Dong Gao, Ziang Zhen Gao v. Board of Immigration Appeals

482 F.3d 122, 2007 U.S. App. LEXIS 7148

• Court: Court of Appeals for the Second Circuit
• Decided: March 28, 2007
• Docket: Docket 04-4020-ag
• Status: Published

Opinion

UNDERHILL, District Judge:

Petitioners Dong Gao and Xiang Zhen Gao, a married couple, are natives and citizens of the People’s Republic of China. They challenge the decision of the Board of Immigration Appeals (“BIA”) that affirmed the oral ruling of an immigration judge (“IJ”), denying their applications for asylum and withholding of removal under the Immigration and Naturalization Act, and directed their removal to China. In re Dong Gao & Xiang Zhen Gao, Nos. A76 132 444, A77 297 011 (B.I.A. July 16, 2004), aff'g Nos. A76 132 444, A77 297 011 (Immig. Ct. N.Y. City Sept. 24, 2002). Xiang Zhen Gao’s (“Mrs.Gao”) asylum claim is based primarily on her forced sterilization after the couple violated China’s family planning laws. Dong Gao (“Mr.Gao”) has applied for asylum derivatively. See Zhou Yun Zhang v. INS, 386 F.3d 66, 71-72 (2d Cir.2004) (discussing availability of political asylum for husband based on wife’s forced abortion or sterilization).

The IJ denied the Gaos’ petitions for asylum based solely on his determination that the couple lacked credibility. That adverse credibility determination was grounded exclusively on three purported inconsistencies in their testimony. None of those three subsidiary factual findings was supported by substantial evidence. Rather, the IJ relied on misstatements of the facts in the record and speculation concerning purported testimony of Mr. Gao from a previous asylum hearing — testimony that he did not admit to having given and of which there is no record. In a short opinion, the BIA affirmed the IJ’s decision for the reasons he set forth but corrected one of the IJ’s erroneous statements of fact. Accordingly, the petition for review is granted, the agency’s decision is vacated, and the case is remanded for further proceedings consistent with this opinion.

BACKGROUND

Both Mr. and Mrs. Gao were born in the village of Longmen in Changle County, Fujian Province, China. They were married in a traditional marriage ceremony on March 20, 1992, and registered their marriage on November 21, 1992. The Gaos have two children.

Mrs. Gao was pregnant with her first child in 1993. During that pregnancy, she went into hiding in order to prevent the government from learning that she was pregnant. She testified that she gave birth to that child, a daughter, at her aunt’s house in Hoshan, a neighboring village, on July 16, 1993. A midwife delivered the baby. After giving birth, Mrs. Gao left her daughter at her aunt’s house in order to prevent the government from learning that she had a child.

The Gaos testified that Mrs. Gao again went into hiding at her aunt’s house when she was pregnant for a second time in May of 1994. Mrs. Gao explained that she was afraid that the government officials would discover her pregnancy and force her to undergo an abortion because she already had one child.

In July or August 1994, while she was in hiding, the village family planning cadre came to the Gaos’ house in search of Mrs. Gao because they suspected that she already had given birth to one child and was pregnant again.

Mrs. Gao testified that she feared she would be forced to undergo an abortion if discovered by the government officials. She explained that she went to the county hospital where a cousin worked, pretended to be more than ten days past her due date, and requested a Caesarian section. Mrs. Gao had a C-seetion, and the Gaos’ son was born at the Changle County hospital on November 28, 1994. Mrs. Gao testified that she remained at the hospital for one week following her son’s birth.

Two weeks after the birth, Mr. Gao tried to register their son with local authorities. Before they were permitted to register the birth, the village cadre required the Gaos to pay a fine for violating family planning policies, that is, for having two children without the mandated spacing. The Gaos paid that fine on February 15,1995.

After their payment of the fine, in April 1995, the village cadre notified the Gaos that one of them would have to be sterilized. Initially, the Gaos successfully bribed a friend of Mr. Gao’s father, who worked in the village government, in order to avoid an involuntary sterilization. In October of 1995, however, more than ten government officials surrounded the Gaos’ house in order to take one of them to be sterilized. Eventually, they took away Mrs. Gao, detained her overnight, and took her the following day to the county hospital. She underwent a compelled bilateral tubal ligation without being fully anaesthe-tized. 1

In 1999, government officials demanded that Mrs. Gao pay another fine for violating the family planning policies. They threatened her with arrest if she did not pay the additional fine. She refused to pay, then left her home, and came to the United States because she feared arrest and imprisonment.

In the meantime, in August 1998, Mr. Gao entered the United States and thereafter applied for asylum based on China’s coercive family planning policy, including the forced sterilization of his wife. He testified at an asylum hearing on November 5, 1999; that day, the IJ issued an order denying Mr. Gao’s application for asylum and denying his application for withholding of removal. On January 10, 2000, while Mr. Gao’s appeal of that decision was pending before the BIA, Mrs. Gao entered the United States. She also applied for asylum based on China’s family planning policy and her forced sterilization.

