BERZON, Circuit Judge:

The primary question we decide today concerns whether one's membership in a "disfavored group" — that is, a group of individuals in a certain country or part of a country, all of whom share a common, protected characteristic, many of whom are mistreated, and a substantial number of whom are persecuted — is pertinent in determining whether an applicant for withholding of removal is eligible for that form of relief. The question arises because we have recognized that membership in a disfavored group is relevant to whether an applicant has a well-founded fear of future persecution for purposes of an asylum claim, but have never determined the role of disfavored group analysis in the context of a claim for withholding of removal. We do so now.

Under the Immigration and Nationality Act ("INA"), eligibility for asylum is established by demonstrating "`a well-founded fear of persecution'" on account of "`race, religion, nationality, membership in a particular social group, or political opinion.'" Al-Harbi v. INS, 242 F.3d 882, 888 (9th Cir. 2001) (quoting 8 U.S.C. § 1101(a)(42)(A)). To be "well-founded," the applicant's fear of persecution must be "both `subjectively genuine' and `objectively reasonable.'" Id. (internal citation omitted); see also 8 C.F.R. § 208.13(b)(2)(i). Because asylum is a discretionary form of relief, the standard for objective reasonableness is fairly low: Even a ten percent chance of future persecution may establish a well-founded fear. INS v. Cardoza-Fonseca, 480 U.S. 421, 440, 107 S.Ct. 1207, 94 L.Ed.2d 434 (1987). Eligibility for asylum is also conditioned upon a range of other factors, including — most relevantly for the petitioner in this case — the statutory requirement that the application for asylum be filed within one year of the applicant's arrival in the United States. 8 U.S.C. § 1158(a)(2)(B).

The INA also provides two additional, non-discretionary forms of relief to aliens who risk certain kinds of harm if removed to their home countries. First, withholding of removal under 8 U.S.C. § 1231(b)(3) provides relief to applicants who fear persecution according to the same substantive criteria as asylum, but with a higher standard of objective reasonableness; future persecution must be "more likely than not." 8 C.F.R. § 208.16(b)(2). Second, withholding or deferral of removal is available under the Convention Against Torture (commonly called "CAT relief). Applications for CAT relief must meet the same standard of objective reasonableness as do applications for normal withholding under 8 U.S.C. § 1231(b)(3), but substantive criteria differ. To be eligible for CAT relief, aliens must show that it is "more likely than not" that they will be tortured (rather than persecuted on a protected ground) if returned to their home countries. 8 C.F.R. § 208.16(c)(2).

The petitioner in this case, Matthew Henoch Wakkary, sought all three types of relief, and the agency found him ineligible for each one. We conclude, first, that the determination that Wakkary's asylum claim is time-barred must be reconsidered as the Board of Immigration Appeals ("BIA") applied the wrong legal standard in determining whether Wakkary filed his asylum application within a "reasonable period" after the expiration of his immigration status. See 8 C.F.R. § 208.4(a)(5). Second, we hold that the BIA should have considered Wakkary's country-conditions evidence regarding the widespread mistreatment of Chinese and Christians in Indonesia when assessing the likelihood that he would face future persecution for withholding of removal purposes, and so remand for reconsideration of the withholding decision. Finally, we hold that the BIA's determination that Wakkary did not demonstrate eligibility for CAT relief is supported by substantial evidence.

I. BACKGROUND

A. Factual background

Wakkary was born in 1974 in Medam, Indonesia. His mother is ethnically Chinese, and his father is ethnically Indonesian. His father has served as a Pentecostal Christian pastor in Medam for over twenty years. Wakkary, like his father, is a Pentecostal Christian pastor who has been active in the church.

As Wakkary's country-conditions evidence shows, and as we have recognized in other, similar cases, Indonesia's ethnic Chinese minority has suffered a long history of abuse and mistreatment at the hands of the native Indonesian majority. See Sael v. Ashcroft, 386 F.3d 922, 923 (9th Cir. 2004). Comprising approximately three percent of the country's population, Chinese Indonesians are considered relatively more affluent than the native Indonesian population, and are widely resented because of their presumed affluence. Former president Suharto accused the Chinese minority of backing an attempted coup in 1965. During his reign from 1967 to 1998, Suharto used the alleged disloyalty of the Chinese minority as a pretext to implement a range of discriminatory policies imposing various legal disadvantages on individuals of Chinese ethnicity. Although many of these official restrictions have been recently relaxed, popular resentment, discrimination, and, at times, violence against members of the Chinese minority have persisted. The continued occurrence of attacks on Christian churches reflects a religious dimension to the ethnic tensions. Many of Indonesia's Chinese minority are Christian, while the majority of native Indonesians are Muslim.

As a person of mixed Chinese ethnicity and a practicing Christian, Wakkary has personally experienced manifestations, some relatively minor and others somewhat more serious, of this widespread anti-Chinese and anti-Christian sentiment.

In 1985, at age eleven, Wakkary encountered a group of ten native Indonesian youths in the street who told him, "Chinese[,] you stop." The youths stole Wakkary's sandals and beat him. Wakkary told his parents about the incident, but they did not report it to the police because they believed that the police "couldn't do anything."

In 1990, at age sixteen, Wakkary and two friends, one ethnic Chinese and one ethnic Indian, encountered a group of native Indonesian youths. The youths forced Wakkary to hand over his watch and money at knifepoint, and then beat Wakkary and one of his friends with a stick. Again, Wakkary did not report this incident to the police because "[t]his happened many times in our city. [Native Indonesians] continued to do []thing[s] to the Chinese. [The police] could not defend me."

