FOR PUBLICATION

UNITED STATES COURT OF APPEALS

FOR THE NINTH CIRCUIT

MIGUEL ANGEL VELASQUEZ- No. 21-70093

SAMAYOA,

Petitioner, Agency No.

A044-804-654

v.

MERRICK B. GARLAND, Attorney OPINION General, Respondent.

On Petition for Review of an Order of the Board of Immigration Appeals

Argued and Submitted March 10, 2022 Pasadena, California

Filed June 24, 2022

Before: Richard C. Tallman and Michelle T. Friedland, Circuit Judges, and Edward R. Korman, * District Judge.

Opinion by Judge Friedland

* The Honorable Edward R. Korman, United States District Judge for the Eastern District of New York, sitting by designation. 2 VELASQUEZ-SAMAYOA V. GARLAND

SUMMARY **

Immigration

Granting Miguel Angel Velasquez-Samayoa’s petition for review of the Board of Immigration Appeals’ decision affirming denial of protection under the Convention Against Torture, and remanding, the panel held that the Board erred by failing to adequately consider Velasquez-Samayoa’s aggregate risk of torture from multiple sources, and erred in rejecting Velasquez-Samayoa’s expert’s credible testimony solely because it was not corroborated by additional country conditions evidence.

Velasquez-Samayoa asserted that, if he were removed to his native country of El Salvador, he would be identified as a gang member based on his gang tattoos and face a significant risk of being killed or tortured—either by Salvadoran officials or by members of a rival gang with the acquiescence of the Salvadoran government. Relying on Matter of J-F-F-, 23 I. & N. Dec. 912 (A.G. 2006), the Board concluded that Velasquez-Samayoa failed to demonstrate a clear probability of torture because he did not establish that every step in a hypothetical chain of events was more likely than not to happen.

The panel concluded that the Board erred by failing to assess Velasquez-Samayoa’s aggregate risk of torture. Discussing Cole v. Holder, 659 F.3d 762 (9th Cir. 2011), the panel explained that when an applicant posits multiple theories for why he might be tortured, the relevant inquiry is whether—considering all possible sources of and reasons for ** This summary constitutes no part of the opinion of the court. It has been prepared by court staff for the convenience of the reader.

VELASQUEZ-SAMAYOA V. GARLAND 3

torture—the total probability that the applicant will be tortured exceeds 50 percent. For example, if an applicant is at risk of torture from police, death squads, and gangs, he need not prove that each group, treated individually, would more likely than not torture him. The panel explained that Cole’s approach is consistent with Matter of J-F-F-, which provides that, when an applicant posits a single theory for why he would be tortured, but the torture will come about only if several hypothetical events all occur in sequence, an applicant must show that the individual probability of each event occurring is greater than 50 percent.

Here, the Board considered Velasquez-Samayoa’s two separate theories of torture—either by Salvadoran officials or by members of a rival gang—as a single hypothetical chain of events and denied his CAT claim because the probability of that hypothetical chain occurring was not high enough. The panel concluded that in doing so, the Board misapplied Cole and Matter of J-F-F-. The panel explained that the Board should not have considered Velasquez- Samayoa’s claim as a single hypothetical chain of events, when—as the Board itself acknowledged—he posited two “alternative” and distinct theories for why he would be tortured if he were removed to El Salvador. By requiring Velasquez-Samayoa to show that every step in two hypothetical chains was more likely than not to occur, the Board increased Velasquez-Samayoa’s CAT burden. The panel explained that Velasquez-Samayoa was not required to show that he was more likely than not to be tortured under both theories, nor was he required to show that he was more likely than not to be tortured under any single theory considered individually. Rather, the law required him to show only that, taking into account all possible sources of torture, he is more likely than not to be tortured. 4 VELASQUEZ-SAMAYOA V. GARLAND

The panel concluded that the Board also erred by disregarding credible testimony from Velasquez-Samayoa’s expert Dr. Thomas Boerman. The panel explained that, although the agency may reject credible testimony if it is outweighed by other more persuasive evidence, when the agency has credited an expert, it cannot reject that expert’s testimony for the sole reason that it is not corroborated by additional evidence. The panel wrote that the mere fact that Dr. Boerman’s testimony was not corroborated by country conditions evidence was not a valid reason for rejecting that testimony, as expert testimony can itself provide evidence of country conditions.

