`
`UNITED STATES COURT OF APPEALS
`FOR THE NINTH CIRCUIT
`
`
`
`No. 23-95
`Agency No.
`A216-276-233
`
`OPINION
`
`
`JASWINDER SINGH,
`
` Petitioner,
`
` v.
`
`MERRICK B. GARLAND, Attorney
`General,
`
` Respondent.
`
`
`On Petition for Review of an Order of the
`Board of Immigration Appeals
`
`Argued and Submitted December 5, 2023
`Seattle, Washington
`
`Filed October 4, 2024
`
`
`Before: N. Randy Smith, Gabriel P. Sanchez, and Salvador
`Mendoza, Jr., Circuit Judges.
`
`Opinion by Judge Sanchez;
`Dissent by N. Randy Smith
`
`
`
`
`
`
`2
`
`SINGH V. GARLAND
`
`SUMMARY*
`
`
`
`Immigration
`
`The panel granted Jaswinder Singh’s petition for review
`of a decision of the Board of Immigration Appeals affirming
`the denial of his application for asylum, holding that the
`Board erred in its adverse credibility determination and its
`internal relocation analysis.
`The panel held that in concluding that Singh was not
`credible the agency misapplied Matter of R-K-K-, 26 I. & N.
`Dec. 658 (BIA 2015), which permits consideration of
`strikingly similar affidavits submitted by asylum applicants
`in unrelated proceedings as a basis for an adverse credibility
`determination. First, the agency erred by relying solely on
`non-unique factual similarities between Singh and other
`unknown declarants from India, without considering
`appropriate factors such as the use of similar words or
`phrases, distinct language and grammar, syntax, and
`narrative structure, or other cues that would suggest the
`affidavit was plagiarized. Additionally, there were serious
`concerns about the government’s unreliable chart and its
`methodology
`in
`selecting
`declarations
`for
`comparison. Finally, the government’s submission of
`declarations with redacted identifying information about the
`declarants, including their names and the location and dates
`of their attacks, raised due process concerns because Singh
`had no way to determine who the declarants were or explain
`
`
`* This summary constitutes no part of the opinion of the court. It has
`been prepared by court staff for the convenience of the reader.
`
`
`
`
`
`SINGH V. GARLAND
`
`
`
`3
`
`why their factual allegations may have been similar to his
`own.
`Addressing the agency’s internal relocation analysis, the
`panel held that the agency erred by failing to hold the
`government to its burden of proving that Singh could safely
`engage in Mann Party activities in areas outside of
`Punjab. The agency also relied on improper speculation in
`concluding that violence against Mann Party members,
`particularly outside of Punjab, was rare. Finally, by focusing
`solely on whether officials within Punjab would follow
`Singh to a new region, the agency failed to consider whether
`Singh may face persecution outside Punjab from local
`authorities, or other actors, based on his future political
`activities.
`The panel remanded for the BIA to conduct a renewed
`credibility determination under Matter of R-K-K-, and to
`conduct a sufficiently individualized analysis of whether
`Singh could safety and reasonably relocate outside of Punjab
`if he continues to engage in Mann Party activities.
`Dissenting, Judge N.R. Smith wrote that because Singh
`did not meaningfully challenge any of the IJ’s internal
`relocation findings in his opening brief, those issues should
`be deemed forfeited and his petition should fail on that basis
`alone. Moreover, in addressing the agency’s internal
`relocation analysis, the majority raised new arguments Singh
`never made and then found contrary evidence in the record
`to refute the IJ’s findings, while ignoring the IJ’s findings of
`fact, and more importantly, the appropriate substantial
`evidence standard of review.
`Judge N.R. Smith wrote that in addressing the agency’s
`credibility determination, the majority cherry picked facts,
`manufactured arguments, ignored the standard of review,
`
`
`
`
`
`4
`
`SINGH V. GARLAND
`
`and improperly substituted its decision for that of the IJ. The
`majority also read Matter of R-K-K- too narrowly in
`concluding
`that
`it applied only
`to applications and
`declarations containing similar language, grammar, and
`spelling. Additionally, the majority inappropriately created
`new requirements for the government’s methodology in
`selecting declarations for comparison.
`
`
`COUNSEL
`
`Alexandra K. Jacobs (argued), Law Office of Robert B.
`Jobe, San Francisco, California; Jagbir S. Terkiana, Terkiana
`Inc., Sunnyvale, California; for Petitioner.