Ethnic tension in Indonesia became especially acute in mid-May of 1998. Widespread anti-Chinese rioting broke out, with native Indonesian mobs in Jakarta and elsewhere in the country attacking Chinese individuals and destroying their homes and businesses. According to the newspaper accounts Wakkary submitted to the agency, over a thousand Chinese individuals were killed during this period. Christian churches also became the targets of vandalism, looting, and arson. Many of the crimes perpetrated against the Chinese and Christian minorities went unpunished. Indeed, as the U.S. State Department's 2000 Country Report indicates, an official factfinding team found evidence that the Indonesian security forces not only tolerated the attacks, but may have been involved in planning and executing them.

During this period, Wakkary was traveling by car one day with his father, his mother, his aunt, and his brother and his brother's wife to a Bible-school gathering in another city. At the city limits, between twenty and thirty ethnic Indonesians with weapons blocked the car and asked "where[they] were going and what people [they were]." Wakkary's father (an ethnic Indonesian) said they were from the city of Manado, and the native Indonesians allowed them to pass after paying a small donation to build a mosque. Further on, as the car approached the next city, another group of ethnic Indonesians approached the car, "yelling loudly, ['A]re there any Chinese or Christians in that car[?']" The mob attempted "to get through the glass window into the car." Wakkary's father, an ethnic Indonesian, told the mob, "[`W]e are Indonesian indigenous[,]'" and the car was allowed to pass through.

When Wakkary and his family arrived at the Bible school, "there was crying" and "a lot of . . . people shouting." Wakkary noticed groups of native Indonesians looting Chinese-run shops nearby. He and others feared that the Bible school would also be attacked, so they stopped choir practice to "avoid any provocation." He and some other leaders of the group "asked [the people] to stay in the room" while they "guard[ed] the complex." Wakkary was not harmed during the incident.

Soon after the incident at the Bible school, Wakkary applied for, and was granted, an R-1 visa authorizing him to enter the United States as a temporary religious worker. His departure to the United States was delayed by several days due to unrest in Jakarta and large crowds of Chinese Indonesians at the airport attempting to board flights out of the country. Wakkary took refuge in a complex near the airport with other Chinese Indonesians, hiding from the armed rioters. Several days later, he ventured back to the airport and boarded a flight to the United States. He was admitted to the United States on May 16, 1998.

In September 2000, Wakkary's father told him by phone that Wakkary's friend Kalep, also a Christian, was shot and killed by two native Indonesians while driving Pastor Munthe, a Christian pastor, to church. Wakkary confirmed his father's account of Kalep's murder through the Internet. Wakkary also learned that several weeks before Kalep's murder, a bomb had exploded on the lawn outside Pastor Munthe's church, and that there had been an attempted bombing of Pastor Munthe's car. Wakkary believes that these acts were perpetrated by native Indonesian extremists who targeted Kalep and Pastor Munthe on account of their religion. Wakkary's country-conditions evidence confirms that hundreds of churches throughout Indonesia were attacked and bombed in 1999, 2000, and 2001.

On March 31, 2001, Wakkary was readmitted to the United States on a new R-1 visa, and he has remained in this country ever since. His R-1 visa authorized him to remain here only until April 11, 2002. He ultimately filed his asylum application on October 15, 2002. After an interview with the asylum office, his application was denied, and on November 25, 2002, the then-Immigration and Naturalization Service initiated removal proceedings by issuing a Notice to Appear.

B. Proceedings before the IJ

Wakkary appeared before an Immigration Judge ("IJ") for his removal hearing on July 17, 2003. The IJ determined, first, that Wakkary's asylum claim was not timely filed, so he was ineligible for that form of relief. Moving on to Wakkary's claims for withholding of removal under 8 U.S.C. § 1231(b)(3) and CAT relief — for which there are no filing deadlines — the IJ rejected Wakkary's claims on the merits. He denied withholding of removal because he found that the thefts and beatings that Wakkary experienced in 1985 and 1990 were "[a]pparently . . . random encounters" that were "remote in time and circumstances to [his] departure" from Indonesia, while the more recent events that occurred in the context of the 1998 riots were "isolated circumstance[s]" that were "not directed against [Wakkary] personally." The IJ also went on to deny CAT relief because he found that Wakkary had failed to show that it is more likely than not that he would be subjected to torture upon return to his native country.

C. Appeal to the BIA

On appeal, the BIA "adopt[ed] and affirm[ed]" the IJ's decision that Wakkary had not timely filed for asylum and that he had not met his burden of proof for withholding of removal or CAT relief. The BIA also added its own commentary regarding this Court's "disfavored group" holding in Sael v. Ashcroft, 386 F.3d 922 (9th Cir. 2004), which was decided after Wakkary's hearing before the IJ. The BIA characterized Sael as prescribing a "reduced burden of proof for. asylum applicants who belong to disfavored groups. Based on that understanding of Sael, the BIA considered Sael inapplicable in the context of a claim for withholding of removal, which requires that the applicant meet a higher burden of proof than is required of an asylum applicant. The BIA therefore refused to apply Sael's disfavored group analysis to Wakkary's withholding and CAT claims.

Wakkary timely filed a petition for review with this Court, arguing that the BIA erred (1) in holding that his asylum claim was time-barred; (2) in rejecting his claim for withholding of removal, either (a) because the record compels a finding that he suffered past persecution or, in the alternative, (b) because he faces a clear probability of future persecution in light of the widespread mistreatment of Chinese and Christians and his own past experiences; and (3) in rejecting his CAT claim. We have jurisdiction over his petition for review under 8 U.S.C. § 1252.

II. ANALYSIS

In cases in which, as here, the BIA "adopt[s] and affirm[s]" the decision of the IJ but also adds its own analysis, the scope of our review extends to the decisions of both the IJ and the BIA. Kataria v. INS, 232 F.3d 1107, 1112 (9th Cir. 2000). We review the agency's legal determinations de novo, and factual findings for substantial evidence. Hernandez-Gil v. Gonzales, 476 F.3d 803, 804 n. 1 (9th Cir. 2007). Because the IJ found Wakkary credible, we treat Wakkary's testimony as true. See Vukmiromc v. Ashcroft, 362 F.3d 1247, 1251 (9th Cir. 2004).