On appeal, the tape recordings and transcript containing the testimony of Mr. Gao’s hearing and the IJ’s oral decision, necessary for the BIA’s review, were missing. In re Dong Gao, No. A76 132 444 (B.I.A. May 9, 2001). On May 9, 2001, the BIA returned the case to the IJ for further action, ordering the IJ to “take such steps as are necessary and appropriate to enable preparation of a complete transcript of the proceedings and a decision of the Immigration Judge including a new hearing, if necessary.”

On remand, the Gaos’ cases were consolidated, and they both testified at a removal hearing on September 24, 2002. In an oral decision issued that same day, the IJ ruled against them in reliance on three identified inconsistencies in their testimony, involving: (1) the birthplace of their first child, (2) problems encountered during Mrs. Gao’s second pregnancy, and (3) the duration of Mrs. Gao’s hospitalization following her forced sterilization. In his oral ruling, the IJ referenced Mr. Gao’s testimony from the first hearing although the tape recording and transcript had not surfaced, and there was no record of that testimony in evidence.

On July 16, 2004, in a per curiam decision, the BIA affirmed the IJ’s decision for the reasons stated therein with one exception. The BIA noted that, with respect to the third identified inconsistency, the discrepancy between the Gaos’ testimony “did not relate to hospitalization after a sterilization procedure, but to hospitalization after the birth of the couple’s second child.”

DISCUSSION

I. Scope of Review

When the BIA issues an opinion, that decision becomes the focus of our review. See Yan Chen v. Gonzales, 417 F.3d 268, 271 (2d Cir.2005). When the BIA adopts, modifies, or supplements the decision of an IJ, the scope of the Court’s review will vary. The extent to which we review the IJ’s decision directly depends on the approach employed by the BIA in affirming the IJ. See Ming Xia Chen v. Board of Immigration Appeals, 435 F.3d 141, 144 (2d Cir.2006) (discussing BIA’s various “techniques” and the corresponding scope of appellate review).

For example, when the BIA issues an opinion and does not adopt the IJ’s decision to any extent, we review only the BIA’s decision. See Yan Chen, 417 F.3d at 271. At the other extreme, when the BIA expressly adopts the findings and reasoning of the IJ, we review only the IJ’s decision, as if it were the BIA’s. See Chun Gao v. Gonzales, 424 F.3d 122, 124 (2d Cir.2005).

On occasion, the BIA will reference certain portions of the IJ’s decision, sometimes rejecting one or more aspects of the IJ’s reasoning. For instance, when the BIA affirms the IJ’s decision in all respects except one, the Court reviews the IJ’s decision “minus-the single argument for denying relief that was rejected by the BIA.” Xue Hong Yang v. U.S. Dep’t of Justice, 426 F.3d 520, 522 (2d Cir.2005). Likewise, when the BIA issues a short opinion affirming the IJ decision in part and modifying it in part, we will review the IJ’s decision as modified by the BIA. Ming Xia Chen, 435 F.3d at 144. That review will be confined to the rationale of the IJ on which the BIA relied. See Zerrei v. Gonzales, 471 F.3d 342, 346 (2d Cir.2006).

Other times, the BIA will adopt the IJ’s reasoning and supplement it, affirming the result on additional grounds. In those instances, the Court reviews the IJ’s decision as supplemented by the BIA. See Yu Yin Yang v. Gonzales, 431 F.3d 84, 85 (2d Cir.2005). Similarly, when the BIA adopts the entirety of an IJ’s decision, then adds alternative grounds, this Court may review the underlying IJ decision. See Ming Xia Chen, 435 F.3d at 144 (noting that the Court was not confined to reviewing the BIA’s two alternative grounds for denying relief when BIA had also adopted IJ’s determinative adverse credibility finding).

When the BIA agrees with an IJ’s ultimate credibility determination but emphasizes particular aspects of the IJ’s reasoning, the Court reviews both the BIA’s and the IJ’s opinions, including those portions of the IJ’s decision that the BIA did not explicitly discuss. See Yun-Zui Guan v. Gonzales, 432 F.3d 391, 394 (2d Cir.2005).

In the instant case, the IJ denied relief based solely on an adverse credibility determination that in turn relied on three subsidiary factual findings based on purportedly inconsistent testimony. The BIA affirmed the IJ’s decision “for the reasons stated therein, with the following exception,” then noted that the petitioners had not testified inconsistently concerning Mrs. Gao’s hospitalization following the sterilization — as the IJ had found — but had testified inconsistently with respect to the duration of her hospital stay after giving birth to their son.