`Roberta O. Roberts (argued) and Craig W. Kuhn, Trial
`Attorneys; Eden L. Patton, Law Clerk; Jonathan A. Robbins,
`Assistant Director; Office of Immigration Litigation; Brian
`M. Boynton, Principal Deputy Assistant Attorney General;
`Civil Division, United States Department of Justice,
`Washington, D.C.; for Respondent.
`
`
`
`
`
`
`
`SINGH V. GARLAND
`
`
`
`5
`
`OPINION
`
`SANCHEZ, Circuit Judge:
`
`Jaswinder Singh, a citizen and native of India, appeals
`the Board of Immigration’s (“BIA”) decision affirming the
`Immigration Judge’s (“IJ”) denial of his petition for
`asylum.1 He argues that the agency erred by misapplying
`Matter of R-K-K-, 26 I. & N. Dec 658 (BIA 2015), which
`permits immigration judges to consider strikingly similar
`affidavits submitted by asylum applicants in unrelated
`proceedings as a basis
`for an adverse credibility
`determination.
` Singh also challenges
`the agency’s
`alternative finding that he could reasonably relocate within
`India. We agree that the agency erred in both respects.
`As Matter of R-K-K- itself and decisions from other
`circuit courts have emphasized, an analysis of inter-
`proceeding similarities involves more than just comparing
`factually similar events. Matter of R-K-K- addresses
`similarities in the way events are described in the affidavits,
`such as the use of identical words or phrases, distinct
`language and grammar, or other cues that suggest the
`affidavit was plagiarized. The agency misapplied Matter of
`R-K-K- by relying solely on non-unique factual similarities
`between Singh and other unknown declarants from India to
`make an adverse credibility finding. The agency further
`
`1 The agency also denied Singh’s claims for withholding of removal and
`relief under the Convention Against Torture. Because Singh’s opening
`brief only challenges the denial of asylum, any other challenges are
`waived. Hernandez v. Garland, 47 F.4th 908, 916 (9th Cir. 2022) (Issues
`not “specifically and distinctly” argued in opening brief may be deemed
`forfeited. (quoting Velasquez-Gaspar v. Barr, 976 F.3d 1062, 1065 (9th
`Cir. 2020))).
`
`
`
`
`
`6
`
`SINGH V. GARLAND
`
`erred in its relocation analysis by failing to hold the
`government to its burden of proof that Singh could safely
`engage in Mann Party activities in areas outside of Punjab.
`We grant the petition and remand to the BIA for further
`proceedings consistent with this opinion.
`I.
`Before coming to the United States, Singh lived in the
`Punjab region and worked as a farmer. On April 13, 2016,
`Singh joined the Shiromani Akali Dal (Amritsar) (“Mann
`Party”). The Mann Party advocates for a separate state of
`Khalistan and for the rights of Sikhs. As a member of the
`Mann Party, Singh attended rallies, hung posters, engaged in
`community service, and participated in recruitment efforts.
`Singh testified that he was attacked twice in 2017
`because of his Mann Party activities. On August 20, 2017,
`Singh was hanging posters for a Mann Party blood donation
`camp when he was approached by a car with four members
`of the opposition Indian National Congress Party (“INC”),
`whom he recognized by the INC logo on the car and from an
`INC rally he had observed. The four men exited the car and
`asked Singh why he was putting up posters. Singh explained
`that the Mann Party was organizing a blood donation drive.
`The men told him to stop putting up posters, join the INC,
`and offered to give him money and illicit drugs to sell if he
`did so. When Singh refused, the four men threw him on the
`ground and beat him with wooden sticks, hockey sticks, and
`baseball sticks for four to five minutes. The men only
`stopped when a group of six or seven witnesses intervened
`to help Singh. The INC members told Singh that if they
`found him again, they would kill him. Singh received
`medical treatment, including bandages and injections for the
`pain, and remained on bed rest for 20–25 days. He sustained
`
`
`
`
`
`SINGH V. GARLAND
`
`
`
`7
`
`internal injuries, bruising, and a head injury that required
`stitches. Singh, his father, and a third individual went to the
`police to report the incident. The police refused to file a
`complaint because the attackers were members of the INC,
`the political party in power at that time. Singh was told that
`he would be arrested if he came back to the police.