A. Wakkary's Asylum Claim

The INA generally requires that an alien file for asylum within one year of arriving in the United States. 8 U.S.C. § 1158(a)(2)(B). The INA allows for exceptions, however, for aliens who show "extraordinary circumstances relating to the delay in filing." Id. § 1158(a)(2)(D). According to the applicable regulations, one example of such an "extraordinary circumstance[]" is when an "applicant maintained . . . lawful immigrant or nonimmigrant status . . . until a reasonable period before the filing of the asylum application." 8 C.F.R. § 208.4(a)(5)(iv).

Wakkary was residing in this country on a valid religious worker visa until April 11, 2002. Under the applicable regulation, his maintenance of such status qualifies as an "extraordinary circumstance[.]" See 8 C.F.R. § 208.4(a)(5). Nevertheless, noting that Wakkary submitted his asylum application on October 15, 2002 — six months and some days after his status expired — the IJ held his delay inexcusable. The IJ stated:

The claim that respondent was delaying the actual filing of the application[] while he attempted to obtain supporting materials is simply not persuasive, or otherwise capable of excusing the [one-year] filing deadline. Even so, respondent should have filed the application immediately [upon the expiration of his R-1 visa] and then requested time to gather supporting materials while the application was under consideration by the Immigration officers.

The IJ therefore held that Wakkary's asylum application was untimely. The BIA affirmed without providing further elaboration.

Neither the IJ nor the BIA cited the relevant statutory or regulatory provisions, or even used their terminology, in reaching the conclusion that Wakkary's application was time-barred. The IJ stated that Wakkary "should have filed the application immediately" when his visa expired. Thus, the IJ effectively held Wakkary to the usual one-year filing deadline, ignoring the fact that he had maintained lawful nonimmigrant status until April 11, 2002, and stating further that Wakkary's "attempt[s] to obtain supporting materials" were by definition not "capable" of justifying filing any later than that. For the reasons we explain below, we hold that the IJ's determination is based on an incorrect understanding of the applicable regulation.

The regulations provide that, to be excused from the one-year filing deadline, an applicant must first demonstrate extraordinary circumstances, and then show "that those circumstances were directly related to the alien's failure to file the application within the one-year period, and that the delay was reasonable under the circumstances." 8 C.F.R. § 208.4(a)(5). As noted above, Wakkary's maintenance of lawful status qualifies as an extraordinary circumstance under the regulation. The question then becomes whether his delay in filing was "reasonable under the circumstances" within the meaning of 8 C.F.R. § 208.4(a)(5), a determination that logically depends upon both (1) the reasons given to justify the delay, and (2) the length of time that passed between the expiration of his status and the filing of his asylum application.

We recently confronted a similar question — whether an alien filed his asylum application within a reasonable period of the expiration of his status — in Husyev v. Mukasey, 528 F.3d 1172 (9th Cir. 2008). There, the petitioner waited exactly 364 days after the expiration of his visa to file an asylum application, and he gave no explanation for his delay. Id. at 1182. We concluded that Husyev's asylum application was not filed within a reasonable period. Our conclusion was guided in part by the Preamble to the permanent regulations, which describes the Department of Justice's "expectat[ion]" that "waiting six months or longer after expiration or termination of status would not be considered reasonable." Id. at 1181 (quoting 65 Fed. Reg. 76, 121-01, 76, 123-24 (Dec. 6, 2000)). In light of the Preamble's emphasis on taking a flexible, case-based approach to the reasonable circumstances determination, see 65 Fed.Reg. at 76123, Husyev read the Preamble's six-month period not as a hard cut-off, but as a "presumptive deadline" that would apply "[i]n the absence of any special considerations." Husyev, 528 F.3d at 1182. We emphasized that each case must be assessed "on the basis of all the factual circumstances." Id. at 1182 n. 4.

Wakkary's filing fell just days outside the presumptive six-month deadline. Unlike the petitioner in Husyev, he presented a cogent and well-documented reason for his delay: taking time to gather identity documents and supporting documents he considered vital to his claim. Contrary to the IJ's blanket statement, nothing in the statute or regulations suggests that gathering supporting documents is per se an invalid reason for filing outside the one-year deadline where extraordinary circumstances have been established. On the contrary, the regulations make clear that the reasonableness determination must be made "under the circumstances," on a case-by-case basis. 8 C.F.R. § 208.4(a)(5); see also 65 Fed.Reg. at 76,123. We therefore reject, as unsupported by the statute or regulations, the IJ's holding that gathering supporting documents can never be a valid reason for delay.

The question that remains is whether Wakkary's particular circumstances render his delay of just over six months "reasonable" under the regulations. In Husyev, we held that while "six months may serve in default as a reasonable presumptive deadline, we do not foreclose other reasonable periods, and exceptions thereto, that may be set out by the agency, nor do we preclude individualized determinations of reasonableness of delay." Id. at 1182 n. 4. Here, of course, the agency did not perform such ah individualized determination. Because the IJ considered the gathering of supporting documents to be necessarily "[in]capable" of justifying a reasonable delay, he had no occasion to engage in the fact-based inquiry of the particular circumstances for which the regulations call. We therefore remand so that the agency may determine in the first instance whether, under the particular circumstances of this case, Wakkary applied for asylum within a "reasonable period" as the regulations require. See Fakhry v. Mukasey, 524 F.3d 1057, 1064 (9th Cir. 2008) ("Because the agency applied the wrong legal standard and has not considered the issue using the correct standard, we remand for a determination of whether Fakhry demonstrated changed circumstances and, if so, whether he applied for asylum within the requisite `reasonable period.'").

