TABLE OF CONTENTS

I.   THE STREAMLINING REGULATIONS ........................................234
     A.  Background ......................................................234
     B.  Statutory and Regulatory Scheme .................................236
     C.  Constitutional Challenges .......................................238

II.  THE AGENCY'S DENIAL OF RELIEF .......................................245
     A.  Dia's Testimony .................................................245
     B.  Burden and Standard of Review ...................................247
     C.  The Immigration Judge's Decision ................................250
         1.  Past Persecution ............................................251
         2.  Procurement of a Passport and Visa ..........................256
         3.  Future Persecution ..........................................259

III. CONCLUSION ..........................................................260

Judge Rendell filed the opinion of the Court in which Chief Judge Scirica and Judges Nygaard, Barry, Fuentes, and Smith joined amd Judges McKee, Ambro and Becker joined as to Part II. Judge Alito, joined by Judges Solviter and Roth, filed an opinion concurring as to Part I and dissenting as to Part II. Judge Stapleton filed an opinion dissenting, in which Judges McKee, Ambro and Becker joined. Judge McKee filed an opinion concurring as to Part II.

In 2001, the Immigration and Naturalization Service (INS) charged Saidou Dia, a native of the Republic of Guinea, with removability for illegal entry into the United States. Dia conceded removability but applied for relief, seeking asylum, withholding of removal, and relief under the United Nations Convention Against Torture, and Other Cruel, Inhuman or Degrading Treatment or Punishment.1 Dia alleged that he had been, and would be, persecuted in Guinea due to his actual and imputed political opinions. The Immigration Judge (IJ) rejected Dia's allegations, finding that she was "not convinced that [Dia] has suffered past persecution" or that Dia would be persecuted and/or killed if he returned to Guinea. The IJ based her conclusion solely on her determination that Dia was not credible. Dia appealed to the Board of Immigration Appeals (BIA), which summarily affirmed the IJ's decision under its streamlining regulations. This petition for review followed. Our jurisdiction arises under 8 U.S.C. § 1252.

Two issues are before the court for consideration en banc:

First, we will review whether the streamlining regulations promulgated by the Attorney General are either inconsistent with the INA, or violative of Dia's due process rights under the Fifth Amendment. See U.S. Const. amend. V.

Second, we will review the adverse credibility determination made by the Immigration Judge and summarily affirmed by the BIA.

As to the first issue, we determine that the streamlining regulations are valid.

As to the second issue, however, we conclude that the IJ's analysis of Dia's credibility was based on reasoning that was at best unexplained and at worst speculative. Accordingly, it was not supported by substantial evidence.

We will grant the petition for review, vacate the order, and remand to the BIA to give the IJ the opportunity to explain or bolster her analysis.

In upholding the IJ's determination denying Dia relief from removal, the BIA did not issue an opinion, but, instead, issued an "affirmance without opinion" (AWO) under its streamlining regulations. See 8 C.F.R. § 3.1(a)(7) (2002). The streamlining regulations have recently been the subject of many unsuccessful attacks. See, e.g., Khattak v. Ashcroft, 332 F.3d 250, 253 (4th Cir.2003) (rejecting the argument that the regulations are "impermissibly retroactive"); Albathani v. INS, 318 F.3d 365, 377 (1st Cir.2003) (rejecting a due process challenge); Capital Area Immigrants' Rights Coalition v. United States Dep't of Justice, 264 F. Supp. 2d 14, 39 (D.D.C.2003) (rejecting a challenge under the Administrative Procedure Act). Dia, with able support of amici, broadly attacks the streamlining regulations on two grounds: (1) as inconsistent with the INA; and (2) as violative of his due process rights.

The Supreme Court has "`long recognized the power to expel or exclude aliens as a fundamental sovereign attribute exercised by the Government's political departments largely immune from judicial control.'" Fiallo v. Bell, 430 U.S. 787, 792, 97 S. Ct. 1473, 52 L. Ed. 2d 50 (1977) (quoting Shaughnessy v. United States ex rel. Mezei, 345 U.S. 206, 210, 73 S. Ct. 625, 97 L. Ed. 956 (1953)). With limited exceptions, Congress, in the INA, charges the Attorney General "with the administration and enforcement of [the INA] and other laws relating to the immigration and naturalization of aliens." 8 U.S.C. § 1103(a)(1) (2002). Pursuant to this power, Congress has mandated that the Attorney General "shall establish such regulations; ... issue such instructions; and perform such other acts as he deems necessary for carrying out his authority under [the INA]." 8 U.S.C. § 1103(a)(3) (2002). Congress has further authorized that "[t]he Attorney General may provide by regulation for any other conditions or limitations on the consideration of an application for asylum not inconsistent with this Act." 8 U.S.C. § 1158(d)(5)(B) (2002).

The Attorney General has delegated to the BIA many of his responsibilities under the immigration laws, see 1 Charles Gordon, Stanley Mailman, & Stephen Yale-Loehr, Immigration Law and Procedure § 3.02[1] (rev. ed. 2003) (stating that the BIA "exercises so much of the Attorney General's authority under the immigration and nationality laws as the Attorney General may delegate to it"), and has further delegated supervision of the BIA to the Department of Justice's Executive Office of Immigration Review. Id. The BIA, established by regulation, has existed in various guises and has held various responsibilities since 1922. Id. at § 3.05[1]. Initially, immigration laws were enforced by the Secretary of Labor, under whose supervision the administrative immigration appellate body was known as the "Board of Review." Id. After Congress transferred the responsibility for immigration enforcement to the Attorney General in 1940, the Board of Review was renamed the Board of Immigration Appeals. In its present form, the BIA has been described as "a quasi-judicial body with exclusively appellate functions." Id.