`On October 19, 2017, Singh was attacked again by four
`INC members while returning from a religious meeting. The
`INC members told Singh to leave the Mann Party, threw him
`to the ground, and started beating him. One person held a
`weapon to his head. Laborers in a nearby field observed
`what was happening and came to help. One of the attackers
`threatened that if they found Singh again, they were going to
`shoot and kill him. Singh received medical treatment for his
`injuries.
`Following these attacks, INC members came looking for
`Singh at his home three or four times, asking his friends and
`others about Singh’s whereabouts. They could not locate
`Singh because he was hiding on his farmland. With the help
`of his father and his father’s friend, Singh entered the United
`States on or around January 14, 2018. Singh is still a
`member of the Mann Party and testified that he would
`continue to work for the party if he were to return to India.
`Singh timely filed his application for asylum on June 22,
`2018. At the end of Singh’s hearing, the IJ noted that he was
`“concerned about this case because it seems to mirror many
`cases coming from the same region that I have encountered,
`and . . . I have some credibility concerns on that basis.” The
`IJ requested supplemental briefing from
`the parties
`concerning Matter of R-K-K- and allowed additional
`documentary evidence to be submitted on this issue. Along
`with its Matter of R-K-K- brief, the government submitted
`
`
`
`
`
`8
`
`SINGH V. GARLAND
`
`twenty redacted declarations from other asylum seekers
`from India who alleged political persecution by the INC, the
`Bharatiya Janata Party (BJP), or the Shiromani Akali Dal
`Badal party, based upon their membership in a variety of
`different political parties, including the Indian National Lok
`Dal Party and the Mann Party (“RKK Declarations”).
`The government also prepared a demonstrative chart
`comparing the factual allegations in Singh’s affidavit to
`categories of allegations present in many of the RKK
`Declarations. The demonstrative chart displayed columns
`such as “Respondent hanging posters,” “Respondent then
`approached by opposing party,” “Opposing party attempts
`to recruit,” “Respondent declines,” “Respondent then
`punched, kicked, threatened,” “Report to police after first
`attack,” “Police refuse complaint,” “Police threaten with
`arrest,” “Respondent then attacked while traveling home
`from event,” “Attacked with sticks of some kind,” “Saved
`by strangers (usually farmers),” “Internal injuries,” and
`“Family sends them to USA.” Below each column, the chart
`depicted with a checkmark which of the twenty RKK
`Declarations shared in that factual similarity. Based on these
`comparisons, the government’s supplemental brief asserted
`that “[Singh’s] core facts of persecution are identical, or
`nearly identical” to those contained in the redacted
`declarations. Although the twenty declarations varied from
`each other and from Singh’s declaration in different ways,
`the government did not explain how it had chosen these
`particular categories or what criteria it used to conclude that
`the declarations were “nearly identical” to Singh’s evidence.
`Singh filed his response to the government’s Matter of
`R-K-K- submission, arguing
`that
`there were only
`“generalized non-unique
`similarities” between
`the
`declarations that did not warrant an adverse credibility
`
`
`
`
`
`SINGH V. GARLAND
`
`
`
`9
`
`determination. Singh argued that “[i]n India recruiting other
`opposition members of a political party is not unique and
`neither is [the] threat of violence or violence upon declining”
`those recruitment efforts. Nor was it unique or distinctive
`that opposition party members would use “sticks and rods as
`weapons” or that the police in Punjab would refuse to take a
`report of the violence. Singh also identified several
`dissimilarities between his declaration and the RKK
`Declarations, pointing out that only three of the twenty
`declarations involved allegations of persecution of Mann
`Party members by members of the opposition INC Party.
`On August 9, 2021, the IJ denied Singh’s applications
`for asylum, withholding of removal, and relief under CAT.
`The IJ found Singh not credible based on “similarities
`between the respondent’s testimony and that of respondents
`in other removal proceedings.” The IJ noted that Singh and
`a varying combination of the other declarants “were low-
`level workers of their political party,” were hanging posters
`for a party event, were approached by members of an
`opposing party usually in a car, were asked what they were
`doing and instructed to leave the party and join the opposing
`party, were asked to sell drugs, were attacked or threatened
`twice with slaps, punches, kicks, or hits with wooden or
`hockey sticks, were aided by witnesses, “went to a doctor…
`and received medical treatment,” went to the police and were
`refused after the first attack, and fled India with family
`assistance.