B. Wakkary's Withholding of Removal Claim

As noted above, asylum is a discretionary form of relief. See Cardoza-Fonseca, 480 U.S. at 424, 107 S.Ct. 1207. Therefore, even if the agency on remand determines that Wakkary's asylum application was filed within a reasonable time and goes on to conclude on the merits that he is eligible for relief, the Attorney General retains the discretion not to grant that relief. We therefore consider Wakkary's remaining arguments regarding his claim for withholding of removal.

Wakkary submits that the BIA erred in two respects with regard to his claim for withholding of removal: first, in concluding that he had not suffered past persecution, and second, in concluding that he failed to show that future persecution was more likely than not.

1. Past persecution

Persecution is "an extreme concept that does not include every sort of treatment our society regards as offensive." Nagoulko v. INS, 333 F.3d 1012, 1016 (9th Cir. 2003) (internal quotation marks omitted). "[M]ere discrimination," by itself, is not the same as persecution. Fisher v. INS, 79 F.3d 955, 962 (9th Cir. 1996) (en banc). Severe and sustained discrimination, or discrimination in combination with other harms, however, may compel a finding of past persecution. See, e.g., Krotova v. Gonzales, 416 F.3d 1080, 1087 (9th Cir. 2005) (granting petition for review where petitioner testified that she and her daughter were attacked in the street; her brother was beaten; a close family friend was murdered by anti-Semites; and she was sexually harassed, denied promotions and state-sponsored child-care, and unable to practice her religion because she was Jewish); Duarte de Guinac v. INS, 179 F.3d 1156, 1161-62 (9th Cir. 1999) (granting petition for review where petitioner testified that during his conscripted service in the army, he was repeatedly beaten and subjected to race-based insults); Korablina v. INS, 158 F.3d 1038, 1044-45 (9th Cir. 1998) (granting petition for review where petitioner testified that her Jewish coworker disappeared and her Jewish boss was fired; that she personally received threats and anti-Semitic slurs; and that she witnessed violent attacks on other Jews and experienced one herself).

Wakkary's personal experiences at the hands of native Indonesians — being beaten by youths and robbed of his sandals and pocket money in 1985 and 1990 (seventeen and twelve years, respectively, before he filed his asylum application), and being accosted by a threatening mob while his family was driving to Bible school in 1998 — are instances of discriminatory mistreatment. We cannot say, however, that a reasonable factfinder would be "compel[led]" to conclude that these experiences, without more, cumulatively amount to past persecution. INS v. Elias-Zacarias, 502 U.S. 478, 481 n. 1, 112 S.Ct. 812, 117 L.Ed.2d 38 (1992). Compare Nagoulko, 333 F.3d at 1016-18 (evidence that petitioner was insulted and harassed because of her religion, was fired from her job but able to find other work, and witnessed the beatings of her coworkers, did not compel the conclusion that she suffered past persecution). Further, although harm to a petitioner's close relatives, friends, or associates may contribute to a successful showing of past persecution, see Korablina, 158 F.3d at 1044-45, Wakkary has not demonstrated that the murder of Kalep and the attempted murder of Pastor Munthe were part of "a pattern of persecution closely tied to" Wakkary himself, as we have required. Arriaga-Barrientos v. INS, 937 F.2d 411, 414 (9th Cir. 1991). We therefore conclude that the BIA's determination that Wakkary did not prove past persecution is supported by substantial evidence.

2. Future persecution

Even though Wakkary did not demonstrate past persecution, his likelihood of future persecution may still be sufficient to merit withholding of removal. See Avetova-Elisseva v. INS, 213 F.3d 1192, 1201-02 (9th Cir. 2000). To establish eligibility for withholding of removal in the absence of past persecution, an applicant must demonstrate both that he has a subjective fear of persecution in the future, and that this fear is objectively reasonable — which, in the withholding context, means that the chance of future persecution is "more likely than not." 8 C.F.R. § 208.16(b)(2).

In the asylum context, the INA's implementing regulations map out two routes by which an asylum-seeker can show that the objective risk of future persecution is high enough to merit relief. First, he may show that there is a "reasonable possibility" that he will be "singled out individually for persecution" if removed. 8 C.F.R. § 1208.13(b)(2)(iii). Second, in the alternative, he may show that there is a systematic "pattern or practice" of persecution against the group to which he belongs in his home country, such that, even without any evidence of individual targeting, his fear of persecution is deemed reasonable. Id. The regulations governing withholding of removal have the same bifurcated structure, providing for both pattern or practice claims and individualized targeting claims. 8 C.F.R. § 208.16(b)(2)(i)-(ii). Wakkary submits that he has demonstrated the requisite risk of persecution via both routes. We address the two routes in reverse order.

a. Pattern or practice of persecution

To establish the objective reasonableness of one's fear of future persecution via the "pattern or practice" route, an applicant for withholding of removal must demonstrate, first, that "in [his home] country there is a pattern or practice of persecution of a group of persons similarly situated to the applicant on account of race, religion, nationality, membership in a particular social group, or political opinion," and second, that by reason of his "inclusion in and identification with such group of persons such that it is more likely than not that his . . . life or freedom would be threatened upon return to that country." 8 C.F.R. § 208.16(b)(2)(i)-(ii). In Wakkary's case, the IJ never reached the second question, because he determined that Wakkary failed at the first step — that is, he did not demonstrate that a pattern or practice of persecuting Chinese Christians existed in Indonesia. The record before us does not compel us to overturn the IJ's ruling on this point, for the following reasons.