Acknowledging that the agency may point to other persuasive evidence in the record that contradicts a credible expert’s testimony, the panel concluded that to the extent the agency articulated a finding that other evidence in the record outweighed Dr. Boerman’s testimony, such a finding was unsupported by substantial evidence.

The panel remanded for the agency to properly assess the aggregate risk that Velasquez-Samayoa will be tortured if he is removed to El Salvador and, as part of that assessment, to properly consider Dr. Boerman’s testimony.

VELASQUEZ-SAMAYOA V. GARLAND 5

COUNSEL

Jean Reisz (argued) and Niels Frenzen, University of Southern California, Gould School of Law, Immigration Clinic, Los Angeles, California, for Petitioner.

Sarah E. Witri (argued), Trial Attorney; Jennifer P. Levings, Senior Litigation Counsel; Brian M. Boynton, Acting Assistant Attorney General; Shelley R. Goad, Assistant Director; United States Department of Justice, Civil Division, Office of Immigration Litigation, Washington, D.C.; for Respondent.

OPINION

FRIEDLAND, Circuit Judge:

Miguel Angel Velasquez-Samayoa seeks protection under the Convention Against Torture (“CAT”). Velasquez- Samayoa asserts that, if he were removed to his native country of El Salvador, he would be identified as a gang member and therefore would face a significant risk of being killed or tortured—either by Salvadoran officials or by members of a rival gang with the acquiescence of the Salvadoran government. The Board of Immigration Appeals (“BIA”) upheld a decision by an Immigration Judge (“IJ” and, collectively with the BIA, the “Agency”) concluding that neither potential source of torture poses a sufficient risk to entitle Velasquez-Samayoa to CAT relief. Velasquez- Samayoa argues before our court that the Agency failed to assess the aggregate risk that he will be tortured— considering all sources together. We agree. We grant the petition for review and remand to the Agency to reconsider 6 VELASQUEZ-SAMAYOA V. GARLAND

Velasquez-Samayoa’s CAT claim, applying the correct legal standards.

I.

Velasquez-Samayoa came to the United States when he was two or three years old and has lived in this country for more than forty years since. He became a lawful permanent resident in 1995, when he was seventeen. He has not visited El Salvador since approximately 1993, and his entire family now lives in the United States.

When he was about fifteen and residing in the Los Angeles area, Velasquez-Samayoa joined the White Fence gang. The White Fence gang is a rival of the Mara Salvatrucha (“MS-13”) gang in Los Angeles, and Velasquez-Samayoa often fought with MS-13 members. During his time in the White Fence gang, Velasquez- Samayoa had the letters “WF” tattooed on his body in two places—both letters on his neck and one letter on each of his legs. Each letter on his neck is approximately six inches tall and three inches wide and is written in a calligraphy script common for gang tattoos.

In 1998, Velasquez-Samayoa was convicted of multiple felonies. He was sentenced to eighteen years in prison for those crimes. Upon moving prison facilities about seven years into his sentence, he was stabbed over twenty times by gang members and was placed in protective custody for the remainder of his sentence. Velasquez-Samayoa later testified that, with the help of a prison psychologist, he “realize[d he] had wasted [his] entire life with gangs.”

After Velasquez-Samayoa’s release from prison, officers from the Department of Homeland Security detained him and initiated removal proceedings. An IJ determined that

VELASQUEZ-SAMAYOA V. GARLAND 7

Velasquez-Samayoa was removable based on his felony convictions. Velasquez-Samayoa does not dispute that, in light of that determination, the only relief he is eligible to seek is deferral of removal under the CAT.

An IJ held a hearing on Velasquez-Samayoa’s CAT claim in June 2020. Velasquez-Samayoa testified in support of his claim, as did Dr. Thomas Boerman, who has conducted research on gang activity and violence in El Salvador. The Government stipulated that Dr. Boerman qualified as an expert on those topics. The IJ found that both Velasquez-Samayoa and Dr. Boerman “testified credibly” and therefore “afford[ed] their testimonies full evidentiary weight.”

