Francisco Ornelas-Chavez v. Alberto R. Gonzales, Attorney General

O’SCANNLAIN, Circuit Judge,

dissenting.

In this case our modest task is to decide whether the Board of Immigration Appeals’ (“BIA”) decision to deny Petitioner Ornelas-Chavez’s claims was free of legal error and supported by substantial evidence. Because neither sympathy for the petitioner nor distrust of the BIA or the Immigration Judge (“IJ”) should compromise a careful assessment of the administrative decisions and record evidence, I respectfully dissent.

I

First, the court holds that the BIA applied the wrong legal standard to OrnelasChavez’s claim to withholding of removal under 8 U.S.C. § 1231 (b)(3) because it imposed a strict “reporting requirement” to establish government persecution. Maj. Op. at 1056-58. With respect, it did no such thing.

A

Any fair reading of the BIA’s decision demonstrates that it applied the proper legal standard. In relevant part, the BIA stated:

The respondent suffered an incident .of harm, a detention of several hours apparently at this [sic] request of his father, at the hands of government agents.. This single incident does not rise to the level of persecution. All of the other harm suffered by the respondent occurred at the hands of private citizens. The respondent did not report any of these incidents to government authorities. '

... Accordingly, where the respondent never reported his incidents of harm to government authorities, and where the background evidence in the record is inconclusive, the Immigration Judge properly found that the respondent did not prove that the Mexican government is unwilling or unable to control those who harmed or may harm him.

(Emphasis added.) The majority, adopting Ornelas-Chavez’s argument wholesale, seems to fixate on the last sentence of the first quoted paragraph: “The respondent did not report any of these incidents to government authorities.” Indeed, with only a passing nod to the sentence’s context, the majority distorts entirely the BIA’s reasoning.

The BIA did not apply nor did it adopt a strict “reporting requirement.” True, it observed that there was no evidence that Ornelas-Chavez reported the alleged incidents. However, as the majority concedes, it then considered this dearth of evidence in combination with the background country conditions evidence, which failed to demonstrate a pattern of indifference by the government. The BIA concluded, in light of these categories of evidence, that the IJ could appropriately find “that the respondent did not prove that the Mexican government is unwilling or unable to control those who harmed or may harm him.”

Under our circuit’s case law, this is an entirely proper mode of analysis. The persecution of which Ornelas-Chavez complains must be shown to have been perpetrated by the government, or else OrnelasChavez must show that the government “is unwilling or unable to control those elements of its society responsible for targeting a particular class of individuals.” Ave *1062 tova-Elisseva v. INS, 213 F.3d 1192, 1196 (9th Cir.2000) (citation and internal quotation marks omitted). Accordingly, whether the petitioner has reported the incidents to the authorities is clearly relevant to, even if not dispositive of, the ultimate question of whether the government was “unwilling or unable to control” the persecutors. See, e.g., Baballah v. Ashcroft, 367 F.3d 1067, 1078 (9th Cir.2004) (citations omitted).

In this light, it is clear that the BIA’s consideration of general country conditions stood to confirm its conclusion that Ornelas-Chavez did not meet his burden. That the BIA considered such background evidence itself conclusively demonstrates that the inquiry was not terminated by any rigid reporting requirement. Ultimately, I think it clear that the BIA asked and answered the right question: whether Ornelas-Chavez’s alleged persecutors were parties whom the Mexican authorities were unwilling or unable to control. The BIA was correct to treat Ornelas-Chavez’s failure to report the alleged persecution as relevant to that question.

In arriving at its errant conclusion, the majority observes that the BIA “cited only two pieces of evidence,” and since one of them was inconclusive, it must have been improper that the other — the alleged “reporting requirement” — was “a sine qua non for the success of Ornelas-Chavez’s withholding of removal claim.” Maj. Op. at 1057, 1058. But where, I wonder, is the flaw in the BIA’s analysis?