We have never specified with any precision how widespread the pattern or practice of persecuting a protected group must be to meet the statutory definition. See Bromfield v. Mukasey, 543 F.3d 1071, 1078 (9th Cir. 2008) (finding a pattern or practice of persecuting gay men in Jamaica); Knezevic v. Ashcroft, 367 F.3d 1206 (9th Cir. 2004) (finding that Croats maintained a pattern or practice of ethnically cleansing Serbs in the petitioners' region of Bosnia-Herzegovina); Mgoian v. INS, 184 F.3d 1029, 1036 (9th Cir. 1999) (finding a pattern or practice of persecution against a family in the Kurdish-Muslim intelligentsia in Armenia). We have used the word "systematic" to describe the sort of persecution that might amount to a pattern or practice, see Kotasz v. INS, 31 F.3d 847, 853 (9th Cir. 1994), but we have never reduced our understanding of the term to a concrete percentage, and we decline to do so here. Further, although we have acknowledged that one of the archetypal examples of a pattern or practice of persecution was "the systematic attempt to annihilate the Jews in Nazi Germany," Kotasz, 31 F.3d at 852 — a case in which the persecution was incremental, but a broad campaign of marginalization, and eventually extermination, was organized by the Nazi government itself — we have recognized that a pattern or practice may also be the work of private actors, so long as the persecution is sufficiently widespread and the government is unable or unwilling to control those actors. See Lolong v. Gonzales, 484 F.3d 1173, 1180 (9th Cir. 2007) (en banc).

Under these general principles, the record in Wakkary's case does not compel the conclusion that there exists a pattern or practice of persecution against Chinese and Christians in Indonesia. Although the record contains evidence of widespread anti-Chinese and anti-Christian discrimination that affects a very large number of individuals, and although it is clear that a certain portion of those individuals suffer treatment that rises to the level of persecution, the record does not establish that the situation in Indonesia is similar to the patterns or practices of persecution described in our prior case law.

In Bromfield, for example, the petitioner's country-conditions evidence showed that "violence against homosexuals [in Jamaica] is widespread, and is perpetrated by both private individuals and public officials such as police officers and prison personnel," and that "Jamaican law criminalizes homosexual conduct, making it punishable by up to ten-years imprisonment. . . ." Bromfield, 543 F.3d at 1079. And in Knezevic, the record revealed that the petitioners' home town was "targeted for bombing, invasion, occupation, and ethnic cleansing of Serbs by Croats," and specifically, that "a bus of Serbs attempting to return . . . was stoned by Croats; violent mobs of Croats blocked Serb, reentry to the town and destroyed the Serbs' homes; police refused to guarantee protection to Serbs seeking to return; and harassment, looting of vacant homes, cattle theft, and physical violence against Serbs by Croats were widespread." Knezevic, 367 F.3d at 1212.

Applying substantial evidence review, we can only conclude that the record before us does not compel the conclusion that the agency's determination was wrong. We hold that substantial evidence supports the agency's denial of Wakkary's pattern or practice claim.

b. Individualized targeting and disfavored group analysis

In the alternative, Wakkary submits that the agency erred in its assessment of whether he would be individually targeted for persecution by refusing to consider evidence that Wakkary belongs to a "disfavored group." The BIA reasoned that our holding in Sael — regarding the evidentiary relevance of an asylum applicant's membership in a disfavored group — is simply inapplicable to claims for withholding of removal. As a result, the agency ignored a large amount of record evidence that is pertinent to whether Chinese Christians in Indonesia are widely disfavored, discriminated against, and, in a substantial number of instances, persecuted on account of their race and religion.

We disagree with the agency's refusal to consider this evidence in evaluating the future persecution issue. At the same time, we do recognize that the disfavored group mode of analysis needs clarification, as it has been misunderstood by both the agency and some other circuits. Moreover, we note that in practice, the impact of the disfavored group mode of analysis is likely to be of considerably less significance in withholding than in asylum cases, due to the different standards of proof for these two forms of relief. Still, even though this evidence will not often change the outcome of a withholding claim, it is nevertheless relevant, and the BIA erred in not considering it.

i. Disfavored group analysis in asylum cases

We begin by briefly reviewing the development, in the asylum context, of what has come to be called — perhaps unfortunately, as the terminology may be misleading — "disfavored group" analysis. We first recognized the evidentiary relevance of an asylum applicant's membership in a "disfavored group" in Kotasz v. INS, 31 F.3d 847 (9th Cir. 1994). Kotasz explained that although the asylum regulations provide two ways to establish a well-founded fear of future persecution — showing a "pattern or practice" of persecution, or showing a likelihood of being "individually singled out," see 8 C.F.R. § 1208.13(b)(2)(iii) — these two categories of future-fear claims should not be understood to require discrete sorts of evidence.

As Kotasz emphasized, "[g]roup membership is an aspect of nearly all asylum claims, not a special problem limited to pattern or practice cases." Id. at 853. Indeed, Kotasz noted, the most "common" individualized persecution scenario is one in which, "although members of the disfavored group[] are not threatened by [a pattern or practice of] systematic persecution of the group's entire membership, the fact of group membership nonetheless places [the individual] at some risk." Id. The "singled out" path is not reserved solely for those applicants whose would-be persecutors seek them out personally, by name. Rather, Kotasz recognized that one's chances of being singled out from the general population and subjected to persecution is often strongly correlated with the frequency with which others who share the same disfavored characteristics are mistreated and persecuted.

So, in a case in which the asylum applicant attempts to show that he faces a reasonable likelihood of being singled out individually on account of a protected characteristic, "[p]roof that the government or other persecutor has discriminated against a group to which [he] belongs is . . . always relevant," because that proof says something about the chances that he, as a member of that group, will be persecuted. Id. Based on this common-sense evidentiary proposition, Kotasz held that once an applicant establishes that he is a member of a group that is broadly disfavored, "the more egregious the showing of group persecution — the greater the risk to all members of the group — the less evidence of individualized- persecution must be adduced" to meet the objective prong of a well-founded fear showing. Id. (emphases added); see also Mgoian, 184 F.3d at 1035 n. 4.