In other words, the BIA corrected one subsidiary factual finding of the IJ, but upheld his ultimate adverse credibility finding. It is unclear whether the BIA intended to extract the third inconsistency from the IJ’s reasoning, concluding that the remaining two subsidiary grounds were legally sufficient, or whether it instead substituted the “actual inconsistency” it identified in the record for the “profound inconsistency” that the IJ had erroneously identified and relied upon. Because the BIA did not rule independently on the credibility of the petitioners — and indeed it could not 2 — but held that the IJ’s adverse credibility finding was not clearly erroneous, it must have employed the former approach. In other words, we read the BIA’s decision as extracting the one erroneous ground on which the IJ relied, noting that the record did contain a somewhat similar discrepancy, but not relying on that newly articulated inconsistency.

In these circumstances, most akin to cases when the BIA affirms an IJ in all respects but one, the Court must review the IJ’s decision as modified by the BIA, that is, without the erroneous subsidiary fact that the IJ found and upon which he relied in concluding that the Gaos lacked credibility.

II. Standard of Review

In the immigration context, the Court reviews administrative factual findings, including credibility determinations and the subsidiary findings underpinning them, under the substantial evidence standard. See Wu Biao Chen v. INS, 344 F.3d 272, 275 (2d Cir.2003) (per curiam). Following the Supreme Court, this Court has defined “ ‘substantial evidence’ to mean ‘more than a mere scintilla’: it is ‘such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.’ ” Li Zu Guan v. INS, 453 F.3d 129, 135 (2d Cir.2006) (quoting Consolidated Edison Co. v. NLRB, 305 U.S. 197, 229, 59 S.Ct. 206, 83 L.Ed. 126 (1938)). Under the substantial evidence standard, the Court treats factual findings as “conclusive unless any reasonable adjudicator would be compelled to conclude to the contrary.” 8 U.S.C. § 1252(b)(4)(B). Re-latedly, this Court grants “particular deference” in applying the substantial evidence standard to credibility findings based on demeanor. Karaj v. Gonzales, 462 F.3d 113, 116 (2d Cir.2006); see also Jin Chen v. U.S. Dep’t of Justice, 426 F.3d 104, 113 (2d Cir.2005) (“We give particular deference to credibility determinations that are based on the adjudicator’s observation of the applicant’s demeanor, in recognition of the fact that the IJ’s ability to observe the witness’s demeanor places her in the best position to evaluate [credibility].”). The Court is “mindful that the law must entrust some official with responsibility to hear an applicant’s asylum claim, and the IJ has the unique advantage among all officials involved in the process of having heard directly from the applicant.” Zhou Yun Zhang, 386 F.3d at 73.

Nevertheless, “the fact that the BIA has relied primarily on credibility grounds in dismissing an asylum application cannot insulate the decision from review.” Ramsameachire v. Ashcroft, 357 F.3d 169, 178 (2d Cir.2004). To the contrary, an IJ’s adverse credibility finding “constitutes the beginning not the end of our inquiry.” Secaida-Rosales v. INS, 331 F.3d 297, 307 (2d Cir.2003). Although the Court does not review the petitioners’ credibility de novo, our “exceedingly narrow” inquiry must ensure that the IJ’s determinations were not arbitrary or capricious. Zhou Yun Zhang, 386 F.3d at 74. In other words, our review ensures that credibility findings are based upon neither a misstatement of the facts in the record nor bald speculation or caprice. Id.

Notably, when the outcome of an asylum application “rises and falls purely on an IJ’s credibility finding, courts have been particularly concerned that the decision-maker carefully detail the reasoning leading to the adverse finding.” Secaida-Rosales, 331 F.3d at 307. When rejecting an applicant’s testimony, for example, an IJ must provide “specific, cogent” reasons for doing so. Id; accord Ahmad v. INS, 163 F.3d 457, 461 (7th Cir.1999); Senathirajah v. INS, 157 F.3d 210, 216 (3d Cir.1998). “Those reasons must bear a legitimate nexus to the [credibility] finding and must be valid grounds for disregarding an applicant’s testimony.” Secaida-Rosales, 331 F.3d at 307 (internal quotation marks and citations omitted).

Although the substantial evidence standard leaves fact-finding to the agency, “it does not permit an appellate court to defer to unreasoned rulings, or those based on legal error, faulty analysis, or misreadings of the record.” Li Zu Guan, 453 F.3d at 136. Specifically, “[a]dverse credibility determinations based on speculation or conjecture, rather than on evidence in the record, are reversible.” Secaida-Rosales, 331 F.3d at 307 (citing Chen Yun Gao v. Ashcroft, 299 F.3d 266, 272 (3d Cir.2002)); accord Salaam v. INS, 229 F.3d 1234, 1238 (9th Cir.2000). Moreover, if an IJ’s credibility determination analyzing testimony is based on flawed reasoning, it likewise will not satisfy the substantial evidence standard. Secaida-Rosales, 331 F.3d at 307.