Presumably Ornelas-Chavez had two options: to prove that the government was unwilling or unable to control such persecution generally, or to prove that it was unwilling or unable to control his persecutors specifically. See, e.g., Castro-Perez v. Gonzales, 409 F.3d 1069, 1072 (9th Cir.2005) (considering country reports to determine whether the Honduran government was willing or able to control rape generally, and considering the petitioner’s specific situation). Since Ornelas-Chavez failed to do the former, to meet his burden it' was necessary that he have done the latter. And because there is little other evidence in the record to suggest .the government’s disposition towards OrnelasChavez’s particular troubles, the BIA properly deemed rather important in this case — “a sine qua non,” even! — that Ornelas-Chavez failed to report the alleged abuse to government authoritiés. 1 So, the majority concedes that reporting vel non may be relevant, see Maj. Op. at 1057-58; yet it grants the petition because in this particular case the failure to report is too relevant.

The BIA did not, as the majority seems to suggest, improperly consider background country conditions as “establishing] that specific acts of persecution occurred or did not occur.” Maj. Op. at 1056. Although the BIA, having made no adverse credibility determination, was required to accept Ornelas-Chavez’s testimony as factually true, it was not required to adopt his preferred legal conclusion: that the authorities would not have acted to curb the persecution he alleges took place. The background evidence in the record was obviously relevant to the merits of that issue, and the BIA was entitled, if not obligated, to consider it. See Castro-Perez, 409 F.3d at 1072 ; Avetova-Elisseva, 213 F,3d at 1198-99; Andriasian v. INS, 180 F.3d 1033, 1042-43 (9th Cir.1999). 2 *1063 The testimony provided by Ornelas-Chavez, while also relevant, was evidence of an anecdotal, subjective variety and thus not particularly probative. Indeed, the background evidence substantially negates Ornelas-Chavez’s perception. See infra section I.B.l.

Our standard of review in immigration cases is well-established: we must affirm the BIA if the record contains substantial evidence for its decision; we cannot justify granting the petition for review merely by pointing to other record evidence which, in our own view, supports the alien’s claim. See INS v. Elias-Zacarias, 502 U.S. 478 , 481 & n. 1, 112 S.Ct. 812 , 117 L.Ed.2d 38 (1992). The majority cannot say that Ornelas-Chavez’s testimony compels a rational fact finder to conclude that the government was unwilling or unable to control his alleged persecutors. Thus, the majority opts for another, more innovative approach: it holds that the BIA applied the wrong legal standard — simply because there exists supportive evidence that the BIA failed explicitly to discuss. See Maj. Op. at 1057-58. But Ornelas-Chavez’s testimony was not somehow legally distinct from the balance of the record; it was merely additional evidence on the same “unwilling or unable to control” question. Accordingly, the majority’s holding seems little more than a willful circumvention of the Supreme Court’s prescribed standard of review.

In short, the majority goes to great lengths to divine in the BIA’s analysis some kind of per se rule, which it can then merrily reject. As I have sought to demonstrate, such “reporting requirement” is of the majority’s own imagining. Surely one may infer much from the typically brief text of an administrative decisionespeeially when one is all too eager to conjure a meaning that will justify granting the petition for review. Cf. Kumar v. Gonzales, 435 F.3d 1019, 1037 (9th Cir. 2006) (Kozinski, J., dissenting) (“[T]he majority picks apart the [BIA’s] findings piece by piece, scrutinizing [its] every sentence as if it is completely unconnected to the rest of the opinion.”). Such creative invention, however, is not appropriate to our role in reviewing the lawful adjudication of an administrative body.

B

Had the majority given the BIA’s decision a fair reading, it could not then hold that Ornelas-Chavez met his burden of showing that the evidence would compel any reasonable fact finder to conclude that he was more likely than not to suffer persecution upon his return to Mexico. Singh v. INS, 134 F.3d 962, 966 (9th Cir.1998).

1

There is no evidence of past persecution, as defined by the regulations, in this case. The Mexican government was not involved in any persecution of Ornelas-Chavez; the single, six-hour detention by a reluctant local police chief, at the request of Ornelas-Chavez’s father, does not rise to that level. See, e.g., Prasad v. INS, 47 F.3d 336, 339-40 (9th Cir.1995).