Velasquez-Samayoa testified to his fear that, if he were removed to El Salvador, he would be killed or tortured either by Salvadoran officials or by rival gang members. He explained that his conspicuous neck tattoo and/or the records of his criminal convictions would reveal his prior gang affiliation. Velasquez-Samayoa declared: “[B]oth of these parties [i.e., Salvadoran officials or rival gang members] will eventually kill me for these reasons.”

Dr. Boerman agreed that Velasquez-Samayoa faces a “high risk of egregious physical harm and death if returned to El Salvador.” Dr. Boerman testified that, if Velasquez- Samayoa were removed to El Salvador, he would likely be identified as a gang member because of his neck tattoo or because of information the United States would send to the Salvadoran government.

Dr. Boerman testified that, if Velasquez-Samayoa were identified as a gang member by Salvadoran officials, he would likely be tortured in government custody. According to Dr. Boerman, even if Velasquez-Samayoa were not 8 VELASQUEZ-SAMAYOA V. GARLAND

identified by government authorities immediately upon entering the country, he would eventually encounter police officers who would target and potentially kill him because of his obvious gang affiliation.

Dr. Boerman also spoke about the risk that Velasquez- Samayoa would be tortured or killed by rival gang members if he were removed. Dr. Boerman testified that the MS-13 and Barrio 18 gangs, both rivals of the White Fence gang, are present in 95 percent of Salvadoran municipalities. He stated that White Fence is “well-known by MS-13 and Barrio 18 in El Salvador” and that gangs in El Salvador routinely kill or torture members of rival gangs, and do so with impunity. And Dr. Boerman pointed to Velasquez- Samayoa’s status as a middle-aged man—which would make him appear to be a gang leader—and mannerisms he acquired growing up in the United States as other factors that would make him a target for torture.

The IJ issued a written decision denying Velasquez- Samayoa CAT protection and ordering him removed to El Salvador. The IJ acknowledged that Velasquez-Samayoa claimed that he would be either detained and then tortured by Salvadoran officials or tortured by gang members with the acquiescence of the Salvadoran government. The IJ explained that a CAT applicant who brings a “claim for relief based on a future and hypothetical chain of events . . . must establish that every step in the ‘hypothetical chain of events is more likely than not to happen’” (quoting Matter of J-F-F-, 23 I. & N. Dec. 912, 917 (A.G. 2006)). The IJ then proceeded to “analyze[] whether there is sufficient evidence in the record to establish it is more likely than not that each of these steps will occur.” The IJ discussed some of the steps that Velasquez-Samayoa claimed might lead to his torture and determined that he had not shown that they were more likely than not to occur.

VELASQUEZ-SAMAYOA V. GARLAND 9

Velasquez-Samayoa appealed the order of removal to the BIA. The BIA affirmed the IJ’s decision, concluding that the IJ “properly determined that [Velasquez-Samayoa] had not demonstrated that each link in the chain of hypothetical events will more than likely occur.”

Velasquez-Samayoa filed a petition for review in our court. He argues that the Agency’s CAT analysis was fundamentally flawed in that it failed to consider his aggregate risk of torture. He also argues, among other things, that the Agency erred by requiring that the testimony of Dr. Boerman—who was deemed credible by the IJ—be corroborated by evidence from the country conditions reports.

II.

“We review issues of law regarding CAT claims de novo.” Cole v. Holder, 659 F.3d 762, 769 (9th Cir. 2011). We review the factual findings underlying the BIA’s decision that an applicant is not eligible for CAT relief for substantial evidence. Id. at 770. “Under the substantial evidence standard, the court upholds the BIA’s determination unless the evidence in the record compels a contrary conclusion.” Id. (quoting Arteaga v. Mukasey, 511 F.3d 940, 944 (9th Cir. 2007)).

III.

