did not meaningfully support the
11
conclusion that Paredes faced a likelihood of torture based on his possession of non-gang- related tattoos. Id. (“[A] Human Rights Watch report regarding the likelihood of generalized harm to deportees with tattoos does not significantly support the proposition of a likelihood of future torture.”).
The BIA’s decision concluded that the IJ’s finding that Paredes was likely to be tortured lacked sufficient evidentiary support and failed to establish the individualized risk required for CAT protection. Paredes then timely petitioned this Court for review of the BIA’s decision. We possess jurisdiction to consider Paredes’ petition for review pursuant to 8 U.S.C. § 1252(b)(2).
II.
“When, as here, the BIA reverses an immigration judge’s decision and issues its own opinion without incorporating the IJ’s decision, we review only the BIA’s decision.” See Funez-Ortiz v. McHenry, 127 F.4th 498, 504 (4th Cir. 2025). “We review the BIA’s legal conclusions, including the question of whether the BIA has applied the proper standard of review, de novo.” See Ibarra Chevez v. Garland, 31 F.4th 279, 288 (4th Cir. 2022). And “[w]e review factual findings for substantial evidence, treating them as conclusive unless any reasonable adjudicator would be compelled to conclude to the contrary.” See Portillo Flores v. Garland, 3 F.4th 615, 626 (4th Cir. 2021); see also Nasrallah v. Barr, 590 U.S. 573, 576 (2020) (explaining that “judicial review of factual challenges to CAT orders must be highly deferential”).
12
Those factual findings — “including [the] predictions about the likelihood of future mistreatment and government acquiescence — ‘are conclusive unless any reasonable adjudicator would be compelled to conclude to the contrary.’” See Ponce-Flores v. Garland, 80 F.4th 480, 484 (4th Cir. 2023) (quoting 8 U.S.C. § 1252(b)(4)(B)). Under this deferential standard, we may not overturn the BIA’s findings unless the record compels a contrary result. See 8 U.S.C. § 1252(b)(4)(B).
III.
Paredes challenges the BIA’s ultimate determination that he is not more likely than not to face torture if returned to El Salvador. Underpinning that contention are two principal arguments. Paredes first asserts that the BIA applied the wrong legal standard in overturning the IJ’s grant of CAT protection. He argues next that the BIA overstepped its authority by improperly substituting its own judgment for that of the IJ.
As explained herein, we discern no error with the BIA’s reversal of the IJ’s grant of CAT protection. And we are satisfied that, when viewed as a whole, the record evidence does not compel a finding that Paredes is more likely than not to be tortured upon his return to El Salvador.
A.
1.
We turn first to the question that is at the heart of this petition: Whether substantial evidence supports the BIA’s conclusion that Paredes failed to establish a clear probability that he would be subjected to torture if he were to be removed to El Salvador, either by
13
Salvadoran officials or with their acquiescence. Upon thorough review of the record and the BIA’s decision, we are satisfied that the BIA’s determination was supported by substantial evidence.
“Unlike asylum and withholding of removal claims, which require an applicant’s fear of persecution to be based on an enumerated ground, protection under the CAT is available for an applicant who can prove the likelihood of torture regardless of the motivation.” See Lizama v. Holder, 629 F.3d 440, 449 (4th Cir. 2011). 7 To qualify for protection under the CAT, Paredes bears the burden demonstrating that it is “more likely than not” that he would “be tortured if removed to the proposed country of removal.” See 8 C.F.R. § 1208.16(c)(2).
That burden is demanding. Paredes must show not only that he faces harm, but that such harm would rise to the level of “torture” as defined by the regulation — namely, the intentional infliction of severe physical or mental pain or suffering, by or at the instigation of, or with the consent or acquiescence of, a public official or other person acting in an official capacity. See 8 C.F.R. § 1208.18(a). Speculative or generalized fears of violence do not satisfy this standard. See, e.g., Singh v. Holder, 699 F.3d 321, 334-45 & n.16 (4th Cir. 2012) (recognizing that country conditions reports describing “widespread corruption”
7
“Noncitizens who are otherwise ineligible by statute or regulation for asylum or withholding of removal under the INA or the CAT may be eligible for other protection under the CAT, including deferral of removal.” See Kouyate v. Garland, 122 F.4th 132, 141 (4th Cir. 2024). A noncitizen who seeks deferral of removal, like Paredes, “asks the government not to rescind his removal but merely to forbear from or delay removing him” to a designated country where he would be tortured. See Lopez-Sorto v. Garland, 103 F.4th 242, 249 (4th Cir. 2024).
14
do not “constitute a sufficient ground for finding that a particular person would more likely than not be tortured”).