Critically, the evidence also falls short of compelling a finding that any past private “persecution” at issue was perpetrated by parties whom the government was unwilling or unable to control. Ornelas-Chavez has not “convincingly established” — or proven by any standard — that his resort to government authorities would have been futile. Cf. Korablina v. INS, 158 F.3d 1038, 1045 (9th Cir.1998). As the BIA found, the relevant background evidence is, at best, inconclusive. Ornelas-Chavez relies on little more than, a bare assumption that the police would have done nothing had he reported the abuse — a showing that we have previously deemed insufficient. See Castro-Perez, 409 F.3d at 1072 .

*1064 The administrative record provides little support for Ornelas-Chavez’s claim of past persecution. .The evidence regarding general country conditions, while perhaps revealing societal disfavor of homosexuals, does not demonstrate that the authorities would have been inattentive to the primary incidents of persecution alleged here: the rape and physical abuse of a young child. Even if the majority were willing to overlook the relevance of Ornelas-Chavez’s age — allowing him to gild his allegations in the armor of sexual orientation and gender identity — there is no evidence in the record that the Mexican authorities would have viewed the incidents in similar terms and thus permitted the abuse.

Even assuming that a few individual police officers were beholden to OrnelasChavez’s father because they rented a house on his land, Ornelas-Chavez makes no allegation that this potential bias somehow seeped into all (or even most) persons holding Mexican authority. It is true that Ornelas-Chavez’s teacher failed to take any action in response to his allegations of sexual contact. But at best, this testimony establishes one teacher’s ineptitude; it does not compel a finding that the Mexican authorities generally would have been indifferent to the rape of a young child. 3 Lastly, I note that when, at the prison at which he worked, Ornelas-Chavez informed his supervisor of the abuse by his coworkers, the supervisor exhibited a willingness to control the alleged conduct by offering to adjust Ornelas-Chavez’s shift so that he could avoid his would-be persecutors. Ornelas-Chavez cannot prove that this measure would have been inadequate because he chose to resign his employment. That the supervisor “took no action against the co-workers,” Maj. Op. at 1054, is of little relevance.

The various items of evidence OrnelasChavez put forth fail to support his claim because his subjective perceptions of government officials are not relevant — much less dispositive. The critical inquiry as to government responsiveness is objective, and we have always considered the “unwilling or unable to control” requirement in that context. See, e.g., Malty v. Ashcroft, 381 F.3d 942, 948 (9th Cir.2004). Ornelas-Chavez’s personal belief that the relevant Mexican authorities would have been unwilling to shield him from persecution is grossly insufficient.

In sum, while I do not discount nor make light of the harms Ornelas-Chavez has apparently suffered, I must also eon- *1065 elude that the evidence does not compel a finding — as it must do under our limited standard of review — that the harms were occasioned by persons whom the government was unwilling or unable to control.

2

Having failed to produce compelling evidence of past persecution by actors whom the government was unwilling or unable to control, Ornelas-Chavez would not be entitled to a presumption' of future persecution. See 8 C.F.R. § 1208.13 (b)(1). His claim to withholding of removal would then fail if he also falls short of independently establishing a clear probability of future persecution. See Lim v. INS, 224 F.3d 929, 938 (9th Cir.2000).

a

On this record, there simply is insufficient showing that homosexuals are currently subject to official persecution in Mexico. The U.S. Department of State’s June 1997 country report documented “no evidence of systematic official persecution of homosexuals,” and Ornelas-Chavez’s other evidence was not to the contrary. A report from a private group, dated May 2000, stated that “repression by ... authorities is now the exception, not the rule,” and noted that societal attitudes are trending towards tolerance.