`The IJ also found Singh to be evasive and non-
`responsive, particularly concerning his testimony about how
`he had crossed the border. The IJ reviewed Singh’s
`corroborating evidence, comprised of declarations from
`family and community members and county conditions
`evidence, and found it insufficient to overcome his
`
`
`
`
`
`10
`
`SINGH V. GARLAND
`
`credibility concerns. Finally, the IJ concluded that even if
`Singh had testified credibly, the government met its burden
`to show that Singh could relocate to another area of India to
`avoid persecution.
`On January 3, the BIA dismissed Singh’s appeal, citing
`Matter of Burbano, 20 I & N. Dec. 872, 874 (BIA 1994).
`The BIA held that the IJ properly applied Matter of R-K-K-
`by providing Singh “meaningful notice” of the alleged
`similarities and a “reasonable opportunity to explain before
`making a credibility determination based on the totality of
`the circumstances.” It found the IJ’s adverse credibility
`finding was not clearly erroneous, rejecting Singh’s
`argument that the “striking similarities” were “actually
`generalized and not unique factual circumstances.” Finally,
`the BIA found that even if Singh had testified credibly, the
`IJ properly determined that the government met its burden
`of rebutting the presumption of future persecution. It found
`that the IJ’s relocation findings were not clearly erroneous
`because it was unlikely that Singh would be targeted outside
`of Punjab as “a low-level supporter of the Mann Party and a
`non-violent advocate.” This petition for review followed.
`II.
`We have jurisdiction under 8 U.S.C. § 1252. We review
`questions of law de novo, and factual findings under the
`substantial evidence standard. Regalado-Escobar v. Holder,
`717 F.3d 724, 726–27 (9th Cir. 2013). The agency’s
`findings of fact are considered “conclusive unless any
`reasonable adjudicator would be compelled to conclude to
`the contrary.” Ruiz-Colmenares v. Garland, 25 F.4th 742,
`748 (9th Cir. 2022) (cleaned up). Where the BIA cites
`Matter of Burbano and also provides its own review of the
`
`
`
`
`
`SINGH V. GARLAND
`
`
`
`11
`
`evidence and the law, the court reviews both the IJ and the
`BIA’s decision. See id.
`
`III.
`“Under the REAL ID Act, . . . the IJ is authorized to base
`an adverse credibility determination on the totality of the
`circumstances and all relevant factors.” Manes v. Sessions,
`875 F.3d 1261, 1263 (9th Cir. 2017) (per curiam) (internal
`quotation marks omitted). A finding of adverse credibility
`may be premised on an applicant’s “‘demeanor, candor, or
`responsiveness’ as well as the consistency between an
`applicant’s statements and other evidence in the record.”
`Iman v. Barr, 972 F.3d 1058, 1065 (9th Cir. 2020) (quoting
`8 U.S.C. § 1158(b)(1)(B)(iii)). “Speculation and conjecture
`cannot form the basis of an adverse credibility finding,
`which must instead be based on substantial evidence.” Shah
`v. INS, 220 F.3d 1062, 1071 (9th Cir. 2000).
`Whether an IJ may consider notable similarities between
`affidavits submitted by asylum applicants in unrelated
`proceedings has long percolated throughout our nation’s
`courts and immigration proceedings. In Mei Chai Ye v. U.S.
`Dep’t of Justice, the Second Circuit first considered the
`credibility implications of two unrelated asylum applications
`that originated from China, each of which contained
`affidavits with striking similarities in their narrative
`structure, wording, and grammar. 489 F.3d 517, 522 (2d Cir.
`2007). Both petitioners were represented by the same
`attorney. Id. at 520–21. In comparing the two asylum
`affidavits, the IJ found “twenty-three separate places at
`which the two affidavits were strikingly similar in language
`and grammatical structure,” with “similar phrasing and …
`vocabulary,” and where the “identical portions appeared . . .
`in the exact same order in both affidavits.” Id. at 521–22.
`
`
`
`
`
`12
`
`SINGH V. GARLAND
`
`The IJ reasonably concluded that the affidavits were “so
`blatantly similar in both form and substantive details” such
`that they “could not have been the result of honest applicants
`inserting truthful information into standardized templates.”
`Id. at 525.