Paredes attempts to meet this burden by offering several independent sources of potential harm that, “whether standing alone or considered cumulatively,” establish the requisite likelihood of torture and thus entitle him to relief under the CAT. See Petitioner’s Br. 29-30. First, Paredes points to his tattoos. He has nine in total, including a dragon that resembles the number 8. Paredes says that MS-13 members have questioned him about his dragon tattoo, suspecting that he is affiliated with the rival 18th Street gang. He fears that if he is returned to El Salvador, government officials — or, alternatively, vigilante groups acting at the behest of Salvadoran law enforcement — will make similar assumptions, and thus cause him to be tortured. Second, Paredes points to his criminal history. He argues that deportees with criminal histories are treated harshly in El Salvador, and claims that he will either be targeted by the police or singled out for extrajudicial punishment by government-sanctioned vigilante death squads.
Undoubtedly, Paredes’ theories of harm are concerning, and should be taken seriously. But those claims — though grounded in decidedly difficult personal circumstances — fall short of the requisite showing. Take first, Paredes’ claim regarding his tattoos. His tattoos are innocuous: A dragon, stars, and the names of loved ones. And Paredes himself attested to the fact that none of his tattoos signify an association with a gang. Moreover, the record is devoid of specific, individualized evidence establishing that generic tattoos are commonly associated with gang membership in El Salvador. And even the country conditions evidence submitted by Paredes in support of his application —
15
though confirming that tattoos can, in some circumstances, raise suspicion — stops short of supporting the inference that individuals with non-gang-related tattoos are routinely subjected to torture by gangs or state actors.
The same is true of Paredes’ claims regarding his criminal record. We recognize that deportees with criminal histories may face stigma, just as we recognize that there is anecdotal evidence reflecting that Salvadoran officials have harmed deportees with tattoos in its efforts to curb gang-related violence. It is certainly “possible, or even probable” — as the BIA properly acknowledged — that Paredes “will have some level of contact with law enforcement” upon his return to El Salvador. See BIA Decision 6. But although Paredes offered evidence of generalized harm perpetrated by Salvadoran officials, he fails to link these assertions to his individual circumstances. Without that showing, Paredes’ likelihood of torture remains speculative.
Paredes, then, must rely on a theory of acquiescence to demonstrate that he is likely to be tortured. A public official has acquiesced to torture if, “prior to the activity constituting torture, the official has awareness of such activity and thereafter breached his or her legal responsibility to intervene to prevent such activity.” See Lizama, 699 F.3d at 449 (internal quotations and alterations omitted). “Such ‘awareness’ can be shown either through ‘actual knowledge or willful blindness.’” See Moreno-Osorio v. Garland, 2 F.4th 245, 256 (4th Cir. 2021) (quoting 8 C.F.R. § 1208.18(a)(7)).
In this respect, Paredes posits that Salvadoran law enforcement will learn of his prior convictions and suspect that he was involved in a gang. He claims that those officials will either engage in his extrajudicial torture or, alternatively, inform vigilante groups of their
16
suspicions. Those vigilante groups, Paredes says, will then target him for torture with the acquiescence of Salvadoran law enforcement. But here, again, Paredes’ theories of harm are unsupported by the record evidence.
While it is true that some evidence suggests that “the Salvadoran police . . . have been known to carry out extrajudicial killings,” there is nothing in the record that meaningfully links these anecdotes to Paredes’ particular circumstances. See BIA Decision 5; see also, e.g., Ibarra Chevez v. Garland, 31 F.4th 279, 290 (4th Cir. 2022). By the same token, such “vague language” and sweeping assumptions about impunity within the Salvadoran government does not do the work necessary to sustain Paredes’ claim for CAT protection — particularly when considered in tandem with evidence of “significant efforts to eliminate police corruption” in El Salvador. See BIA Decision 5; Ponce-Flores v. Garland, 80 F.4th 480, 485 (4th Cir. 2023); see also Lizama, 629 F.3d at 449-50 (rejecting contention that “gangs or other criminal entities have the approval or acquiescence of the government of El Salvador” and noting that “the [Salvadoran] government’s current focus is . . . providing great security for the public against gang violence”).
Although there is no question that El Salvador has struggled with violence, the law demands more than a troubling backdrop. See Singh, 699 F.3d at 321, 334-35 & n.16 (upholding denial of CAT protection where country reports identified widespread abuse but did not show that petitioner himself would more likely than not be tortured). It demands a concrete, individualized showing of risk that “he would be singled out for torture.” See Kerr v. Garland, 66 F.4th 462, 471 (4th Cir. 2023). Undoubtedly, the harms Paredes fears
17
are serious. But the record evidence he offered simply does not bridge the gap between generalized country conditions and the kind of particularized risk required under the CAT.