Velasquez-Samayoa’s sole claim for relief is deferral of removal under the CAT. To prevail on that claim, the burden is on Velasquez-Samayoa “to establish that it is more likely than not that he . . . would be tortured if removed to the proposed country of removal.” 8 C.F.R. § 1208.16(c)(2) (providing the standard for withholding of removal under the CAT); see id. § 1208.17(a) (providing the standard for 10 VELASQUEZ-SAMAYOA V. GARLAND

deferral of removal under the CAT). To qualify for CAT protection, an applicant must “show only a chance greater than fifty percent that he will be tortured if removed.” Cole v. Holder, 659 F.3d 762, 770 (9th Cir. 2011) (quoting Hamoui v. Ashcroft, 389 F.3d 821, 827 (9th Cir. 2004)).

In Cole, we made clear that the standard for granting CAT relief does not change when an applicant contends that he would face risks of torture from multiple distinct sources if he were removed to his country of origin. Such an applicant must establish that, “taking into account all possible sources of torture, he is more likely than not to be tortured.” Id. at 775. Thus, in assessing a CAT claim from an applicant who has posited multiple theories for why he might be tortured, the relevant inquiry is whether the total probability that the applicant will be tortured—considering all potential sources of and reasons for torture—exceeds 50 percent. See Quijada-Aguilar v. Lynch, 799 F.3d 1303, 1308 (9th Cir. 2015) (“CAT claims must be considered in terms of the aggregate risk of torture from all sources, and not as separate, divisible CAT claims.”). For example, if an applicant is at risk of torture from “police, death squads, and gangs,” he “need not prove that each group, treated individually, would more likely than not torture him.” Cole, 659 F.3d at 775.

The Attorney General has instructed that, if an applicant would be tortured only if a single “hypothetical chain of events” comes to fruition, CAT relief cannot be granted unless each link in the chain is “more likely than not to happen.” Matter of J-F-F-, 23 I. & N. Dec. 912, 917-18 (A.G. 2006). The Attorney General’s rationale for that rule was that “a [single] chain of events cannot be more likely

VELASQUEZ-SAMAYOA V. GARLAND 11

than its least likely link.” 1 Id. at 918 n.4. For example, in J-F-F-, the CAT applicant’s theory of torture was premised on the suppositions

that [the applicant] needs medication in order to behave within the bounds of the law; that such medication is not available in the Dominican Republic; that as a result [the applicant] would fail to control himself and become “rowdy”; that this behavior would lead the police to incarcerate him; and that the police would torture him while he was incarcerated.

Id. at 917. The Attorney General determined that, among other things, the applicant failed to show he was “more likely than not to be denied access” to his medication, so he had failed to meet his burden to show that he was more likely than not to be tortured on the theory he had presented. Id. at 919.

Our holding in Cole and the rule articulated in J-F-F- are logically consistent. Cole provides that, when an applicant posits multiple theories for why he would be tortured, the Agency should consider the aggregate risk posed by all sources and grant CAT relief if the cumulative probability of torture is greater than 50 percent. J-F-F- provides that, when an applicant posits a single theory for why he would be tortured, but the torture will come about only if several hypothetical events all occur in sequence, an applicant must

1 J-F-F- must be applied carefully—if it is applied at all. Adjudicating a CAT claim will generally involve speculation about the likelihood of future events, and it will not be possible, or even desirable, to quantify probabilities precisely. 12 VELASQUEZ-SAMAYOA V. GARLAND

show, at a minimum, that the individual probability of each event occurring is greater than 50 percent. It is possible that an applicant might bring a CAT claim in which he posits multiple theories for why he might be tortured, but one or more of those theories entails a hypothetical chain of events occurring in sequence. In such a case, the rule articulated in J-F-F- will often be of little practical help to the Agency. Even if the Agency were to determine that the applicant did not carry his CAT burden on the basis of any one theory, considered individually, the Agency’s work would not be done. It would still have to assess whether the applicant’s aggregate risk of torture—considering all theories collectively—entitled him to CAT relief.

Here, the BIA misapplied our precedent and J-F-F-. The BIA acknowledged that Velasquez-Samayoa posited at least two separate theories for why he might be tortured if he were removed to El Salvador. Under the first theory, the BIA explained, he would be identified as a gang member by Salvadoran officials, and he would be killed or tortured in the custody of those officials. Under the second theory, the BIA explained, he would be identified as a gang member by rival gangs, and he would be killed or tortured by those gangs.