`Mei Chai Ye cautioned, however, about “the “dangers
`inherent in relying on inter-proceeding similarities.” Id. at
`526. “To assume that one asylum applicant is responsible
`for, or even aware of, the striking similarities that appear in
`an unrelated applicant’s submissions
`is much more
`problematic” than relying on internal inconsistencies within
`the same proceeding. Id. at 519-20. The Second Circuit
`explained that there are many ways in which substantially
`similar affidavits from unrelated asylum applicants may
`have an innocent explanation: (1) “both applicants have
`inserted truthful information into a similar standardized
`template”; (2) “the different applicants employed the same
`scrivener, who wrote up both stories in his own rigid style”;
`(3) the other applicant “plagiarized the truthful statements of
`the petitioner”; or (4) “the similarities resulted, not from the
`original documents themselves, but rather from inaccurate or
`formulaic translations—which unaffiliated applicants would
`not be in a position to discover or contest.” Id. at 520, 526.
`Because the IJ carefully followed procedural safeguards that
`gave the petitioner an opportunity to explain or contest the
`similarities, and the IJ took pains to ensure that these
`similarities were not the result of mere coincidence, the court
`upheld the agency’s adverse credibility determination. Id. at
`527. In doing so, Mei Chai Ye urged the BIA to develop
`appropriate guidelines
`for analyzing
`inter-proceeding
`similarities in future immigration cases. Id. at 526.
`The BIA formalized those guidelines in Matter of R-K-
`K-. The respondent there was an asylum seeker from India
`
`
`
`
`
`SINGH V. GARLAND
`
`
`
`13
`
`who alleged two incidents of arrest and abuse by the police
`in 2010. 26 I. & N. Dec. at 662–63. But his testimony rang
`some alarm bells. His brother, who had been granted asylum
`in 2009, used the “same or remarkably similar language” to
`describe acts of abuse by police years earlier. Id. at 661–63.
`The IJ observed that “both applications shared identical
`wording, typographical and spelling errors, and spacing
`irregularities in describing the same events.” Id. at 664.
`Both declarations also used “the same distinctive
`descriptions of the alleged events,” each alleging that police
`forced them to undress, beat them on their back and buttocks
`with sticks or batons, and pulled their legs apart during
`nighttime detentions. Id. at 663–65. In addition, the
`respondent’s declaration used plural pronouns to describe
`just himself, whereas his brother’s application properly used
`those same plural pronouns to describe violence against
`himself and his wife. Id. at 663.
`On appeal, the BIA addressed whether “in making an
`adverse credibility determination, an [IJ] can consider
`significant similarities between statements submitted by
`applicants in different proceedings,” and what procedural
`safeguards it should adopt “to preserve the fairness of the
`proceedings.” 26 I. & N. Dec. at 659, 661. As in Mei Chai
`Ye, the BIA acknowledged the difficulty that these questions
`present. It noted that “some inter-proceeding similarities are
`so significant that, when left unexplained, they cannot be
`ignored,” while cautioning that other “innocent similarities
`may be mistakenly interpreted as evidence of falsity.” Id. at
`661. Thus, the BIA resolved that IJs “may rely on inter-
`proceeding similarities as part of an adverse credibility
`determination, but we must also review such determinations
`with ‘an especially cautious eye.’” Id. (quoting Mei Chai Ye,
`489 F.3d at 519–20).
`
`
`
`
`
`14
`
`SINGH V. GARLAND
`
`Matter of R-K-K adopted a three-part framework for
`adjudicating cases that involve inter-proceeding similarities:
`(1) “[T]he [IJ] should give the applicant meaningful notice
`of the similarities that are considered to be significant”;
`(2) “[T]he [IJ] should give the applicant a reasonable
`opportunity to explain the similarities”; and (3) “[T]he [IJ]
`should consider the totality of the circumstances in making
`a credibility determination. Each of these steps must be done
`on the record in a manner that will allow the [BIA] and any
`reviewing court to ensure that the procedures have been
`followed.” Id. Applying this framework, the BIA
`determined that the IJ’s adverse credibility ruling was not
`clearly erroneous. The “nearly identical wording” in the
`applications properly raised credibility concerns and the
`textual and narrative similarities were “too numerous and
`obvious to be coincidental.” Id. at 665. The BIA further
`concluded that the IJ did not err in finding that the
`respondent’s explanations
`for
`the “nearly verbatim
`statements” were insufficient and “not persuasive,” and the
`IJ had given proper consideration to the totality of the
`circumstances and afforded
`the
`respondent ample
`opportunity to address these concerns. Id. at 665–66.