2.
The BIA thus determined that Paredes’ fears, while genuine, were based on possibility — not probability. In this respect, the BIA’s application of Matter of J-F-F- was appropriate. That decision reaffirms that predictive findings regarding torture must be grounded in a plausible chain of events, not “a series of suppositions.” See 23 I&N Dec. 912, 917 (A.G. 2006); see also Lopez-Sorto, 103 F.4th at 253 (clarifying that J-F-F- stands for the proposition that “a party cannot rely upon ‘a series of suppositions’ to establish a risk of torture”). Yet Paredes argues that the BIA used J-F-F- to break his claim into smaller pieces, treating each potential source of harm — vigilante death squads, Salvadoran law enforcement, and gangs — as too speculative to sustain his claim for CAT protection when viewed alone.
Paredes is correct that the BIA cannot ignore the cumulative effect of different sources of harm when considering a claim for CAT protection. If a person faces danger from several directions — say, from gangs, government officials, and vigilante groups — the BIA has to step back and ask whether, taken together, those threats make it more likely than not that the person will be tortured if removed. See, e.g., Rodriguez-Arias v. Whitaker, 915 F.3d 968, 973 (4th Cir. 2019); Ibarra Chevez, 31 F.4th at 289.
But Paredes is mistaken about how we have interpreted and applied Matter of J-F-
F-. He claims that we have limited J-F-F- to cases involving “one causal chain per source of torture.” See Petitioner’s Br. 27. We have not. Instead, we have clarified that “a chain-
18
of-events analysis could[] be used on [the] first step — i.e., to determine the likelihood of torture from a single source” prior to aggregating the likelihood of torture from multiple sources. See Lopez-Sorto, 103 F.4th at 254. Put simply, a “chain-of-events analysis” can be used to assess the likelihood of torture from a single source before adding up the risks from multiple sources. But the prediction of torture from any single source must be grounded in a plausible chain of events, not “a series of suppositions.” Id. at 253–54.
To be sure, aggregation is required under this framework. Our precedent requires that “the risks of torture from all sources should be combined when determining whether a CAT applicant is more likely than not to be tortured in a particular country.” See Rodriguez-Arias, 915 F.3d at 973. So, if a petitioner for CAT protection alleges multiple sources of torture, the question becomes where the cumulative probability of torture — across all those threats — exceeds 50%. See Ibarra Chevez, 31 F.4th at 289. If an applicant can show that the probability of torture exceeds 50%, whether because of multiple sources of harm or a credible chain of events, then he is entitled to protection under the CAT.
Crucially, there is no rigid formula for how the BIA must perform that analysis.
Indeed, “[t]here is nothing in the regulations, agency guidance, or our precedents that require[s]” the BIA “to employ any specific methodology when considering and aggregating the likelihood of risk from multiple entities, nor do they require any specific statistical or quantitative analysis when evaluating the ‘more likely than not’ standard in the regulations.” See Ibarra Chevez, 31 F.4th at 290 (quoting 8 C.F.R. § 1208.16(c)(2), (c)(3)). Nor is the BIA required to “make it perfectly clear that it
19
was performing an aggregate analysis.” Id. (citing Iraheta-Martinez v. Garland, 12 F.4th 942, 960 (9th Cir. 2021)); see also Lopez-Sorto, 103 F.4th at 255 (explaining that BIA “need not be crystal clear” in its aggregation analysis). It just needs to show its work clearly enough so that we can tell it considered the big picture. Id. (explaining that “[i]t is sufficient if the court can tell that” the BIA “combined the threats from each source”).
Consistent with our precedent, the BIA’s reasoning here makes clear that it considered the risk of torture in the aggregate in a way that is “sufficient to enable a reviewing court to perceive” the substance of its analysis. See Ibarra Chevez, 31 F.4th at 292. It acknowledged and discussed the multiple sources of potential harm that Paredes identified — Salvadoran officials, gangs, and vigilante groups — and then weighed those risks in the aggregate. See BIA Decision 4 (“We have considered the extensive evidence, to include current country conditions reports, of a pattern of human rights violations in El Salvador . . . . While the respondent may face some risk from such sources, the cumulative record does not support a conclusion that the risk is significant enough to establish that [Paredes] is more likely than not to be tortured in El Salvador.”). Indeed, “[b]y addressing each risk individually and then considering all three together,” it is evident that the BIA properly “aggregated the risks facing” Paredes. See Lopez-Sorto, 103 F.4th at 255 (internal quotation marks omitted). The BIA’s decision thus satisfies our standard.