Since Kotasz, we have found petitioners from a number of groups to have established that they are members of disfavored groups for asylum purposes. See, e.g., El Himri v. Ashcroft, 378 F.3d 932, 937 (9th Cir. 2004) (stateless Palestinians born in Kuwait); Hoxha v. Ashcroft, 319 F.3d 1179, 1182-83 (9th Cir. 2003) (ethnic Albanians in Kosovo); Singh v. INS, 94 F.3d 1353, 1359-60 (9th Cir. 1996) (ethnic Indians in Fiji). We first applied disfavored group analysis to the asylum claim of a member of Indonesia's Chinese Christian minority in Sael v. Ashcroft, 386 F.3d 922 (9th Cir. 2004).

In Sael, we found that the record evidence, documenting centuries of popular and official discrimination against the Chinese and Christian minorities in Indonesia, "establish[ed] that ethnic Chinese are significantly disfavored in Indonesia." 386 F.3d at 927. Accordingly, Sael, as a Chinese Christian woman, had to "demonstrate a `comparatively low' level of individualized risk in order to prove that she ha[d] a well-founded fear of future persecution." Id. We concluded that Sael met her burden of showing individualized risk by adducing evidence of her past personal experiences of threats, vandalism, and threatened violence by native Indonesians. Id. at 927-29. These experiences, "even [though] not sufficient to compel a finding of past persecution," were "indicative of individualized risk [of future harm]." Id. at 928-29. When viewed in the context of the country-conditions evidence of widespread discrimination against Chinese Christians, Sael's evidence of individualized risk compelled the conclusion that her own fear of future persecution was objectively well-founded. Id. at 929.

ii. The application of disfavored group analysis to withholding claims

Until today, we have applied the disfavored group approach only to asylum claims. In Sael, we expressly reserved the question whether this mode of analysis is applicable in the context of withholding of removal claims as well. Sael, 386 F.3d at 930 n. 10. Wakkary's case presents that question squarely. When it dismissed Wakkary's appeal, the BIA held that Sael prescribes a "reduced burden of proof for asylum seekers belonging to disfavored groups, which it considered to be incompatible with "the higher, and less elastic, clear probability standard required for withholding of removal." In so ruling, the BIA misunderstood Sael's holding.

Disfavored group analysis does not prescribe a lower-than-usual burden of proof for the asylum claims of members of certain judicially-anointed groups. An individual seeking asylum must always show that he faces at least a ten percent chance of future persecution, whether he attempts to meet his burden by showing a pattern or practice or by showing a likelihood that he will be individually singled out. See Al-Harbi 242 F.3d at 888. We have never held, in Sael, Kotasz, or elsewhere, that a member of a disfavored group must show only, say, a five or nine percent likelihood of persecution to be eligible for asylum.

Instead, the "lesser" or "comparatively low" burden to which we averted in Kotasz, 31 F.3d at 854, and Sael, 386 F.3d at 927, refers not to a lower ultimate standard, but to the lower proportion of specifically individualized evidence of risk, counterbalanced by a greater showing of group targeting, that an applicant must adduce to meet that ultimate standard under the regulations' "individually singled out" rubric. The more evidence of group targeting an asylum applicant proffers, the less evidence of individually specific evidence he needs to reach that ten percent. In other words, the disfavored group concept simply describes the basic evidentiary proposition that an asylum applicant's membership in a group whose members are shown to have been widely targeted for discrimination, a substantial number of whom are shown to have been persecuted, is relevant evidence in assessing whether his fear of being personally targeted for persecution in the future rises to the requisite level of objective reasonableness.

With this clarification, it should be apparent that our disfavored group cases do not invent a "judicially created alternative to the statutory and regulatory scheme," Kho v. Keisler, 505 F.3d 50, 55 (1st Cir. 2007), or a "lower threshold of proof," Firmansjah v. Gonzales, 424 F.3d 598, 607 n. 6 (7th Cir. 2005), as some other circuits have mistakenly assumed. See also Lie, 396 F.3d at 538 n. 4. Rather, our cases merely outline a method for weighing indicators of future individual risk, a method consistent with that employed in other circuits. The First Circuit, for example, recognized "that in evaluating each claim on its facts, it may be that evidence short of a pattern or practice will enhance an individualized showing of likelihood of a future threat to an applicant's life or freedom." Kho, 505 F.3d at 55. Although the First Circuit regarded that possibility as "a different matter" from our disfavored group approach, in fact that fairly obvious evidentiary point is the essence of our disfavored group cases. In other words, when asking how likely it is that an individual applicant will be "singled out" in the future on the basis of his group membership, it is indisputably relevant (though of course not dispositive) how others in his group are treated. For, as the Fourth Circuit explained,

an applicant will typically demonstrate some combination of [individual risk and group risk] to establish a well-founded fear of persecution. "The more egregious the showing of group persecution[,] the less evidence of individualized persecution must be adduced." Conversely, a stronger showing of individual targeting will be necessary where the underlying basis for the applicant's fear is membership in a diffuse class against whom actual persecution is hap-hazard and rare.

Chen v. INS, 195 F.3d 198, 203-04 (4th Cir. 1999) (internal alterations omitted) (quoting Kotasz, 31 F.3d at 853); see also Makonnen v. INS, 44 F.3d 1378, 1383 (8th Cir. 1995).