Credibility determinations that are based on the IJ’s analysis of testimony, as opposed to demeanor, are granted less deference. See Jin Chen, 426 F.3d at 113; cf. Secaida-Rosales, 331 F.3d at 307 (distinguishing IJ credibility determinations that are based on analysis of testimony from those based on witness demeanor, which merit greater deference). The Court reviews such credibility determinations only with the “high degree of deference that is normally given to an IJ’s factual determinations.” Lin Zhong v. U.S. Dep’t of Justice, 461 F.3d 101, 123 n. 26 (2d Cir.2006).

Finally, the Court will confine its review to the reasons given by the agency and will not search the record for alternative reasons to affirm. See Karaj, 462 F.3d at 116.

III. The Decision Under Review

We vacate the agency’s rejection of the Gaos’ petitions for asylum and withholding of removal based on China’s coercive family planning policies because that rejection rested exclusively on an adverse credibility finding that was not supported by substantial evidence. The IJ’s analysis of the petitioners’ testimony and his reliance on purported inconsistencies in that testimony were clearly erroneous.

In concluding that the Gaos lacked credibility, the IJ relied solely on three purported inconsistencies in the petitioners’ testimony. Those inconsistencies related to: (1) the birthplace of the Gaos’ “first child” and (2) whether Mrs. Gao had any “problems” during her second pregnancy. Neither of those factual findings is supported by substantial evidence in the record.

The IJ also relied on (3) what he characterized as a “profound inconsistency” concerning the duration of Mrs. Gao’s hospitalization following her forced sterilization. Although the BIA noted that the third inconsistency identified by the IJ was in error, the BIA’s remarks concerning a different “actual inconsistency,” relating to Mrs. Gao’s hospitalization following the birth of her son, do not provide additional support for the IJ’s credibility determination.

A. The Three Purported Inconsistencies

1. Birthplace of the “First Child”

First, the IJ found that Mrs. Gao was “internally inconsistent” concerning the birthplace of her first child. Relatedly, with respect to the first child’s place of birth, he found that Mr. Gao was unable to reconcile his purported testimony from the 1999 hearing with his testimony in 2002.

When explaining Mrs. Gao’s internal inconsistency regarding the first child, the IJ mistakenly discussed the birth of the Gaos’ son, who is their second child:

The respondent was internally inconsistent with regards to the place of her first child’s birth. Respondent testified that her son was born at her aunt’s house. She then recanted that testimony and indicated that her son was born in the hospital.

That description mischaracterizes Mrs. Gao’s testimony, and the IJ’s finding that Mrs. Gao was “internally inconsistent” is not based on substantial evidence.

During direct examination, Mrs. Gao testified that she gave birth to her daughter — her first born — at her aunt’s house and that she delivered her second child, a son, at the county hospital. On cross, she was first asked, “Now you indicated that your first child was born at your aunt’s home, correct?” Mrs. Gao replied affirmatively. She was then asked, “And are you aware that previously your husband testified at his hearing that your first child was born in the Changles Hospital?”

At that point, there was confusion with the interpretation of Mrs. Gao’s testimony. The response was initially recorded: “I didn’t know but my daughter was born at my aunt’s house.” After her attorney objected to the translation, the tape recording was played and the interpreter stated that he had not accurately interpreted her response. He stated: “No, actually, the applicant said my son, my son was born at my aunt’s house.” Her attorney again objected to the interpretation, stating that he believed she had said, “[t]he first, the first child,” then that he believed she had said “daughter.”

The tape recording was played a second time; there was additional confusion, but the interpreter stood by his second interpretation: “The correct translation should be my son was born at my aunt’s house.” Upon further questioning, Mrs. Gao, however, repeated her direct testimony: “My daughter was born at my aunt’s house. My son was born at Changles County Hospital.”

The government lawyer then repeated to Mrs. Gao that her husband had testified at his initial asylum hearing in 1999 that their first-born child, their daughter, was born in a hospital. In response, Mrs. Gao stated again, consistently, that her daughter was born at her aunt’s house. When asked why her husband would have stated otherwise, she replied: “Maybe he was nervous at the time.”

In short, it is not accurate to say that Mrs. Gao “testified that her son was born at her aunt’s house [and] ... then recanted that testimony and indicated that her son was born in the hospital.” On both direct and cross-examination, Mrs. Gao testified that her daughter was born at her aunt’s house and her son was born at the hospital. There was one instance, in response to a question concerning the birth of her “first child,” when the interpreter— after listening to the tape recording of her testimony twice — asserted that Mrs. Gao’s response was: “my son was born at my aunt’s house.” Still, when the question was repeated, Mrs. Gao testified consistently with her direct testimony and earlier testimony on cross-examination. 3

Moreover, the IJ’s analysis of that testimony — referencing her “first child” but then discussing the testimony surrounding the birth of the Gaos’ son — was clearly erroneous.