After expressly recognizing that Velasquez-Samayoa presented “alternative” theories of torture, the BIA stated that “the Immigration Judge properly determined that [Velasquez-Samayoa] had not demonstrated that each link in the chain of hypothetical events will more than likely occur” (citing Medina-Rodriguez v. Barr, 979 F.3d 738, 750-51 (9th Cir. 2020); J-F-F-, 23 I. & N. Dec. at 917-18). The BIA proceeded to consider some of the steps that might lead to Velasquez-Samayoa’s torture under one theory or the other and opined generally that those steps were not more likely than not to occur. In the process, the BIA referred to

VELASQUEZ-SAMAYOA V. GARLAND 13

Velasquez-Samayoa’s two alternative theories of torture as a “claimed chain of events that would lead to his torture.” Finally, the BIA concluded that the IJ properly denied Velasquez-Samayoa’s CAT claim, citing Medina-Rodriguez for the statement that “the evidence does not establish that any step in the claimed hypothetical chain of events is more likely than not to happen, let alone that the entire chain will come together to result in the probability of torture.” See 979 F.3d at 750-51.

The BIA erred by failing to assess Velasquez-Samayoa’s “overall risk of being tortured.” Cole, 659 F.3d at 775. The BIA considered his two separate theories of torture as a single hypothetical chain of events and denied his CAT claim because the probability of that hypothetical chain occurring was not high enough. But the BIA should not have considered his claim as a single hypothetical chain of events, when—as the BIA itself acknowledged—he posited two “alternative” and distinct theories for why he would be tortured if he were removed to El Salvador. By requiring Velasquez-Samayoa to show that every step in two hypothetical chains was more likely than not to occur, the BIA increased his CAT burden. Velasquez-Samayoa was not required to show that he was more likely than not to be tortured under both theories, nor was he required to show that he was more likely than not to be tortured under any single theory considered individually. See Quijada-Aguilar, 799 F.3d at 1308. The law requires him to show only that, “taking into account all possible sources of torture, he is more likely than not to be tortured.” Cole, 659 F.3d at 775. Thus, the BIA should have assessed whether aggregating the risks posed by Velasquez-Samayoa’s two theories results in a probability greater than 50 percent that he will be tortured. See id. (“The BIA erred by . . . never assessing [the applicant’s] overall risk of being tortured.”). 14 VELASQUEZ-SAMAYOA V. GARLAND

Despite the BIA’s reliance on it, our decision in Medina- Rodriguez does not support the Agency’s approach. In that case, the applicant claimed that he would likely be tortured if removed to Mexico for two reasons: (1) he had a physical disability, and the Mexican healthcare system subjected individuals with disabilities to abuse, and (2) he had twenty tattoos, which would attract the attention of cartels, which would result in the applicant’s being tortured by a cartel. Medina-Rodriguez, 979 F.3d at 750-51. That second theory of torture risk relied on multiple events all occurring in sequence. In assessing that second theory, we held that “[t]he evidence does not establish that any step in this hypothetical chain of events is more likely than not to happen, let alone that the entire chain will come together to result in the probability of torture.” Id. at 751 (alteration in original) (quoting J-F-F-, 23 I. & N. Dec. at 917-18). We cited J-F-F- to explain why the applicant’s second theory was not sufficient in and of itself to sustain his burden of showing he was more likely than not to be tortured if removed to Mexico. In assessing the applicant’s risk of torture on account of his first theory, we decided that the BIA properly rejected that theory because there was no evidence in the record to support the proposition that individuals in Mexico faced torture because of physical disabilities. Medina-Rodriguez, 979 F.3d at 750. Because we had already determined that the applicant did not face any risk of torture stemming from his physical disability, our decision that the applicant could not meet his CAT burden on the basis of his only other theory was dispositive. Here, by contrast, the Agency never made a determination that Velasquez-Samayoa faced no risk of torture on the basis of either of his theories, so the Agency’s application of J-F-F- was inadequate to assess whether he had met his CAT burden.

VELASQUEZ-SAMAYOA V. GARLAND 15

IV.

Velasquez-Samayoa argues that the Agency also erred by disregarding credible expert testimony. We agree.