The Attorney General promulgated the streamlining regulations in 1999 when the Board was faced with a crushing caseload, the number of cases having increased exponentially in a little over a decade. See Executive Office of Immigration Review: Board of Immigration Appeals Streamlining, 64 Fed. Reg. 56,135, 56,136 (Oct. 18, 1999) (to be codified at 8 C.F.R. pt. 3). See generally Board of Immigration Appeals: Procedural Reforms to Improve Case Management, 67 Fed. Reg. 54,878, 54,878-79 (Aug. 26, 2002) (to be codified at 8 C.F.R. pt. 3). Under the regulations, "the Chairman [of the BIA] may designate certain categories of cases as suitable for review" by designated Board members "who are authorized to affirm decisions of Immigration Judges ... without opinion." 8 C.F.R. § 3.1(a)(7)(i) (2002). The single BIA member to whom the case is assigned may affirm an IJ's decision in a single sentence without an opinion if he or she determines that the result was correct, and that "(A) the issue on appeal is squarely controlled by existing Board or federal court precedent and does not involve the application of precedent to a novel fact situation; or (B) the factual and legal questions raised on appeal are so insubstantial that three-Member review is not warranted." 8 C.F.R. § 3.1(a)(7)(ii) (2002).2 Each AWO is exactly the same. It reads: "The Board affirms, without opinion, the results of the decision below. The decision is, therefore, the final agency determination." See 8 C.F.R. § 3.1(a)(7)(iii) (2002); see also Executive Office of Immigration Review: Board of Immigration Appeals Streamlining, 64 Fed. Reg. at 56,137-38 ("The decision rendered below will be the final agency decision for judicial review purposes.... [T]he Immigration Judge's decision becomes the decision reviewed."). Such an order does not necessarily imply approval of all of the reasoning of the IJ's decision, but does signify that the reviewing Board member considered that any errors by the IJ were harmless or immaterial. Id. If the single BIA member decides that the decision is inappropriate for affirmance without an opinion, the case is assigned to a three-member panel for review and decision. 8 C.F.R. § 3.1(a)(7)(iv) (2002). That panel, however, is also authorized to determine that a case should be affirmed without an opinion. Id.

We are "confronted [with] questions implicating an agency's construction of the statute which it administers." INS v. Aguirre-Aguirre, 526 U.S. 415, 424, 119 S. Ct. 1439, 143 L. Ed. 2d 590 (1999) (citation and internal quotation marks omitted). For this reason, we apply the principles of deference described in Chevron U.S.A., Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 842, 104 S. Ct. 2778, 81 L. Ed. 2d 694 (1984). See Aguirre-Aguirre, 526 U.S. at 424, 119 S. Ct. 1439 ("It is clear that principles of Chevron deference are applicable to this statutory scheme."). We initially ask whether "the statute is silent or ambiguous with respect to the specific issue" we confront. Chevron, 467 U.S. at 843, 104 S. Ct. 2778. If it is, "the question for the court is whether the agency's answer is based on a permissible construction of the statute." Id. at 843, 104 S. Ct. 2778; see also Aguirre-Aguirre, 526 U.S. at 424, 119 S. Ct. 1439. In doing so, we bear in mind that "judicial deference to the Executive Branch is especially appropriate in the immigration context where officials `exercise especially sensitive political functions that implicate questions of foreign relations.'" Aguirre-Aguirre, 526 U.S. at 425, 119 S. Ct. 1439 (quoting INS v. Abudu, 485 U.S. 94, 110, 108 S. Ct. 904, 99 L. Ed. 2d 90 (1988)); see also Abdulrahman v. Ashcroft, 330 F.3d 587, 597 (3d Cir.2003) (quoting this section of Aguirre-Aguirre); Abdulai v. Ashcroft, 239 F.3d 542, 551 (3d Cir.2001) (acknowledging "the narrow scope of our review" under such circumstances).

The streamlining regulations easily pass the first step of the Chevron inquiry. The INA "is silent ... with respect to" streamlined administrative appeals. Chevron, 467 U.S. at 843, 104 S. Ct. 2778. The next question is whether the streamlining of administrative appeals "is based on a permissible construction of the statute." Id. If, as Dia contends, the streamlining regulations are inconsistent with the INA, they certainly are not based on a permissible construction of the statute. So, we must look at what the INA says regarding the BIA in particular, and administrative appeals in general.3 In so doing, we can discern nothing in the INA with which the streamlining regulations are inconsistent. See Abdulai, 239 F.3d at 555 ("[N]othing in the INA specifically requires the Board to explain its decisions."). In fact, the INA says nothing whatsoever regarding the procedures of an administrative appeal, or, for that matter, any other procedures employed by the BIA.

As Dia points out, the INA refers to the BIA in its "definitions" section, in connection with its definition of the term "order of deportation."4 8 U.S.C. § 1101(a)(47)(A) (2002). The relevant provision reads, in pertinent part:

The ["order of deportation"] shall become final upon the earlier of (i) a determination by the Board of Immigration Appeals affirming such order; or (ii) the expiration of the period in which the alien is permitted to seek review of such order by the Board of Immigration Appeals.