With respect to her husband’s testimony on the issue, the IJ commented that Mr. Gao had testified at the 1999 hearing that “the first child” was born at a hospital, although at the later hearing he testified “the first child” was born at the aunt’s house.

The respondent’s husband testified at a previous hearing that the first child was born at a hospital, although he testified today that the first child was born at the aunt’s home. When confronted with this significant variance, respondent’s husband simply stated that that was due to the fact that he was nervous.

Again, the IJ’s finding is not supported by substantial evidence in the record. At the 2002 hearing, Mr. Gao testified his “first child” was born at the aunt’s house. Wfiien the government lawyer asked why he had previously testified that she had been born at the hospital, he replied first: “At that time, I didn’t, I didn’t remember.” He later replied: “Because I was very nervous when I was in front of the judge. Maybe I made a mistake.”

The IJ never set forth any basis for the finding that Mr. Gao “testified at a previous hearing that the first child was born at a hospital,” and there is no evidence in the record to support that finding. The IJ appears simply to have adopted the assumption implicit in the cross-examiner’s question; an assumption underlying a question, however, is not evidence. Moreover, Mr. Gao’s two comments, notably, “Maybe I made a mistake,” do not amount to an admission that he testified inconsis tently at the 1999 hearing with respect to the first child’s place of birth. Cf. Lin Zhong v. U.S. Dep’t of Justice, 461 F.3d at 126 (noting — in reviewing credibility determination based in part on household registry not admitted into evidence — that an IJ cannot rely on non-record evidence in making ultimate credibility determination); In re S-M-J-, 21 I. & N. Dec. 722, 728 (BIA 1997) (stating that “any evidence relied upon by the Immigration Judge must be included in the record so that the Board can meaningfully review any challenge to the Immigration Judge’s decision on appeal”).

Strict evidentiary rules do not govern immigration proceedings. See Felzcerek v. INS, 75 F.3d 112, 116 (2d Cir.1996). Nevertheless, “it is black letter law that questions asked by counsel are not evidence.” Washington v. Schriver, 255 F.3d 45, 61 (2d Cir.2001). Thus, absent other evidence, an IJ cannot rely on facts set forth only in a lawyer’s question. Cf. United States v. DeFillipo, 590 F.2d 1228, 1239-40 (2d Cir.1979) (noting “the traditional evi-dentiary rule that a question (which assumes a fact) may become improper on cross-examination, because it may by implication put into the mouth of an unwilling witness, a statement which he never intended to make, and thus incorrectly attribute to him testimony which is not his”) (internal quotation marks and citation omitted). 4

Accordingly, we conclude that substantial evidence does not support the IJ’s finding of an inconsistency between the purported 1999 testimony and Mr. Gao’s testimony at the 2002 hearing.

2. Problems During Second Pregnancy

The second claimed inconsistency on which the IJ relied related to what transpired when Mrs. Gao was pregnant for the second time. Both Mr. and Mrs. Gao testified consistently on this subject at the 2002 hearing. The two described the family planning cadre who came to their home in search of a pregnant Mrs. Gao, suspecting the couple already had a child. The IJ, however, again relied on the supposed testimony of Mr. Gao at his first hearing.

Here, the IJ based his credibility determination on the Gaos’ inability to reconcile Mr. Gao’s earlier purported testimony with their testimony in 2002. The IJ relied on the fact that Mrs. Gao “could not explain” her husband’s supposed testimony from the 1999 hearing, which she did not attend, and that Mr. Gao “when confronted with his previous testimony that they had no problems or difficulties with Chinese officials during the second pregnancy, explained that he said it because he didn’t want to elaborate and the events were too upsetting to explain.”

The IJ asserted that Mr. Gao had previously testified “that they had no difficulties during her second pregnancy, had no confrontations with anyone, and had no problems with the Chinese government.” Again, there was no record of any such testimony in evidence. Rather, during the 2002 hearing, the government attorney and the IJ repeatedly stated simply that Mr. Gao had indicated in 1999 that the couple had “no problems” during the second pregnancy:

Q: Can you explain why, sir, at your asylum hearing before the Court, you indicated that you and your wife had no problems during your wife’s second pregnancy?

A: The second pregnancy, do you mind to repeat again?

Q: You indicated at your hearing that you and your wife had no problems during your wife’s second pregnancy. Can you explain why your testimony is different today?

A: You say I didn’t have any problem when I was at my first asylum hearing?

Judge: No, sir. No. No. You said at your last hearing that you and your wife had no problems during her second pregnancy. Why would you say that?