“The regulations implementing CAT explicitly require the IJ to consider ‘all evidence relevant to the possibility of future torture.’” Aguilar-Ramos v. Holder, 594 F.3d 701, 705 n.6 (9th Cir. 2010) (quoting 8 C.F.R. § 208.16(c)(3)). “If the [Agency] rejects expert testimony, it must state ‘in the record why the testimony was insufficient to establish the probability of torture.’” Castillo v. Barr, 980 F.3d 1278, 1283 (9th Cir. 2020) (quoting Cole v. Holder, 659 F.3d 762, 772 (9th Cir. 2011)). Although the Agency may reject credible testimony if it is “outweighed by other more persuasive evidence,” Garland v. Ming Dai, 141 S. Ct. 1669, 1681 (2021) (emphasis added), when the Agency has credited an expert, the Agency cannot reject that expert’s testimony for the sole reason that it is not corroborated by additional evidence, see Castillo, 980 F.3d at 1284 (“If an expert’s opinion could only be relied upon if it were redundant with other evidence in the record, there would be no need for experts.”).

The BIA improperly rejected Dr. Boerman’s testimony. The BIA affirmed the IJ’s decision that Dr. Boerman was a credible witness. Yet the BIA proceeded to reject key pieces of Dr. Boerman’s testimony regarding why Velasquez- Samayoa faced a high risk of torture, agreeing with the IJ that “the country conditions evidence did not corroborate the expert’s testimony that [Velasquez-Samayoa] will be perceived as a gang leader based on his age and other characteristics.” The mere fact that Dr. Boerman’s testimony is not corroborated by country conditions evidence is not a valid reason for rejecting that testimony— expert testimony can itself provide evidence of country 16 VELASQUEZ-SAMAYOA V. GARLAND

conditions. See id. To be sure, the Agency may point to other persuasive evidence in the record that contradicts a credible expert’s testimony. See Ming Dai, 141 S. Ct. at 1681. Here, the Agency did not do so.

The BIA did refer to the IJ’s additional explanation for rejecting Dr. Boerman’s testimony that Velasquez-Samayoa was likely to be targeted based on his age. The IJ stated that “the documentary evidence establishes that [in El Salvador] children and young adults are most vulnerable to gang- related violence.” To the extent the Agency was articulating a finding that other evidence outweighs Dr. Boerman’s testimony, we reject that finding as unsupported by substantial evidence. The IJ cited to the declaration of another one of Velasquez-Samayoa’s experts, as well as to several pages about gang-related violence in El Salvador from a country conditions report. But neither piece of evidence contradicts or even casts doubt on Dr. Boerman’s testimony regarding the risk Velasquez-Samayoa faces due to his relatively advanced age. Dr. Boerman testified that there is a “very short life expectancy” among gang members and that “if you are considered a gang member, who lives to the age of 40, by definition, you’re a leader, and if you’re a leader, you are far more of concern to everyone.” Dr. Boerman’s point was that relatively few gang members live to be forty, so—given that Velasquez-Samayoa is more than forty years old and that Dr. Boerman opined that Velasquez- Samayoa will be readily identified as a gang member—he is likely to be targeted as a gang leader. The fact that younger people in El Salvador are more likely overall to be associated with gangs than older people in El Salvador, and therefore more likely to be affected by gang violence, does not refute Dr. Boerman’s testimony that Velasquez-Samayoa is at a high risk of torture because he is likely to be identified as an older gang member.

VELASQUEZ-SAMAYOA V. GARLAND 17

V.

For the reasons above, we grant the petition for review and remand the case to the Agency to properly assess the aggregate risk that Velasquez-Samayoa will be tortured if he is removed to El Salvador and, as part of that assessment, to properly consider Dr. Boerman’s testimony. 2 Because the Agency must, on remand, consider “all evidence relevant to the possibility of future torture,” 8 C.F.R. § 208.16(c)(3), we need not address Velasquez-Samayoa’s other arguments that the Agency failed to consider all his evidence.

GRANTED AND REMANDED.

2 Because Velasquez-Samayoa has asserted at least two distinct theories for why he would be tortured if he were removed to El Salvador, the Agency should bear in mind, when assessing the aggregate risk of torture, that certain pieces of evidence—such as his prominent neck tattoo—may be relevant to both theories.