`In setting out a framework to guide immigration judges
`on inter-proceeding similarities, it is clear that Matter of R-
`K-K- addresses similarities in the use of language, grammar
`and syntax, and narrative structure between affidavits that
`are so striking they call into question whether the affidavit
`has been plagiarized. In its discussion of the framework for
`analysis, the BIA explained:
`
`Identification of a substantial number of
`instances where the same or remarkably
`similar language is used to describe the same
`
`
`
`
`
`SINGH V. GARLAND
`
`
`
`15
`
`kind of incident or encounter would tend to
`raise credibility questions that should be
`further addressed. This is particularly true
`where there is additional material in both
`statements that “wouldn’t necessarily have to
`be mentioned but [was] mentioned.” Mei
`Chai Ye v. U.S. Dep’t of Justice, 489 F.3d at
`521 (quoting the Immigration Judge). But
`the presence of even a relatively few
`similarities could raise the same credibility
`concerns if, in the context of the overall
`asylum claim, distinct language was used or
`unique factual circumstances were repeated
`without reasonable explanation. See, e.g.,
`Dehonzai v. Holder, 650 F.3d 1, 8 (1st Cir.
`2011) (rejecting the applicant’s explanation
`that language describing a beating with a
`bundle of wires attached to a tennis ball,
`which was “virtually
`identical”
`to his
`cousin’s statement, was “mere coincidence”).
`
`Id. at 661 (emphasis added).
`Accordingly, Matter of R-K-K- is concerned with more
`than just factually similar events. It instructs that the telltale
`signs of a canned or plagiarized affidavit are the ways in
`which events are described in the affidavit, such as the use
`of identical phrases or words, the same grammatical
`mistakes and punctuation, the use of distinctive language, or
`the unnecessary addition of extraneous detail. In Dehonzai,
`upon which Matter of R-K-K- relies, the First Circuit upheld
`an adverse credibility determination of an asylum applicant
`that “virtually copied” a story attributed to his alleged cousin
`in an Amnesty International Report. 650 F.3d at 6, 8. Not
`
`
`
`
`
`16
`
`SINGH V. GARLAND
`
`only did the affidavit contain unique factual details about the
`alleged beating, claiming that his persecutors “beat [him]
`with a bundle of electric wire with tennis ball at the end, the
`ball continually struck my back [sic],” but the language “was
`virtually identical to the language attributed to [his cousin]
`within the Amnesty International report.” Id. at 8.
`Here, the IJ did not rely on any similarities in language,
`grammar, or narrative structure between Singh’s affidavit
`and any of the twenty redacted declarations submitted by the
`government below. As the government concedes, Singh’s
`affidavit substantially differs in its use of language, wording,
`and structure to describe the events in question. Instead, the
`IJ’s “principal concern” was the alleged factual similarities
`between Singh’s
`testimony and
`that of
`the RKK
`Declarations. The IJ found that because several unknown
`declarants from India had alleged that opposition party
`members attempted to recruit them while they were putting
`up political posters, they were beaten by sticks or rods, they
`were turned away by police and threatened with arrest, and
`they had to flee to the United States, Singh’s own account of
`factually similar events lacked credibility.
`We conclude that the agency misapplied Matter of R-K-
`K-. Relying exclusively on broad factual similarities to
`trigger credibility suspicion runs counter to the special
`caution required under Matter of R-K-K- and its express
`focus on finding striking similarities in the language,
`grammar and structure of related affidavits. 26 I & N. Dec.
`at 661–62. There are several problems with the way the
`agency approached its inter-proceeding analysis.
`First, by focusing exclusively on broad factual
`similarities between the declarations, the IJ erred in applying
`Matter of R-K-K- too expansively. As required by Matter of
`
`
`
`
`
`SINGH V. GARLAND
`
`
`
`17
`
`R-K-K-, the IJ did not identify any linguistic or grammatical
`similarities between Singh’s declaration and the RKK
`Declarations that would suggest Singh’s affidavit had been
`plagiarized, much less “strikingly similar” words or phrases
`or grammatical cues. While the agency has an interest in
`rooting out “canned” applications, see Mei Chai Ye, 489 F.3d
`at 524, relying solely on non-unique factual similarities to
`deem an applicant not credible undermines a fundamental
`principle of asylum law that persecution sometimes occurs
`through widespread or systematic actions by the government
`or by its acquiescence to third-party harm. See Mgoian v.