8 U.S.C. § 1101(a)(47)(B) (2002). Under this provision, an order of deportation is not "final" until either the BIA has passed on it, or the time for seeking BIA review has expired. The statute also provides that the statutory right to judicial review of orders of deportation is only available for a "final order." 8 U.S.C. § 1252(b)(9) (2002); Gao v. Ashcroft, 299 F.3d 266, 271 (3d Cir.2002). Based on these two provisions, Dia argues that the BIA at the relevant time was an entity provided for by statute, and no longer existed solely by regulation.

But even assuming that the BIA could not be eliminated without statutory authorization, we are hard pressed to conclude much more from the definitional statement at § 1101(a)(47)(B). It says absolutely nothing about procedures to be employed by the BIA, or the right to, or manner of, review generally; it only speaks to review by the BIA and its "affirming" the "order" of deportation. 8 U.S.C. § 1101(a)(47)(B). Based on the fact that § 1101(a)(47)(B) contains the only mention of the BIA in the INA, it seems clear that Congress has left all procedural aspects of the BIA, especially how it hears cases, entirely to the Attorney General's discretion. Id.

The statute's references to an "administrative appeal" do not alter this conclusion. Only two statutory provisions of the INA reference the term "administrative appeal." These provisions mandate that the procedure established for applying for asylum

shall provide that —

...

(iii) in the absence of exceptional circumstances, final administrative adjudication of the asylum application, not including administrative appeal, shall be completed within 180 days after the date an application is filed;

(iv) any administrative appeal shall be filed within 30 days of a decision granting or denying asylum, or within 30 days of the completion of removal proceedings before an immigration judge under section 1229a of this title, whichever is later.

8 U.S.C. § 1158(d)(5)(A)(iii) & (iv) (2002) (emphasis added). Although these provisions contemplate some type of an administrative appeal in connection with applications for asylum, they fail to provide any guidance as to the procedural trappings of that appeal.

Similarly unpersuasive is Dia's citation to 8 U.S.C. § 1229a(c)(4), which provides: If the immigration judge decides that the alien is removable and orders the alien to be removed, the judge shall inform the alien of the right to appeal that decision and of the consequences for failure to depart under the order of removal, including civil and criminal penalties.

Id. (emphasis added). Even if we were to assume that this provision contemplates that an alien will have the opportunity for an administrative appeal, neither this provision nor any other provision of the INA references the procedural requirements of an administrative appeal or outlines a scheme inconsistent with the streamlining regulations. Instead, it only speaks generally of an "administrative appeal" and "the right to appeal," and of the BIA only in the context of a "final" order. To conclude from this language in the INA that the streamlining regulations are not a "permissible construction of the statute" under Chevron, 467 U.S. at 843, 104 S. Ct. 2778, would require a sizable leap that we cannot make. The Supreme Court has forcefully emphasized that "[a]bsent constitutional constraints or extremely compelling circumstances the administrative agencies should be free to fashion their own rules of procedure and to pursue methods of inquiry capable of permitting them to discharge their multitudinous duties." Vermont Yankee Nuclear Power Corp. v. Natural Res. Def. Council, Inc., 435 U.S. 519, 543, 98 S. Ct. 1197, 55 L. Ed. 2d 460 (1978) (citation and internal quotation marks omitted). This "basic tenet of administrative law," id., has even more force in the immigration context where our deference is especially great. See Abdulai, 239 F.3d at 552 ("In light of the INA's enormously broad delegation to the Attorney General, we would be extremely reluctant to hold that his interpretation is unreasonable."); see also 1 Gordon, Mailman, & Yale-Loehr, Immigration Law and Procedure § 3.02[2] ("[T]he theory of the [INA] is that all responsibility to enforce or administer the immigration laws is vested in the Attorney General, and that she may delegate or assign any of such powers in any manner she deems appropriate."). We therefore hold that, in promulgating the streamlining regulations, the Attorney General did not run afoul of the INA.5

Dia next attacks the streamlining regulations as a deprivation of his constitutional right to due process under the Fifth Amendment.6 See U.S. Const. amend. V. We have plenary review over constitutional challenges to immigration procedures. Abdulrahman, 330 F.3d at 597. We agree with our sister courts of appeals that have passed on this issue and conclude that the streamlining regulations do not violate the Due Process Clause of the Constitution. See Denko v. INS, 2003 WL 22879815, at *8 (6th Cir. Dec.8, 2003); Falcon Carriche v. Ashcroft, 350 F.3d 845, 850 (9th Cir. 2003); Georgis v. Ashcroft, 328 F.3d 962, 967 (7th Cir.2003); Mendoza v. United States Att'y Gen., 327 F.3d 1283, 1288 (11th Cir.2003); Soadjede v. Ashcroft, 324 F.3d 830 (5th Cir.2003); Albathani, 318 F.3d at 377.