In short, the BIA did not misunderstand J-F-F- or misuse it to reject Paredes’s claim.
It applied it as a guardrail against speculation, not as a barrier to proper aggregation. The BIA’s ruling reflects the kind of holistic, evidence-based analysis that the law requires, and its conclusion — that Paredes failed to demonstrate a likelihood of torture when all sources
20
of potential harm are considered together — is both legally sound and supported by the record.
As such, the BIA’s conclusion that Paredes failed to establish a likelihood of torture — whether by gangs, rogue actors, or the Salvadoran government — was not only reasonable but compelled by the evidentiary gaps in the record. Because our role is not to reweigh the evidence, but rather to determine whether the BIA’s decision was supported by substantial evidence, we are therefore bound to affirm the BIA’s determination the Paredes is not entitled to protection under the CAT.
B.
In addition to contesting the BIA’s ultimate determination, Paredes also challenges the way the Board reached its decision. More specifically, he argues that the BIA impermissibly undertook a de novo review of the record, instead of reviewing the IJ’s decision for clear error. In Paredes’ view, the BIA improperly reassessed the credibility of Dr. Bishop, reweighed the evidence related to the risk of torture presented by his tattoos, and reinterpreted factual inferences that were within the exclusive purview of the IJ. We disagree.
The BIA is constrained to review an IJ’s factual findings for clear error. See 8 C.F.R. § 1003.1(d)(3)(i) (“Facts determined by the immigration judge . . . shall be reviewed only to determine whether the findings of the immigration judge are clearly erroneous.”); see also Funez-Ortiz v. McHenry, 127 F.4th 498, 505-04 (4th Cir. 2025). That means the BIA cannot discard an IJ’s factual determinations “simply because [it is] convinced that [it] would have decided the case differently.” See Glossip v. Gross, 576
21
U.S. 863, 881 (2015) (quoting Anderson v. Bessemer City, 470 U.S. 564, 573 (1985)). Nor can it selectively examine evidence in an effort to “substitute[] its own judgment for that of the IJ.” See Crespin-Valladares v. Holder, 632 F.3d 117, 128 (4th Cir. 2011) (citing Kabba v. Mukasey, 530 F.3d 1239, 1246 (10th Cir. 2008)). But the BIA can reverse when, after reviewing the entire record, it is “left with the definite and firm conviction that a mistake has been committed.” See United States v. U.S. Gypsum Co., 333 U.S. 364, 395 (1948); see also Garland v. Ming Dai, 593 U.S. 357, 367 (2021) (“[T]he BIA . . . has experience with the sort of facts that recur in immigration cases and the ability to directly override the IJ’s factfindings.”).
That is what the BIA did here: It accepted the IJ’s credibility determinations and agreed with certain factual premises — namely, that Paredes suffers from mental health conditions, that El Salvador experiences high levels of crime and corruption, and that tattoos can sometimes lead to the targeting of individuals. It then examined the IJ’s prediction about Paredes’ likelihood of future harm, and then asked whether that predictive finding rested on solid evidentiary ground. It thus determined that the IJ had relied too heavily on generalized reports and anecdotal accounts, not concrete evidence that Paredes himself faced a specific risk of torture.
The BIA declined to adopt the IJ’s determination that Paredes would more likely than not face torture upon removal to El Salvador, concluding that the finding lacked plausibility when viewed against the record as a whole. That determination does not amount to the kind of second-guessing Paredes suggests. Rather, it reflects the exercise of
22
the BIA’s authority to correct factual findings premised on a misunderstanding of the evidentiary record. We find no error in the BIA’s determination in this regard.
C.
We are not unsympathetic to Mr. Paredes’s situation. His personal history, mental health struggles, and fears about returning to a country plagued by instability warrant serious consideration. We also recognize that Paredes could face a difficult return: He may encounter stigma, economic hardship, or inadequate treatment for his mental health disorders. Nothing in the record, however, compels the conclusion that he would suffer severe harm intentionally inflicted with the awareness of the Salvadoran government, or with its consent.
We emphasize that our role as a reviewing court is not to reweigh the evidence or to substitute our own views for those of the agency; it is, instead, to ensure that the BIA acted within the bounds of the law, applied the correct legal standards, and reached a decision supported by substantial evidence. The BIA did so here, and we are thus constrained to deny Paredes’ petition for review.
IV.
For the foregoing reasons, we deny Paredes’ petition for review of the BIA decision reversing the IJ’s grant of CAT protection.
PETITION FOR REVIEW DENIED