The record evidence further demonstrates extensive legal, political, and cultural advances by homosexuals in Mexico, including the enactment of legislation prohibiting discrimination on the basis of sexual orientation, wide acceptance of participation by homosexuals in two of the three major political parties, and the election of openly homosexual government representatives. To the extent private persecution occurs, the record suggests that the government’s response has changed. One news report in the record, for example, notes that the Mexico City police “have set up a unit specializing in dealing with homophobic crimes, and are to get sensitivity training.”

b

Nor was the evidence so compelling that any rational fact finder must conclude that Ornelas-Chavez, individually, would more likely than not be subject to persecution by actors whom the government was unwilling or unable to control. 8 C.F;R. § 1208.16(b)(l)-(2); see Sael v. Ashcroft, 386 F.3d 922, 925 (9th Cir.2004); Hoxha v. Ashcroft, 319 F.3d 1179, 1185 (9th Cir.2003). While the record reveals societal disfavor of homosexuality generally, it exposes few incidents of actual persecution. These general country conditions do not compel reversal of the BIA. More directly, there is simply no evidence that OrnelasChavez is likely to be singled out by the government or by private forces whom the government is unwilling or unable to control. 4

Read as a whole, the record is not such that a reasonable fact finder must conclude that Ornelas-Chavez is more likely than not to be persecuted in the future if he is returned to Mexico. 5 I would affirm *1066 the BIA’s decision because it is supported by substantial evidence.

II

I must also dissent from the majority’s holding that the BIA applied an incorrect legal standard as to the level of government involvement required for relief under the Convention Against Torture (“CAT”). Maj. Op. at 1058-61. Again, the majority’s reading of the relevant administrative decision is partisan and inaccurate. Further, there is absolutely no evidence of torture of homosexuals by the Mexican government (or by private parties with the government’s consent or acquiescence) in this record. Random incidents of private violence simply cannot support the notion that official torture of homosexuals is systematic or even common. As such, the IJ’s decision on the CAT claim is supported by substantial evidence.

A

The applicable regulations define “torture” as “severe pain or suffering, whether physical or mental” when intentionally inflicted “with the consent or acquiescence of a public official or other person acting in an official capacity.” 8 C.F.R. § 208.18 (a)(1) (emphasis added). In this case, the IJ found that Ornelas-Chavez had “provided no evidence of past torture or any mental or physical intentionally inflicted severe pain or suffering that is sanctioned by a public official.” (Emphasis added.) The majority seizes upon the word “sanctioned” to assert an erroneous application by the IJ of a legal standard “higher” than that prescribed in the regulations. Maj. Op. at 1058.

While the IJ may have employed imprecise diction in that passage, the majority is again too keen to find fault in the agency decision we now review. Earlier in its order, the IJ explained, in part, that “[t]he Respondent presented no evidence that the authorities refused to protect him or that the authorities did not protect him, since he never reported any of these incidents to the authorities.” Id. at 1059 . This phraseology indicates that the IJ considered whether the authorities were aware of the alleged abuses such that they could possibly “consent or acquiesce.” Thus, the IJ necessarily found that Ornelas-Chavez’s evidence fell short of even the proper “consent or acquiescence” standard. Although the word “sanctioned” arguably suggests affirmative approval by the government, there is no indication in the opinion that the IJ actually reviewed the evidence for conduct beyond the government’s “tacit or passive acceptance” or “implied consent to an act.” I also find it significant that the IJ’s errant word choice was simply part of its summary declaration at the end of the opinion, not in any way enmeshed in the actual analysis of the facts.

There was not, as the majority suggests, implicit in the IJ’s finding a standard “also higher than that dictated by CAT.” Maj. Op. at 1059. In so asserting, the majority again lays bare the analytical flaw that infects its entire opinion: it posits that “the IJ’s decision rests on the unwarranted premise that the only way a public official *1067 can have such awareness '... is if the applicant reports the alleged torture to him.” Id. at 1060 . Reporting is not the only way a public official can have such awareness, but in this case there was absolutely no suggestion of an alternative. No government agent witnessed the alleged torture; there was'no suggestion that someone else may have told the government' about it; and there was insufficient evidence that the government is willfully blind to such torture generally. The IJ’s statement, read in context, was clearly an assessment of the facts of this particular case. It was not an erroneous, general statement of law upon which we could grant the petition for review.