A: At that time the cadre had come. They wanted my wife to come out. I remember that point very clearly. Both my, both we were hiding, both—

Interpreter: Your Honor—

Judge: Are you getting any of this?

Interpreter: No.

Judge to Mr. [Gao]: Sir, do you want to repeat this or do you want to stop for a second here?

A: At that time they have notified my mother they wanted my wife to come out. I wouldn’t, I wouldn’t go back home. I was so scared I wouldn’t go back home to sleep.

Q [Ms. Schroeck]: Sir, that doesn’t explain why at your first hearing you indicated you had no problems during your wife’s second pregnancy and now you’re saying you were contacted by the family planning or they came looking for your wife during your wife’s second pregnancy. Can you explain why the testimony is different?

A: At that time I remembered—

Judge: Yes?

A: (No audible response.)

Judge: Do you understand the question, sir. Let me rephrase it. Today you’re visibly upset in recalling that Chinese officials were looking for your wife during her pregnancy. You were so scared, you couldn’t even sleep at your own home. At the last hearing when you were testifying you said you had no problems during the second pregnancy. That’s very different than what you’re saying today. Can you reconcile this or can you explain why that’s so?

A: (No audible response.)

Judge: Why would you say you had no problems at the last hearing and today you’re crying about it?

A: At that time I thought it was a very simple things. I thought it was a . very simple. I shouldn’t detail it, elaborate it and because this is something very heartbroken.

Not even the questions, purporting to describe Mr. Gao’s earlier testimony that the couple had “no problems” during the second pregnancy, imply — as the IJ found — that Mr. Gao had previously testified that the couple experienced “no difficulties during her second pregnancy, had no confrontations with anyone, and had no problems with the Chinese government.” 5

Although Mr. Gao did testify at the 2002 hearing — while crying and “visibly upset” — that he had previously failed to detail or elaborate what transpired during his wife’s second pregnancy, he never admitted previously testifying that the couple had no confrontations with anyone or that they had no problems with the government during that time.

In short, there was no evidence whatsoever that Mr. Gao ever testified that they “had no difficulties during her second pregnancy, had no confrontations with anyone, and had no problems with the Chinese government. ” That finding by the IJ is a complete misreading of the record that cannot survive appellate review, however deferential. Cf. Li Zu Guan, 453 F.3d at 139-40 (concluding that IJ’s credibility determination, based on demeanor, was erroneous as a matter of law because it depended on a misstatement of the record, namely that the petitioner took “two or three water breaks during his testimony ... as an opportunity to formulate a response when confronted with a conflicting inconsistency” when he actually asked for only one water break during the course of his testimony); Chun Gao, 424 F.3d at 132 (vacating and remanding IJ’s decision that rested on adverse credibility determination based in part on misstatements of the record).

3. Hospital Stay Following Sterilization

According to the IJ, the third inconsistency that surfaced during the Gaos’ hearing related to the length of Mrs. Gao’s hospital stay following her sterilization. 6 He stated that Mrs. Gao testified that she stayed in the hospital for a week after the sterilization, whereas her husband testified that he took her home from the hospital on the same day as the sterilization. The IJ described that as a “profound inconsistency.”

In fact, the IJ was confused. The Gaos testified consistently with respect to Mrs. Gao’s hospitalization when she was sterilized. Both testified that Mr. Gao met her at the hospital outside the operating room, following the surgery. Mr. Gao testified that he brought her home that same day; Mrs. Gao was not even asked about the length of time she spent at the hospital following the sterilization. Thus, there was no inconsistency with respect to the forced sterilization, the event that is most integral to the Gaos’ claim for asylum.

As the BIA noted, the Gaos did testify inconsistently with respect to the length of Mrs. Gao’s hospital stay following her C-section. Mr. Gao testified that she stayed in the hospital for one day after the birth of their son, whereas Mrs. Gao testified that she stayed at the hospital for one week. The IJ, however, did not rely on that inconsistency in making his credibility determination. In fact, when Mrs. Gao’s attorney was questioning her with respect to the hospital stay following the C-section, the IJ interrupted, stating: “I really don’t care about that.” Rather, the IJ’s reliance on the “profound inconsistency” was based on his misunderstanding of the testimony concerning the sterilization. The IJ had questioned Mr. Gao three times on the length of his wife’s hospital stay after the sterilization. 7 He asked no questions concerning the hospital visit surrounding the C-section.

B. Lack of Substantial Evidence

The IJ gave only one reason for his credibility determination: “The Court cannot reconcile the three areas of inconsistency. They are most troubling and left unexplained. Therefore, the Court has no choice but to find that the respondent lacked credibility and therefore deny asylum, withholding.”