The basic elements of due process in this context are clear. Although "the Fifth Amendment entitles aliens to due process of law in deportation proceedings," Reno v. Flores, 507 U.S. 292, 306, 113 S. Ct. 1439, 123 L. Ed. 2d 1 (1993), due process is "flexible and calls for such procedural protections as the particular situation demands." Morrissey v. Brewer, 408 U.S. 471, 481, 92 S. Ct. 2593, 33 L. Ed. 2d 484 (1972); see also Marincas v. Lewis, 92 F.3d 195, 203 (3d Cir.1996) ("Precisely what minimum procedures are due under a statutory right depends on the circumstances of the particular situation."). The due process afforded aliens stems from those statutory rights granted by Congress and the principle that "[m]inimum due process rights attach to statutory rights." Marincas, 92 F.3d at 203; see also Meachum v. Fano, 427 U.S. 215, 226, 96 S. Ct. 2532, 49 L. Ed. 2d 451 (1976). Our concern, then, is whether the streamlining regulations afford aliens such as Dia their minimum due process rights. See Albathani, 318 F.3d at 375 (stating that an unadmitted alien present in the United States has only "limited" due process rights); see also Anwar v. INS, 116 F.3d 140, 144 (5th Cir.1997) ("Due process challenges to deportation proceedings require an initial showing of substantial prejudice."). In making this assessment, we look to see if the process at issue fits with the notion that "[t]he fundamental requirement of due process is the opportunity to be heard at a meaningful time and in a meaningful manner." Mathews v. Eldridge, 424 U.S. 319, 333, 96 S. Ct. 893, 47 L. Ed. 2d 18 (1976) (citation and internal quotation marks omitted) (emphasis added).

Dia bases his primary due process argument on statements we made in Abdulai, where we expounded on this "`fundamental requirement of due process'" articulated in Mathews. Abdulai, 239 F.3d at 549 (quoting Mathews, 424 U.S. at 333, 96 S. Ct. 893). In Abdulai, the BIA had issued a two-page per curiam opinion that contained a "terse" application of Board precedent to the specific facts of Abdulai's case. Id. at 547. Abdulai argued, inter alia, that in so doing, "the BIA denied him due process by failing to make an individualized determination of his interests." Id. at 549.

We began our analysis of Abdulai's due process argument by noting that, in the context of the adjudication of claims for relief from removal such as the one before us, due process "requires three things." Id. at 555. "An alien: (1) is entitled to `factfinding based on a record produced before the decisionmaker and disclosed to' him or her; (2) must be allowed to make arguments on his or her own behalf; and (3) has the right to `an individualized determination of his [or her] interests.'" Id. (quoting Llana-Castellon v. INS, 16 F.3d 1093, 1096 (10th Cir.1994) (citation omitted)). Of these three requirements, Abdulai dealt solely with the third requirement, an "individualized determination." In turn, Dia contends that, by issuing an AWO pursuant to the streamlining regulations, the BIA deprived him of his due process right to an "individualized determination" of his interest as that right was recognized in Abdulai.

Although we ultimately concluded in Abdulai that there was no due process violation because we found that the BIA had in fact made an "individualized determination" of Abdulai's application, id. at 550, Dia seizes on what we said as we reasoned toward that conclusion. One such comment was that we had previously "suggested that the BIA denies due process to an alien when it `act[s] as a mere rubberstamp.'" Id. (quoting Marincas, 92 F.3d at 202 n. 7). We also noted that "the question for due process ... is simply whether the Board made an individualized determination of Abdulai's interests." Id. (emphasis added). Dia argues that these observations require that we invalidate the streamlining regulations.

Dia, however, takes our statements in Abdulai out of context. We made those statements in connection with Abdulai's argument that the BIA had not "acknowledg[ed] or address[ed] any of his arguments." Id. at 549. In Abdulai, we necessarily reviewed the BIA's opinion, because the BIA had issued an opinion and the petitioner focused his arguments on that opinion. Id. at 548. We therefore made the statements in Abdulai in the context of a situation in which the BIA had chosen to speak — thus forcing the reviewing court to examine the BIA's reasoning — but had done so in a way that caused us to question whether the BIA had carefully reviewed the specific matter before it. See Mikhael v. INS, 115 F.3d 299, 302 (5th Cir.1997) ("We have authority to review only an order of the BIA, not the IJ, unless the IJ's decision has some impact on the BIA's decision."). The situation here is very different; the BIA did not opine on its own, but, instead, referred us to the IJ's decision.

Contrary to Dia's suggestion, in Abdulai we did not impose a requirement that in all instances the BIA must indicate that it made an individualized determination of the claim for relief. In fact, we noted our approval of decisions of other courts of appeals that have upheld the BIA's right to "`simply state that it affirms the IJ's decision for the reasons set forth in that decision.'" Abdulai, 239 F.3d at 549 n. 2 (quoting Chen v. INS, 87 F.3d 5, 7 (1st Cir.1996)). We also made clear that "[t]here are some situations in which a court of appeals effectively reviews an IJ's decision, but [that Abdulai's was] not one of them." Id. One of those situations arises, we noted, when the BIA "defers" to the IJ. Id. In that situation, "a reviewing court must, as a matter of logic, review the IJ's decision to assess whether the BIA's decision to defer was appropriate." Id.; see also Abdulrahman, 330 F.3d at 591. And so, here, where the BIA directs us to the opinion and decision of the IJ who originally assessed Dia's application, we review the IJ's opinion.