In sum, a fair reading of the decision leads to the conclusion that the word “sanctioned” was inconsequential to the analysis and ultimately harmless. 6 Yet the majority excitedly reverses the IJ on the CAT claim as if one errant word choice, bereft of context and contained in a summary sentence, constituted the smoking gun for which it was searching. See Kumar, 435 F.3d at 1035 (Kozinski, J., dissenting) (“This is the nub of the IJ’s reasoning, which the majority ignores, preferring to nitpick the IJ’s isolated statements.”). Rather than condemn the IJ solely on the authority of Black’s Law Dictionary and Merriam-Webster’s Collegiate, I would look to the mode of analysis it actually employed. The IJ applied the correct standard and committed no error of any consequence to this case.

B

Because, in my view, the IJ did not apply an erroneous standard to his CAT claim, the final step in our review is to consider whether the record evidence is so compelling that any rational finder of fact would conclude that Ornelas-Chavez more likely than not would be tortured, with the consent or acquiescence of the government, upon his return to Mexico. Ochoa v. Gonzales, 406 F.3d 1166, 1172 (9th Cir.2005). Clearly, the record evidence does not even come close to meeting this stringent standard.

1

The majority seeks to curtail such inquiry by suggesting that “the IJ concluded simply that Ornelas-Chavez failed to show that the authorities ‘sanctioned’ his alleged torture because he did not report it,” and that we therefore cannot affirm the IJ’s decision on any other ground even if supported by substantial evidence. Maj. Op. at 1060.

But the IJ’s decision was not so cabined. She found no past torture because, in part, Ornelas-Chavez did not report any abuse. But the majority ignores at least four paragraphs in which the IJ discussed the facts that she felt undermined OrnelasChavez’s claims of future harm. She considered “[Petitioner’s] testimony and evidence offered by the[Petitioner],” and concluded that Ornelas-Chavez had “not shown that he is more likely than not to be tortured if he were to be removed to Mexi *1068 co.” Given that conclusion, the IJ necessarily found many of the facts the majority claims it did not find. 7

As such, we should consider any record evidence that supports the IJ’s finding that Ornelas-Chavez failed to meet his burden of showing a likelihood of torture upon his return to Mexico. See Turcios v. INS, 821 F.2d 1396, 1398 (9th Cir.1987) (“The substantial evidence test is essentially a case-by-case analysis requiring review of the whole record.”). 8

2

As already noted, the regulations define torture for CAT purposes as severe pain or suffering “inflicted by or at the instigation of or with the consent or acquiescence of a public official or other person acting in an official capacity.” 8 C.F.R. § 208.18 (a)(1) (emphasis added). In turn, “[ajcquiescence of a public official requires that the public official, prior to the activity constituting torture, have awareness of such activity and thereafter breach his or her legal responsibility to intervene to prevent such activity.” Id. § 208.18(a)(7) (emphasis added); see Kamalthas v. INS, 251 F.3d 1279, 1282 (9th Cir.2001). In short, the “consent or acquiescence” requirement means that the government must be aware of the allegedly tortuous conduct, or at least willfully blind to it. Zheng v. Ashcroft, 332 F.3d 1186, 1188-89 (9th Cir.2003) (citation omitted).

Past torture does not create a presumption of entitlement to CAT relief, although it is relevant to the ultimate inquiry: whether “it is more likely than not that [the petitioner] would be tortured if removed to the proposed country of removal.” 8 C.F.R. § 208.16 (c)(2); see Mohammed v. Gonzales, 400 F.3d 785, 802 (9th Cir.2005) (citing 8 C.F.R. § 1208.16 (c)(3)). There is no evidence on this record that the Mexican authorities were aware of or willfully blind to the past abuses perpetrated by Ornelas-Chavez’s father, cousins, or others. 9 By his own admission, Ornelas-Chavez never told anyone about the rapes he endured as a child. The lone exception was his confession to a school teacher that he had been sexually abused. However, the teacher was not informed of the alleged torture “prior to” its occurrence such that she could breach a legal duty to prevent it. See 8 C.F.R. § 208.18 (a)(7). 10 The allegation of an at *1069 tempt by Ornelas-Chavez’s co-workers to smother him with a pillow-which, if anything, constitutes only attempted torture-suffers from the same flaw: the supervisor was not aware of the incident until after its occurrence. Moreover,. as suggested above, the record does not compellingly demonstrate that the authorities were willfully blind to such conduct. As such, I must conclude that on this record there is no evidence of past torture, as defined by the regulations, to be considered under Ornelas-Chavez’s CAT claim.