`I.N.S., 184 F.3d 1029, 1035 (9th Cir. 1999) (a petitioner is
`eligible for asylum if they can “show a ‘pattern or practice’
`of persecution against a group of which she is a member”
`(quoting Kotasz v. INS, 31 F.3d 847, 853 (9th Cir. 1994)));
`Knezevic v. Ashcroft, 367 F.3d 1206, 1213 (9th Cir. 2004)
`(“An applicant need not prove that he will be singled out for
`persecution if he can prove a pattern or practice of
`persecution of people similarly situated to the applicant, who
`are members of a protected group.” (internal citations
`omitted)). That multiple asylum applicants from the same
`region of India might describe similar forms of persecution
`does not necessarily imply their accounts are false and
`should be discredited.
`Second, instead of relying on the requisite similarities in
`language and grammar, the agency based its Matter of R-K-
`K- analysis on the government’s unreliable demonstrative
`chart and declarations. The agency’s reliance on this
`submission raises serious questions about its methodology.
`See Mei Chai Ye, 489 F.3d at 527 (“On the other hand, our
`holding indicates that we would view much more skeptically
`an adverse credibility finding by an IJ who, in relying on
`
`
`
`
`
`18
`
`SINGH V. GARLAND
`
`less rigorous
`inter-proceeding similarities, adopted a
`approach than that employed by IJ Vomacka in this case.”)
`To begin, it is unclear how the government selected the
`declarations that formed the basis of its Matter of R-K-K-
`brief. The government provided “twenty redacted affidavits
`from other asylum cases originating from India” that
`purportedly “shared striking similarities”
`to Singh’s
`declaration. Without knowing more about the government’s
`selection process or the broader universe of asylum
`applications from India, it is difficult to draw any
`conclusions about the chosen twenty.
` Suppose the
`government had reviewed one hundred asylum declarations
`and found eighty to describe wholly dissimilar events. The
`IJ—or a reviewing court adopting a more rigorous
`approach—would conclude
`that
`this unrepresentative
`sample tells the court very little about whether Singh’s
`affidavit was filled with canned statements or reflected true
`events.2
`Equally unsound was the IJ’s reliance on broad factual
`similarities that could be present in countless asylum
`applications.
` The government’s demonstrative chart
`displayed columns such as “Respondent hanging posters,”
`“Respondent
`then approached by opposing party,”
`“Opposing party attempts to recruit,” and “Respondent
`declines,” followed by checkmarks indicating that most of
`
`2 Our dissenting colleague suggests that we have imposed a new
`requirement on the government “to disclose how they selected the
`[declarations].” Not so. We merely illustrate the problems that arise
`when the IJ relies on non-unique declarations of unknown provenance to
`make questionable “similarity” findings that are not grounded in
`“remarkably similar” language or grammar. It is the duty of a reviewing
`court to “review such determinations with ‘an especially cautious eye.’”
`Matter of R-K-K-, 26 I & N. Dec. at 661.
`
`
`
`
`
`SINGH V. GARLAND
`
`
`
`19
`
`the RKK Declarations shared in that similarity. Certain
`columns draw arbitrary distinctions about the same event; it
`is unclear why an opposition party’s attempt to recruit and
`respondent’s refusal represents
`two different factual
`circumstances. Other columns are so general as to be
`virtually meaningless, such as “Respondent then punched,
`kicked, threatened,” “Police refuse complaint,” and “Family
`sends them to USA.” All viable claims for asylum are likely
`to allege that the applicant suffered some form of harm or
`threat of harm, the government participated in or acquiesced
`in that harm, and the applicant made their way to the United
`States to seek asylum relief. See 8 U.S.C. §§ 1158(b),
`1101(a)(42)(A). Tallying the number of checkmarks for
`these non-unique categories does not convert this process
`into a rigorous one.
` Indeed, it largely reveals its
`arbitrariness.
`The IJ concluded that “the R-K-K- declarations are not
`merely similar; they contain slight variations on essentially
`identical events occurring in essentially identical order and
`recounted in essentially identical manners.” The IJ did not
`explain, however, which variations or events he found
`“slight” or important, or whether he simply engaged in a
`number-counting exercise.
` There
`is no discernible
`methodology to the IJ’s determination that Singh’s affidavit
`was “essentially identical” to an amalgamation of twenty
`other declarations.3
`
`
`3 Our court has granted several petitions for review by Mann Party
`members who alleged very similar claims of persecution. See Singh v.
`Garland, 57 F.4th 643, 649 (9th Cir. 2022) (attacked twice by INC
`members and suffered internal injuries