Dia, nonetheless, also insists that the streamlining regulations violate his right to an "individualized determination" because they specifically state that an AWO does not necessarily imply approval of all of the reasoning of the IJ's decision. See 8 C.F.R. § 3.1(a)(7)(iii) (2002). But he fails to articulate why or how this is so. We are unaware of any requirement, let alone any constitutional requirement, that an agency adjudicator must commit to writing or otherwise verbalize his or her reasoning, where, as here, the agency has directed us to an opinion for review. In Dia's case, the due process right to an "individualized determination" was accorded to Dia at the IJ level, where the IJ "reasoned" her decision, and the BIA gave the result its imprimatur pursuant to its regulations. Certainly, the BIA could have articulated its reasons for affirming the IJ's order, but just because it had the power to do so, does not mean the Constitution required it to exercise that power.7 See Abdulai, 239 F.3d at 549 n. 3 ("Having the power to do something and being required to do it are not the same thing.").

Equally unavailing is amici's argument that "fundamental rule[s] of administrative law" enunciated by the Supreme Court in SEC v. Chenery Corp., 332 U.S. 194, 196-97, 67 S. Ct. 1575, 91 L. Ed. 1995 (1947), support Dia's argument that the BIA's failure to adopt the reasoning of the IJ, in accordance with the streamlining regulations, violated his constitutional right to due process. In fact, we believe that Chenery actually supports the opposite conclusion. In Chenery, the Court emphasized a "simple but fundamental rule of administrative law ... that a reviewing court, in dealing with a determination or judgment which an administrative agency alone is authorized to make, must judge the propriety of such action solely by the grounds invoked by the agency." Id. at 196, 67 S. Ct. 1575 (emphasis added). The "corollary" of this rule is that the basis of an administrative action "must be set forth with such clarity as to be understandable." Id. The Court therein was concerned with ensuring that a reviewing court may "test" administrative action. Id. Under the streamlining regulations, this requirement is met. The BIA clearly "invokes" the IJ's opinion as the grounds on which the agency's decision rests; we thus "judge the propriety" of the IJ's action in order to "test" the agency's action. As the Court of Appeals for the First Circuit said:

[Petitioner and Amici] both overlook the plain language of Chenery, which refers to agencies in their entirety, not individual components of agencies. Here, the relevant agency — the INS — has presented a statement of reasons for its decision, albeit from the IJ rather than the BIA. Chenery does not require that this statement come from the BIA rather than the IJ.

Albathani, 318 F.3d at 377; see also Nagi El Moraghy v. Ashcroft, 331 F.3d 195, 206 (1st Cir.2003) ("The provision of reasons in the IJ's opinion satisfies the requirement in SEC v. Chenery Corp., 332 U.S. 194, 196-97, 67 S. Ct. 1575, 91 L. Ed. 1995 (1947), that administrative agencies set forth with clarity the basis for their decisions, and the AWO procedure did not prevent there being meaningful review." (citation omitted)); Dominguez v. Ashcroft, 336 F.3d 678, 680 (8th Cir.2003) (stating that "the opinion of the immigration judge is sufficient to satisfy th[e] requirement" in Chenery Corp. that "an agency must set out the basis of its decision").

Dia asserts three other ways in which the AWO violated his due process right: it denied him "meaningful review"; it prevented our court from providing meaningful review; and it was not "fair." Although Dia does not match these due process arguments with any of the three-requirements for due process we outlined in Abdulai, they appear to be variations on his theme that the issuance of an AWO denied him of his right to an "individualized determination." Regardless of their label, we reject these contentions as well.

Dia's claim that the AWO denied him his so-called "due process right to meaningful review" lacks substance. Dia specifically maintains that he has the right to meaningful review by the BIA. Other than pointing generally to the Due Process Clause, however, Dia does not identify the source of this alleged right. We are unaware of any authority supporting a due process right to "meaningful review" by an administrative appellate body.

The "right to meaningful review" that Dia alleges is clearly distinguished from "[t]he fundamental requirement of due process [that] is the opportunity to be heard at a meaningful time and in a meaningful manner," Mathews, 424 U.S. at 333, 96 S. Ct. 893, that we discussed above. The "meaningfulness" requirement of Mathews pertains to "the opportunity to be heard" and the "manner" in which one is heard, not to a review by an administrative appellate body. Id. Moreover, any recognized right to "meaningful review," as we note more fully in the margin, has been confined to the context of review by federal courts, and not extended to review by an administrative appellate body.8 See also, e.g., Awolesi v. Ashcroft, 341 F.3d 227, 232 (3d Cir.2003) ("In order for us to be able to give meaningful review to the BIA's decision, we must have some insight into its reasoning."); Kayembe v. Ashcroft, 334 F.3d 231, 238 (3d Cir.2003) ("When deficiencies in the BIA's decision make it impossible for us to meaningfully review its decision, we must vacate that decision and remand so that the BIA can further explain its reasoning."). Quite clearly, "[a]n alien has no constitutional right to any administrative appeal at all," Albathani, 318 F.3d at 376; see also Guentchev v. INS, 77 F.3d 1036, 1037 (7th Cir.1996), and, therefore, no constitutional right to a "meaningful" administrative appeal.

Nor are we persuaded by Dia's related argument that the streamlining regulations—or, their "opaque" nature, as amici describe them—prevent us as a court of appeals from engaging in a meaningful review of the agency's actions. See Simmons v. Beyer, 44 F.3d 1160, 1169 (3d Cir.1995) (stating that due process requires that a guaranteed "appellate procedure must furnish the components necessary for meaningful review"). We cannot agree with amici's claim that "the summary affirmance process impermissibly strips the federal courts of the ability to properly review critical agency action." The streamlining regulations in no way restrict our ability to review the agency's denial of relief from removal. An agency, not a particular administrative appellate body, must set forth the basis for its order with sufficient specificity to permit meaningful review by this court. See Albathani, 318 F.3d at 377. Here, as we discuss in the next section, we have no doubt as to the basis for the agency's decision as put forth for review by the BIA. The BIA presents for our review the reasoning and decision of the IJ as that of the Attorney General. See Executive Office of Immigration Review: Board of Immigration Appeals Streamlining, 64 Fed. Reg. at 56,137-38.