We see no reason why this evidentiary proposition should be applicable to asylum claims but not to withholding claims. The BIA correctly noted that asylum and withholding of removal have different quantitative standards of proof — ten percent for asylum, and "more likely than not" for withholding. See 8 C.F.R. § 1208.16(b)(1)(iii); Cardoza-Fonseca, 480 U.S. at 449, 107 S.Ct. 1207; Lota v. INS, 204 F.3d 1241, 1244 (9th Cir. 2000).

To repeat, disfavored group analysis does not alter the quantitative standard of proof. Rather, it determines what sorts of evidence can be used to meet that standard, and, quite generally, in what proportions. What matters is that both asylum and withholding have the same qualitative criteria for eligibility, so the same sorts of evidence are relevant to determining whether an applicant's fear is objectively reasonable. Under the regulations, eligibility for both forms of relief depend upon the likelihood that the applicant will be persecuted in his home country "on account of race, religion, nationality, membership in a particular social group, or political opinion." 8 C.F.R. § 1208.13(b)(2)(i)(A) (asylum); id. § 1208.16(b) (withholding). And 8 C.F.R. § 1208.16(b)(2), governing withholding, provides in substantially identical language to 8 C.F.R. § 1208.13(b)(2)(iii) (governing asylum) that a withholding applicant may demonstrate that his fear of future persecution rises to the requisite level of objective reasonableness either by showing a "pattern or practice of persecution," or by showing that he will be "singled out individually." Id. § 1208.16(b)(2).

In sum, if an applicant's membership in a disfavored group is relevant to the objective reasonableness of his future fear in the asylum context, we see no reason, either in logic or in the statutory and regulatory design, why membership in a disfavored group should not similarly be relevant to assessing the likelihood of individual targeting in the withholding context. We hold that it is.

iii. The need to adduce individualized evidence

Even under the disfavored group approach, an applicant for withholding of removal must show that his chance of future persecution is greater than fifty percent. Evidence of group discrimination will go part of the way toward meeting that bar — how far depending upon how "egregious" and pervasive the showing of group discrimination is, see Mgoian, 184 F.3d at 1035 n. 4 — but, absent a "pattern or practice" of persecution, it can never go all the way. As we explained recently in the asylum context in Lolong v. Gonzales, 484 F.3d 1173, 1179, 1181 n. 6 (9th Cir. 2007) (en banc), some evidence of individualized risk is necessary for the petitioner to succeed.

In Lolong, we upheld the BIA's determination that an asylum applicant belonging to Indonesia's Chinese Christian minority failed to show an objectively well-founded fear of future persecution, because she had offered no evidence of her own individual risk. Id. at 1181 n. 6. Rather, Lolong rested her claim solely on evidence that Chinese Indonesian women as a group are targeted for violent attacks and rapes, and on evidence that family members and a friend had experienced anti-Chinese violence. We held that a "general, undifferentiated claim" based solely on the threat to the group as a whole is not sufficient for an individual petitioner to establish the requisite likelihood of persecution under the "singled out individually" rubric. Id. at 1179. Asylum applicants who attempt to show their eligibility for asylum in part by relying on their membership in a disfavored group must "prove something more than their status as . . . members of that group. Id. at 1181 n. 6. Similarly, in Sael, we emphasized that the petitioner's showing that she is a member of a "disfavored group" must be "coupled with a showing that she, in particular, is likely to be targeted as a member of that group." Sael, 386 F.3d at 925. Sael met this burden by adducing evidence that she personally had been threatened, her home vandalized, and her car thronged by a mob who saw that she was Chinese. Id. at 927-28.

An applicant for withholding of removal will need to adduce a considerably larger quantum of individualized-risk evidence to prevail than would an asylum applicant like Sael, assuming their disfavored group evidence is of equal severity and pervasiveness, because the ultimate bar for withholding is higher than the bar for asylum. As a practical matter, then, applying disfavored group analysis to withholding claims will not affect the outcome of most petitioners' cases. Most aliens seek both forms of relief as a matter of course. If an applicant fails to demonstrate eligibility for asylum even with the help of disfavored group analysis (such that his chance of future persecution, even with the evidentiary boost that his membership in a disfavored group provides, is less than ten percent), then he will necessarily have failed to demonstrate eligibility for withholding of removal. See Mansour v. Ashcroft, 390 F.3d 667, 673 (9th Cir. 2004).

There are cases, however, in which the applicant's request for asylum, fails not for lack of proof on the merits, but because his application is found to be time-barred — as Wakkary's was initially, although we have remanded that decision. In such cases, the fact of the applicant's membership in a group proven to be disfavored could mean the difference between meeting withholding's "more likely than not" bar and coming up short. Given the high burden of proof for withholding, it is likely that evidence of group persecution not sufficiently widespread to amount to a "pattern or practice" will relatively infrequently succeed in filling the gap between individually-specific risk evidence and the requisite level of proof. But "infrequently" is not "never." We hold that the BIA erred in precluding consideration of disfavored group evidence entirely with regard to withholding of removal.

iv. Application to Wakkary's withholding claim

At his removal hearing, Wakkary presented some evidence of past mistreatment that he suffered personally. The IJ dismissed these incidents as "random encounters . . . not directed against [Wakkary] personally," which, being random, had little or nothing to say about Wakkary's chances of future persecution.