As to the ultimate question, the likelihood that Ornelas-Chavez will be tortured in the future if removed to Mexico, the evidence utterly falls short. It is first worth noting that the sexual abuse took place during Ornelas-Chavez’s youth, specifically between the ages of six and twelve. Circumstances have so fundamentally changed — he is now an adult — 'that Ornelas-Chavez is rather unlikely to be subjected to future injury in the form of rape. As.noted, the same conclusion applies to the beatings to which OrnelasChavez was subjected by his father. And as to the attempted assault by OrnelasChavez’s co-workers at the state prison, there is simply no indication that these would-be torturers have singled out Ornelas-Chavez such that he would again be a target.

The evidence of general country conditions as it pertains to Mexican homosexuals does not paint a picture so dire as to compel reversal of the BIA decision. See 8 C.F.R. § 1208.16 (c)(3)(iii)-(iv). Quite the contrary, the record contains numerous references to an ever-improving situation. And although there have been some homosexuality-related murders documented over five years, there are, according to this record, as many as 12,600,000 Mexicans of homosexual orientation. 11 Thus, it is impossible to conclude that “more likely than not” Ornelas-Chavez would suffer this fate. Harassment and arbitrary detentions by police officers, to the extent they occur, do not constitute torture. More generally, the evidence does not rise to the level that has compelled us to grant a petition for review in prior cases. There is no record evidence alleging that “torture [is] routinely administered” to any group in Mexico, see Al-Saher, 268 F.3d at 1147; nor is there evideneé that torture is “an institution,” or otherwise so pervasive against homosexuals that we must conclude more likely than not Ornelas-Chavez would fall prey to it, see Khup v. Ashcroft, 376 F.3d 898, 907 (9th Cir.2004).

I would affirm the determination that Ornelas-Chavez failed to meet his burden of showing that he was more likely than not to be tortured, as specifically defined by the regulations, if returned to Mexico.

*1070 III

For all of the foregoing reasons, I respectfully dissent.

. There is no evidence that the authorities had (or should have had) knowledge of the rapes or beatings. Ornelas-Chavez presented evidence that government officials witnessed a couple incidents of harassment, but no incidents of persecution or torture.

. The majority’s attempt to limit the use of background country conditions evidence to credibility determinations, see Maj. Op. at 1056, is entirely without authority.

. Quite the contrary, the administrative record includes a U.S. Department of ■ State country report, dated March 31, 2003, which documents extensive efforts by the Mexican government to promote the well-being of the country’s children. The government has in place the National Institute for the Integral Development of the Family, which fields complaints of abuse against children. A 2000 law provides for imprisonment of up to ten years for the sexual corruption of a minor under 16 years of age; parents convicted of such crimes are stripped of custody over their children. Accomplices to sexual abuse face similar prison terms. In ,2000, the Mexican Congress also passed a constitutional amendment "to protect the rights of children and teenagers and ensure respect for their dignity.” Moreover, the Protection of the Rights of Children and Adolescents Law "provides for the right to life, nondiscrimination, healthy living conditions, protection against threat to liberty and physical abuse, a healthy family life, health services, equal treatment for persons with disabilities, education, pursuit of happiness, and freedom of thought and expression.” Penalties for violation include fines and imprisonment.

The same State Department country report also documents ongoing dangers faced by Mexico's youth, but the record evidence falls far short of compelling the conclusion that the government is "unwilling or unable to control” persecution of the type primarily alleged here.

. It is true that Ornelas-Chavez’s father may be the lone exception. But the record discloses no evidence that his father is seeking him out for the purpose of persecution such that it is likely to happen again. Cf., e.g., Mendoza-Perez v. INS, 902 F.2d 760, 762 (9th Cir.1990). Since Ornelas-Chavez reached adulthood, he has had only one run-in with his father. This incident, a 1998 altercation, was an isolated occurrence; it took place when the two encountered each other by chance at the home of Ornelas-Chavez’s aunt. Ornelas-Chavez himself acknowledged living in Mexico without incident for many years prior. As such, the incident does not compel a conclusion contrary to the one reached by the BIA.