All that is required for our meaningful review is that the agency — as represented by an opinion of the BIA or IJ— put forth a sufficiently reasoned opinion. See Mendoza, 327 F.3d at 1289 ("[T]he meaningful review of the INS's removability determination is not precluded by the brevity of the BIA's summary affirmance decision because an appellate court `will continue to have the IJ's decision and the record upon which it is based available for review.'" (quoting Albathani, 318 F.3d at 377)). While in many instances knowing the BIA's reasoning might prove helpful to our review, the BIA's failure to express it does not amount to a constitutional violation. Neither the Constitution nor Congress guarantee a de novo review by the BIA, Abdulai, 239 F.3d at 549 n. 3, nor do they guarantee a right to a fully reasoned opinion by the BIA. And, as we have noted, we see no constitutional significance in the fact that an AWO does not necessarily imply approval of all of the reasoning of the IJ. We are able to meaningfully review the final determination of the agency, and, in this context, that is all that due process requires.

We are similarly unmoved by Dia's argument that the streamlining regulations violate basic due process requirements of "fairness." See Bridges v. Wixon, 326 U.S. 135, 154, 65 S. Ct. 1443, 89 L. Ed. 2103 (1945) ("Meticulous care must be exercised lest the procedure by which he is deprived of that liberty not meet the essential standards of fairness."). We have made clear that "[w]hen Congress directs an agency to establish a procedure,... it can be assumed that Congress intends that procedure to be a fair one." Marincas, 92 F.3d at 203. What is "fair" within the context of immigration proceedings, however, need not always measure up to the requirements of fairness in other contexts, especially because "[a]liens only have those statutory rights granted by Congress." Id.; see also Mathews v. Diaz, 426 U.S. 67, 79-80, 96 S. Ct. 1883, 48 L. Ed. 2d 478 (1976) ("In the exercise of its broad power over naturalization and immigration, Congress regularly makes rules that would be unacceptable if applied to citizens.").

We find nothing "unfair" in a constitutional sense about the INS's streamlining procedures. An applicant retains a full and fair opportunity to make his case to the IJ, and has a right to review of that decision by the BIA, and then by a court of appeals. See Guentchev, 77 F.3d at 1038 ("The combination of a reasoned decision by an administrative law judge plus review in a United States Court of Appeals satisfies constitutional requirements."); cf. Zubeda, 333 F.3d at 480 ("Justice requires that an applicant for asylum or withholding of deportation be afforded a meaningful opportunity to establish his or her claim."); Abdulrahman, 330 F.3d at 596 (stating that an alien threatened with deportation has a right to a "full and fair hearing"). The fact that the review is done by one member of the BIA and that the decision is not accompanied by a fully reasoned BIA decision may be less desirable from the petitioner's point of view, but it does not make the process constitutionally "unfair." Neither Dia nor amici has provided any reason for us to conclude otherwise.

Our dissenting colleagues who disagree on this point would strike down the regulations, contending that they alter an established administrative scheme under the INA. However, upon further scrutiny, it becomes clear that Judge Stapleton is not really taking issue with the regulations as a perversion of the statute or even agency practice, but rather as a violation of principles of judicial review that we have espoused in our case law.9 Admittedly, the regulations will cause us to review cases affirmed by the Board without opinion. But, they do not force us to venture "through the looking glass" (like Alice in Wonderland), because we have the IJ's reasoning and the record necessary to exercise our function of review. We have always required that the review process be a meaningful one, aided by a reasoned opinion from the agency.10 We do not today cast that principle aside. Rather, we hold that when the issue before us is the validity of an agency's regulations establishing its procedures, unless they violate Congressional dictates or give rise to a due process violation, the regulations must stand, especially where, as here, Congress has specifically delegated the power to establish procedures by regulation.11

Furthermore, Judge Stapleton's dissent focuses on a perceived "inability to review" in the abstract, doing a disservice to, and seemingly ignoring, what we have done and will continue to do in reviewing, in a meaningful manner, the cases that come to us from the BIA. When, after scouring the record, we are still unable to determine the agency's reasoning, we have remanded to the BIA for further explanation.12 In actuality, these regulations do not prevent us from adhering to the very principles that Judge Stapleton's dissent contends are being abandoned. Today we face no such "inability to review" in this case, as we can clearly review, and are reviewing, what the agency did without the aid of the BIA's particular take on the matter.13 Thus, there is no basis for a blanket declaration of invalidity, or a declaration that the regulations are invalid as applied here.

Accordingly, we hold that the Attorney General's implementation of the streamlining regulations and the BIA's issuance of an AWO in this case did not violate either the INA or the Constitution.

We now turn to Dia's substantive attack on the Attorney General's denial of his claim for relief from removal. As outlined above, when the BIA issues an AWO under the streamlining regulations, we review the IJ's opinion and scrutinize its reasoning. Because Dia's credibility was the basis on which the IJ rested her decision to deny relief, the sole issue before us is that credibility determination. In addressing this issue, we first will detail Dia's testimony before the IJ, augmenting it with details from the administrative record. We then will discuss our standard of review. Lastly, we will apply that standard of review to the IJ's opinion, explaining why we must vacate it.