The events were not random in any relevant sense. For each of these incidents, Wakkary provided circumstantial evidence that his harassers were motivated by anti-Chinese and/or anti-Christian sentiment. See Sinha v. Holder, 556 F.3d 774, 779-80 (9th Cir. 2009) (rejecting an IJ's characterization of attacks on Indo-Fijian petitioners, in the context of widespread violence against their ethnic group, as "random," because "individuals of Indo-Fijian ethnicity were the specific targets of the widespread violence" and because the petitioners "presented specific evidence that [their] attackers had racially discriminatory motives," including testimony that their attackers berated them with racist epithets); Baballah v. Ashcroft, 367 F.3d 1067, 1077 (9th Cir. 2003) ("The use of [ethnic] slur[s during the attacks] amply establishes the connection between the acts of persecution and Baballah's ethnicity and religion."). Wakkary's attackers may not have known his name, but it is clear they targeted him, and not some other bystander, because he is Chinese and/or a Christian. The only way the IJ could have considered the incidents random is by ignoring this evidence of motive, as well as the broader evidence of widespread "ongoing hostility towards the Chinese and Christian community." Refusing to weigh Wakkary's individually-focused evidence in light of the risk of mistreatment and persecution that all Chinese Christians in Indonesia face, the IJ saw only random chance at work. In fact, Wakkary has shown that there is a strong correlation between the harm he personally faces and the plight of Chinese Christians in Indonesia generally.

Moreover, Wakkary testified to his position as a pastor and a "religious leader" within the Chinese Christian community — a position that makes him particularly visible and vulnerable to attack on account of his group membership. See Kotasz, 31 F.3d at 854 n. 10 ("[W]here a Christian sect is targeted, priests may be more at risk than ordinary church members. . . ."); Mgoian, 184 F.3d at 1035 n. 4 (noting that evidence of "the alien's individual risk level" may include evidence that "the alien has a special role in the group or is more likely to come to the attention of the persecutors[,] making him a more likely target for persecution.").

The question for the agency on remand will be whether Wakkary has adduced enough evidence of individual risk, in combination with enough evidence that the ethnic and religious group to which he belongs is disfavored in Indonesia, to make out a clear probability of persecution upon return. The BIA never engaged in this inquiry, because it held the disfavored group evidence — that is, the country-conditions evidence concerning the widespread mistreatment and persecution of Chinese and Christian individuals in Indonesia — to be entirely irrelevant to withholding of removal.

Given our limited role in reviewing orders of removal, we may not decide in the first instance whether Wakkary's evidence is sufficient to meet the standard for withholding. Rather, because the BIA misunderstood our disfavored group cases and so did not take into account the disfavored group evidence, we are obliged to remand to the BIA for an appropriate decision based on all the relevant evidence. See INS v. Ventura, 537 U.S. 12, 16-17, 123 S.Ct. 353, 154 L.Ed.2d 272 (2002).

C. Wakkary's Claim for Relief Under the Convention Against Torture

Finally, Wakkary argues that the BIA erred in denying his claim for relief under the CAT. He argues that the BIA erred in failing to apply disfavored group analysis in assessing his CAT claim, or in the alternative, that even without disfavored group analysis, the record compels the conclusion that he is eligible for CAT relief.

To demonstrate eligibility for withholding of removal under the CAT, an alien must show that it is "more likely than not" that a government official or person acting in an official capacity would torture him or aid or acquiesce in his torture by others. Kamaltkas v. INS, 251 F.3d 1279, 1283 (9th Cir. 2001); 8 C.F.R. §§ 208.16(c)(2), 208.18(a)(1). The standard of proof — more than fifty percent — is the same as for regular withholding of removal under 8 U.S.C. § 1231(b)(3).

Unlike regular withholding of removal, though, the regulations that implement the CAT do not specify two routes (a "pattern or practice" route and a "singled out individually" route) by which an applicant may establish that likelihood. 8 C.F.R. § 1208.16(c)(3). Rather, the CAT regulations cast a wide evidentiary net, providing that "all evidence relevant to the possibility of future torture shall be considered, including, but not limited to . . . evidence of gross, flagrant or mass violations of human rights within the country of removal, where applicable, and other relevant information regarding conditions in the country of removal." Id.

In a sense, then, Wakkary is arguing for a rule that is already effectively required by the regulations themselves — given our understanding that disfavored group analysis is simply an evidentiary approach to determine the risk that faces an alien if he is returned to the country from which he fled. Widespread mistreatment of a certain group of people may well be relevant to an applicant's claim that he faces a clear probability of torture upon return, depending upon the facts of the case. That this is so is not a product of Sael so much as it is necessitated by the CAT regulations themselves.

In Wakkary's case, however, the record contains no evidence whatsoever that Wakkary is likely to be tortured, rather than persecuted, by government officials or with their acquiescence on return to Indonesia. The CAT regulations define torture as

any act by which severe pain or suffering, whether physical or mental, is intentionally inflicted on a person for such purposes as obtaining from him or her or a third person information or a confession, punishing him or her for an act he or she or a third person has committed or is suspected of having committed, or intimidating or coercing him or her or a third person, or for any reason based on discrimination of any kind. . . .

8 C.F.R. § 208.18(a)(1). Wakkary has presented some evidence that torture occurs in Indonesia — for example, against certain detainees in military or police custody — but he has offered no evidence that he is likely to find himself in such a position, or that torture is used against members of the Chinese Christian minority. While Wakkary need not show that the torture he fears would be "on account of any of the five protected grounds of race, religion, nationality, political opinion, or particular social group, he must still provide some reason to think that he is likely to be tortured by the actors he fears. There being no such evidence in the record, we affirm the BIA's denial on this point.

III. CONCLUSION

In summary, we conclude that the agency applied an incorrect standard in determining whether Wakkary's asylum claim is time-barred, and so we remand for reconsideration on that claim. Regarding Wakkary's claim for withholding of removal, we conclude that the agency's determinations that Wakkary failed to demonstrate past persecution or a pattern or practice of persecution against Chinese Christians in Indonesia are supported by substantial evidence. We hold, however, that the agency erred in refusing to consider the evidence regarding whether Wakkary belonged to a disfavored group in assessing the likelihood that he would face future persecution for withholding of removal purposes, and so remand to the BIA for reconsideration of the withholding determination. Finally, we conclude that the BIA's denial of CAT relief is supported by substantial evidence.

The petition for review is GRANTED in part, and DENIED in part.