. The showing made by Ornelas-Chavez is not nearly as compelling as those we have considered previously. In Boer-Sedano v. Gonzales, 418 F.3d 1082 (9th Cir.2005), the *1066 petitioner presented evidence that he had been raped repeatedly by a police officer. Id. at 1086. In Hernandez-Montiel v. INS, 225 F.3d 1084 (9th Cir.2000), the petitioner was raped twice by police officers. Id. at 1088 . In the latter case, the petitioner presented expert testimony and documentation showing that the government would not protect a particular subclass: gay men with female sexual identities. Id. at 1094 . Of course, the Hemandez-Montiel case is six years old — and much of the record evidence it cited is far older. The important point, in any case, is that Ornelas-Chavez has not made a remotely similar showing.

. The majority rejects the harmlessness of the misstatement on the basis of the government's argument, under an invalidated BIA decision, that Ornelas-Chavez must have been in the "custody or physical control” of a public official. Maj. Op. at 1059. But the IJ never asserted that Ornelas-Chavez must have been in the custody of a public official; rather, it applied the correct legal standard-that he must have been in the custody or physical control "of a perpetrator.” See Azanor v. Ashcroft, 364 F.3d 1013, 1019 (9th Cir.2004) (citing 8 C.F.R. § 208.18 (a)(6)). The majority, it seems, has seized on the government's curious argument in order to engage in a bit of misdirection. Obviously the government's litigation position has nothing to do with the IJ’s analysis, which, I would have thought, was the actual subject of our review.

. First, the IJ did indeed find that the rapes and beating Ornelas-Chavez suffered in his youth did not qualify as "torture,” because no government actor was aware of such abuses before they occurred. Second, the IJ found that Ornelas-Chavez had failed to prove the likelihood that he would be tortured if returned to Mexico. Third, the IJ’s finding that Ornelas-Chavez did not report any abuses to the government necessarily means that she found he did not inform anyone of torture "prior to” its occurrence.

. The cases cited by the majority say only that we will not apply in the first instance the correct legal standard where the agency applied an erroneous standard, see Azanor v. Ashcroft, 364 F.3d 1013, 1020-21 (9th Cir.2004), and that we will not decide a case on an entirely separate ground which the agency did not consider, see Navas v. INS, 217 F.3d 646 , 658 n. 16 (9th Cir.2000). Those cases are not applicable here.

. For purposes of this discussion, I assume that the rapes and beatings Ornelas-Chavez suffered at the hands of his cousins and father, respectively, constitute torture. See Al-Saher v. INS, 268 F.3d 1143, 1147 (9th Cir.2001) (sustained beatings), amended by 355 F.3d 1140 (9th Cir.2004); cf. Lopez-Galarza v. INS, 99 F.3d 954, 963 (9th Cir.1996) ("The effects of rape appear to resemble the effects of torture.”). I know of no other allegations of past conduct that would qualify under the regulatory definition.

As to the prior knowledge of a public official, it is true that the police detained OrnelasChavez for six hours when he was 16 years old. But this action was in no way connected to an incident of torture.

. For this reason, the majority is clearly incorrect to suggest that the IJ’s finding was unsupported by substantial evidence. Ornelas-Chavez told his teacher that he “had had *1069 sex with a male." But he further explained in a written declaration that he did not tell his teacher exactly what happened — the alleged rape — because his cousins had threatened him. At the hearing, Ornelas-Chavez was asked repeatedly whether he ever told anyone about the alleged rapes and other instances of abuse, and each time he answered "no” without qualification.

. Record evidence included an estimate that between eight and twelve percent of the Mexican population of 105 million persons is homosexual in orientation. See U.S. Department of State Homepage, Background Note: Mexico, at http:// www.state.gOv/r/pa/ei/bgn/35749.htm (last visited June 28, 2006) (providing statistics as to total population).

The record, of course, also includes an observation that the number of homosexuality-related homicides (1) mirrors the homicide rate for the Mexican population generally, and (2) has been in sharp decline since the early 1990s.