Dia, an ethnic Fula, was born in Selouma, Dinguiraye, Guinea. He has had only two years of education, has a limited ability to read and write, and does not speak English. He joined the Rassemblement du Peuple de Guineé ("Rally of the People of Guinea Party" or "RPG") in 1998, at the age of twenty-two. His father had been a member of the RPG before his death in 1997. Dia worked in the field rallying support for RPG's imprisoned leader, Alpha Conde — a member of parliament and a candidate in the 1998 presidential election — and monitored voting polls to help prevent election fraud. The Country Report for Guinea assembled by the U.S. Department of State — included in the administrative record — reveals a country in a state of turmoil. Guinea's political system appears "deeply flawed" and its human rights record even worse. See U.S. Department of State, 2000 Country Reports on Human Rights Practices: Guinea (Feb. 2001) (hereinafter "Country Report"). The Country Report states:

The Government's human rights record was poor; although there were some improvements in a few areas serious problems remained in others. The Government's tight and sometimes partisan control of the electoral process both in the 1998 presidential election and the deeply flawed June municipal elections; its refusal to create an independent electoral oversight mechanism; and its prohibition of nongovernmental broadcast media, effectively restricted citizens' right to change their government. Major human rights abuses include: Extrajudicial killings; disappearances; use of torture, beatings and rape by police and military personnel; and police abuse of prisoners and detainees. Soldiers, police, and civilian militia groups killed, beat, and raped citizens, as well as refugees from Sierra Leone and Liberia. Security forces used arbitrary arrest and detention. Members of the security forces committed abuses with impunity.... Violence and societal discrimination against women, prostitution of young girls, female genital mutilation (FGM), ethnic discrimination and interethnic violence, child labor, reports of trafficking of women and children, and vigilante actions by victims or others persisted.

Id. This report is important because the picture it paints provides a background against which to assess Dia's credibility. See Nagi El Moraghy, 331 F.3d at 204 (stating that State Department reports "provide a context for assessing the credibility of a petitioner's case .... depending on whether or not they corroborate the petitioner's tale"); see also Zubeda, 333 F.3d at 477 ("Official as well as unofficial country reports are probative evidence and can, by themselves, provide sufficient proof to sustain an alien's burden under the INA.").

Dia testified that the problems that directly affected him began on November 22, 2000. At that time, he was approached in his home in Nzerekore, Guinea, by a man named Bangora, who was the chief of his neighborhood, and two other men, who asked him to join the Guinean military to fight Liberian and Sierra Leonean rebels fighting in Guinea. When Dia refused to join the military, Bangora and the other men accused Dia of sympathizing with the rebels. Dia testified that he refused to join the military because members of the military had killed his father and he feared they wanted to kill him as well.14 He also testified that the three men associated the RPG with the rebels, knew that Dia belonged to the RPG, and "wanted to create some problems for [him] so that they could accuse [him] of something." He told the IJ that he believed that Bangora and the two men knew he would not join them because members of the RPG, such as Dia, oppose the government and, thus, refuse to give it aid.

Later that day, Dia, concerned for his welfare, went to his uncle's home outside of town to seek advice. Apparently, his uncle was not at his home and Dia waited three days until his uncle finally returned. After the two consulted, Dia's uncle agreed to return with Dia to town to talk to Bangora. When they arrived, Dia discovered his home burned to the ground.15 Eventually, he found his wife, who was bruised, and daughter at his in-laws' home. In response to questions about her bruises, Dia's wife told him that, on November 24, about twenty-five military men had come to their home searching for Dia, and, upon finding that Dia was not home and hearing Dia's wife's claim not to know where Dia was, the men beat and raped her and burned the house. The men told his wife that Dia was aiding the rebels so that Conde could be released. After consulting with his wife who pled with him to flee the country, Dia decided not to talk with Bangora and to flee from the village, leaving her and their child behind.

For four months, Dia remained in Guinea, living at the home of his friend, Abdoulaye Sow.16 Sow eventually made arrangements for Dia to secure a new Guinean passport and a U.S. visa with his "friend who was a person who ma[de] arrangement[s] for people who want to travel." Dia gave the man his old passport and six photographs. Dia did not learn either the man's name or how the man procured the documents. The man told Dia to tell U.S. immigration officials that he went to work in Italy as a "tomato picker" and was on his way to Honduras to work on a ship. The man gave Dia a new Guinean passport, a U.S. visa, "a letter packet" ostensibly supporting the story that he worked in Italy, and an airline ticket. Dia paid the man 2 million Guinean francs (about $1000 U.S.). Dia also secured the services of a Guinean policeman to help him pass through the police roadblocks. Dia paid that officer 300,000 Guinean francs (about $150 U.S.).

Upon his arrival in the U.S., Dia attempted to enter the country using the story that the smuggler recommended. The INS official, noticing that Dia did not have a ticket to go to Honduras, did not believe his story. The INS then sought to remove Dia. Dia conceded removability, but sought relief from deportation based on asylum, withholding of removal, and relief under the CAT. Dia was twenty-six years old at the time of his hearing before the IJ. Represented by counsel, and speaking in Fulani through an interpreter, he testified at the hearing and called a handwriting expert to support his story that the visa and passport did not contain his handwriting. The IJ issued a written opinion, denying relief because she found that Dia was not credible. As mentioned, the BIA, through a single member of the Board, affirmed without opinion pursuant to 8 C.F.R. § 3.1(a)(7) (2002).