In applying the substantial evidence standard of review, the Court grants “particular deference” to credibility findings based on demeanor. Karaj, 462 F.3d at 116. In this case, the IJ found that the Gaos lacked credibility solely based on purported discrepancies between their 2002 testimony and Mr. Gao’s supposed testimony in 1999—testimony Mr. Gao did not admit to having given and of which there is no record. The IJ’s ruling does not mention the Gaos’ demeanor as a basis for his adverse credibility determination.

“We cannot sustain an adverse credibility finding ... that is ‘based upon ... a misstatement of the facts in the record [or] bald speculation or caprice.’ ” Li Zu Guan, 453 F.3d at 136 (citing Zhou Yun Zhang, 386 F.3d at 74) (second ellipsis and brackets in original). Here, the IJ’s adverse credibility determination is, at best, based upon either his acceptance of government counsel’s assumptions about Mr. Gao’s 1999 testimony 8 or his own, unstated recollection of that three-year-old testimony.

Although the Court will rarely disturb credibility findings, it will when, as here, they are arbitrary or not based on record evidence. See, e.g., Jin Chen, 426 F.3d at 114-15 (vacating order because adverse credibility determination based on: (1) IJ’s evaluation of testimony as “scant of details” and “lacking specificity,” and (2) IJ’s statement that two documents “appeared fabricated,” was not based on substantial evidence); Li Hua Lin v. U.S. Dep’t of Justice, 453 F.3d 99, 111 (2d Cir.2006) (vacating order based on adverse credibility determination because of IJ’s speculation concerning method of forcible sterilization); Bao Zhu Zhu v. Gonzales, 460 F.3d 426, 428 (2d Cir.2006) (vacating agency decision because IJ arbitrarily preferred statements in petitioner’s husband’s asylum application and arbitrarily used those assertions as basis for deeming her not credible); Li Zu Guan, 453 F.3d at 141 (vacating agency decision based on adverse credibility determination because, despite IJ’s correct identification of one inconsistency within the petitioner’s testimony, “no other aspect of his credibility finding [was] entirely beyond reproach”).

Here, the IJ relied on purported inconsistencies in the petitioners’ testimony in determining that the Gaos lacked credibility. Not one of the subsidiary findings, however, was based on an accurate recitation of the record evidence. Although the BIA corrected the factual misstatement with respect to the third supposed inconsistency, the IJ’s credibility determination cannot be saved as a result. 9

The decision we reach in this case is consistent with the Court’s recent decision in Felix Norbert Siewe v. Gonzales, 480 F.3d 160 (2d Cir.2007). There the Court, evaluating a challenge to an IJ’s credibility finding, applied the substantial evidence test to deny the petition for review. Rejecting the petitioner’s assertion that the IJ’s finding that his arrest warrant was inauthentic was based on bald speculation, the Court stated:

The speculation that inheres in inference is not “bald” if the inference is made available to the factfinder by record facts, or even a single fact, viewed in the light of common sense and ordinary experience. So long as an inferential leap is tethered to the evidentiary record, we will accord deference to the finding.

Id. at 168-69. Here, the IJ did not tether any of the inferences leading to his adverse credibility finding to anything in the record. Accordingly, we are left with a credibility finding based only on impermissible speculation.

In summary, the IJ’s oral decision was permeated with factual errors that resulted in legal error: his adverse credibility determination was not based on substantial evidence. Indeed, the IJ identified no sound basis in the record for his adverse credibility determination — the sole ground upon which he relied in denying the Gaos’ claim for asylum.

IV. Conclusion

Because no legitimate basis for questioning the Gaos’ credibility has been identified, we vacate the agency’s decision and remand for further proceedings consistent with this decision and without regard to the agency’s prior adverse credibility finding. See Secaida-Rosales, 331 F.3d at 313. Although we are not permitted to order the taking of additional testimony in connection with our remand, see 8 U.S.C. § 1252(a)(1), the BIA may wish to do so in an effort to arrive at a fully informed decision in this case.

1 . The administrative record includes "extensive documentation" regarding Mrs. Gao's medical history. That documentation includes several supporting documents relating specifically to Mrs. Gao's sterilization. First, a "ChangLe County Certificate of Birth Control Operation” provides that "Comrade GAO, Xiang Zhen of JhangCheng Town (Township) LongMen Village (Neighborhood Committee) was given vasectomy, tubal ligation operation” on October 12, 1995. Second, a radiology report for Mrs. Gao, dated December 12, 1998, showed: "Bilateral fallopian tubes blocked.” Third, a "Marital Birth Control Card,” sets forth "information on birth control measures enforcement,” specifying a tubal ligation dated October 1995. That document also corroborates the Gaos' testimony concerning the birth of their two children, a daughter born in July 1993 and a son born in November 1994, as well as a monetary punishment in February 1995 due to "violation of birth interval.” Finally, the record includes a photograph of Mrs. Gao’s surgical scar, and a report from Edward V. Chan, M.D., of the Mott Medical Group in New York City, confirming that his clinical findings are consistent with a tubal sterilization surgery done outside the United States.

2 . The BIA could not make its own credibility determination. See 8 C.F.R. § 1003. l(d)(3)(iv) (''[T]he Board will not engage in factfinding in the course of deciding appeals.”); see also 8 C.F.R. § 1003. 1(d)(3)© ("Facts determined by the immigration judge, including findings as to the credibility of testimony, shall be reviewed only to determine whether the findings of the immigration judge are clearly erroneous.”).

3 . In light of the IJ's complete reliance on the so-called inconsistencies in testimony, it is worth noting that the 2002 hearing was replete with translation problems. E.g., Judge: “Did you translate it as where?” Interpreter: "I translate when, Your Honor.” Judge: "Tell her you were mistaken.”; Mr. Yu: "Objection to that translation, Judge. I think the question was any contact with the family planning and the translation was were you living at the house.” ... Judge: "That’s a significant error, Mr. Lin. Do you want to listen to the question? ... Was it accurately translated?” Interpreter: "No, it was not. The question was not accurately translated, Your Honor.”

4 . It is curious that the Gaos' attorney consented to government counsel's cross-examination of Mr. Gao based on his testimony from the 1999 hearing. Nevertheless, the Gaos’ willingness to permit improper questioning does not render facts assumed in those questions evidence.

5 . This colloquy is also significant because it took six attempts (by the government lawyer and the IJ) at asking Mr. Gao why his current testimony was different from his 1999 testimony before Mr. Gao responded that he had thought that he should not detail those problems. That testimony undercuts any suggestion that Mr. Gao admitted earlier inconsistencies with respect to what occurred during Mrs. Gao's second pregnancy.

Moreover, the questions concerning “any problems” are ambiguous. When govern ment counsel — immediately after questioning Mrs. Gao about the reasons behind her request for a Caesarian section — initially raised with Mrs. Gao her husband's purported earlier testimony that she had "no problems” during the second pregnancy, even the IJ questioned why that would be inconsistent with Mrs. Gao’s testimony since "she had no problems during the second — ", Apparently, the IJ understood the attorney’s reference to "problems” during the pregnancy to relate to Mrs. Gao’s health, not problems with the government.

6 . Although, as noted above, we are not reviewing this aspect of the IJ’s decision, we discuss this purported inconsistency in order to assist the BIA and/or the IJ following remand.

7 . Perhaps in an effort to “probe for incidental details, seeking to draw out inconsistencies that would support a finding of lack of credibility,” Jin Chen, 426 F.3d at 114, the IJ questioned Mr. Gao repeatedly concerning the length of that hospitalization:

Q: After your wife was sterilized, how long did she spend in the hospital?

A: She went back home after the sterilization. I met her outside the operation room. When she saw me she cried. She rest for awhile and I brought her home.

Q: You brought her home that day?

A: Yes. Yes.

Q: She didn’t spend any time in the hospital?

A: She didn't stay in the hospital. She went back the same day.

That testimony was consistent with Mrs. Gao’s testimony concerning the sterilization at the hospital:

Q: When you were taken for sterilization, did your husband ever come to the hospital after the procedure?

A: After that procedure done, after I come out from the operation room.

8 . It is worth noting that, during the 2002 hearing, government counsel repeatedly mis-characterized Mrs. Gao’s testimony from that same hearing. For example, the government lawyer confused her testimony with respect to the C-section. The attorney stated that Mrs. Gao "testified on direct ... [that] she went [to the hospital] because the family planning cadres [sic] asked her to go and she told them she was overdue 10 to 20 days and that she asked them to schedule the delivery.” Mrs. Gao had not testified on direct that the family planning cadre asked her to go to the hospital; rather she testified that they were searching for her when she was in hiding.

Shortly thereafter, government counsel mis-characterized Mrs. Gao's testimony again: “Well earlier, ma'am you indicated that the cadres [sic] spoke to your husband at your home. Are you changing that testimony?” The IJ sustained the Gaos' attorney’s objection (“I don’t think that’s the testimony. She said that the husband learned it from a neighbor.”).

Third, government counsel stated, incorrectly: "Your aunt lives in the same village as your house with your husband.”

9 . The BIA cannot make credibility determinations any more than we can. Moreover, it is unclear whether the IJ — who "did not care” about the reasons behind Mrs. Gao's week-long hospital stay following the C-section, but questioned Mr. Gao three times concerning the duration of Mrs. Gao's hospital stay following the sterilization — would have relied upon any inconsistency regarding the C-section as a basis for finding the Gaos lacked credibility. In other words, if the IJ had realized that no "profound inconsistency” surrounding the forced sterilization actually existed, it appears likely he would have reached the opposite result.