An alien has the burden of supporting his claim for relief from removal. An alien's credibility, by itself, may satisfy his burden, or doom his claim. Gao, 299 F.3d at 272 ("Aliens have the burden of supporting their asylum claims through credible testimony. Testimony, by itself, is sufficient to meet this burden, if `credible.'" (quoting 8 C.F.R. § 208.13(a)) (citation omitted)); see also Mulanga v. Ashcroft, 349 F.3d 123, 132-34 (3d Cir.2003) (stating that an applicant's credible testimony "may be sufficient to sustain the burden of proof without corroboration"). The IJ here concluded that Dia was not credible based on "the inconsistencies in Dia's testimony and its overall implausibility." This adverse credibility determination — fatal to Dia's claim — was a finding of fact. See Gao, 299 F.3d at 272; see also Mulanga, 349 F.3d at 131-32; Secaida-Rosales v. INS, 331 F.3d 297, 307 (2d Cir.2003) ("Generally, courts have treated credibility questions in deportation proceedings as questions of fact....").

We review the agency's findings of fact under the standard found in the Illegal Immigration Reform and Immigrant Responsibility Act of 1996, Pub. L. No. 104-208, div. C, 110 Stat. 3009 (enacted April 1, 1997) (IIRIRA or "Reform and Responsibility Act"), which provides:

[T]he administrative findings of fact are conclusive unless any reasonable adjudicator would be compelled to conclude to the contrary.

8 U.S.C. § 1252(b)(4)(B).

Since the enactment of the Reform and Responsibility Act, various courts of appeals, including our court, have read this standard to require that the agency support its findings with substantial evidence, as articulated by the Supreme Court in INS v. Elias-Zacarias, 502 U.S. 478, 481-84, 112 S. Ct. 812, 117 L. Ed. 2d 38 (1992).17 There, the Court framed the standard as follows:

The BIA's determination that Elias-Zacarias was not eligible for asylum must be upheld if "supported by reasonable, substantial, and probative evidence on the record considered as a whole." 8 U.S.C. § 1105a(a)(4). It can be reversed only if the evidence presented by Elias-Zacarias was such that a reasonable factfinder would have to conclude that the requisite fear of persecution existed. NLRB v. Columbian Enameling & Stamping Co., 306 U.S. 292, 300, 59 S. Ct. 501, 83 L. Ed. 660 (1939).

Id. at 481, 112 S. Ct. 812. And, in the case relied upon by the Court in Elias-Zacarias for that principle, the Court stated:

Substantial evidence is more than a scintilla, and must do more than create a suspicion of the existence of the fact to be established. "It means such relevant evidence as a reasonable mind might accept as adequate to support a conclusion,"... and it must be enough to justify, if the trial were to a jury, a refusal to direct a verdict when the conclusion sought to be drawn from it is one of fact for the jury.

Columbian Enameling & Stamping Co., 306 U.S. at 300, 59 S. Ct. 501 (citation omitted).

Our court has explicitly stated that "[t]he Reform and Responsibility Act codifies the language the Supreme Court used in Elias-Zacarias to describe the substantial evidence standard in immigration cases."18 Sevoian v. Ashcroft, 290 F.3d 166, 171 (3d Cir.2002). The substantial evidence standard has historically been, and continues to be, the standard governing the relationship between administrative agencies and courts of review.19

The application of the substantial evidence standard is well-established. See, e.g., Allentown Mack Sales & Serv., Inc. v. NLRB, 522 U.S. 359, 366-67, 118 S. Ct. 818, 139 L. Ed. 2d 797 (1998) (indicating that the substantial evidence test requires court to decide "whether on this record it would have been possible for a reasonable jury to reach the Board's conclusion"); FTC v. Indiana Fed'n of Dentists, 476 U.S. 447, 455, 106 S. Ct. 2009, 90 L. Ed. 2d 445 (1986) (noting that the substantial evidence test requires court to "accept Commission's findings of fact if they are supported by `such relevant evidence as a reasonable mind might accept as adequate to support a conclusion'"). Thus, the question whether an agency determination is supported by substantial evidence is the same as the question whether a reasonable fact finder could make such a determination based upon the administrative record. If a reasonable fact finder could make a particular finding on the administrative record, then the finding is supported by substantial evidence. Conversely, if no reasonable fact finder could make that finding on the administrative record, the finding is not supported by substantial evidence.

Thus, where we review an IJ's credibility determination, we must ask whether the determination is supported by evidence that a reasonable mind would find adequate. We look at an adverse credibility determination to ensure that it was "appropriately based on inconsistent statements, contradictory evidences, and inherently improbable testimony ... in view of the background evidence on country conditions." In re S-M-J- (Interim Decision), 21 I. & N. Dec. 722, 1997 WL 80984 (BIA 1997). Where an IJ bases an adverse credibility determination in part on "implausibility" as the IJ did here, such a conclusion will be properly grounded in the record only if it is made against the background of the general country conditions. See Gao, 299 F.3d at 278-79; see also He v. Ashcroft, 328 F.3d 593, 603 (9th Cir. 2003).

Therefore, "[w]hile we defer to the IJ on credibility questions, that deference is expressly conditioned on support in the record," Nagi El Moraghy, 331 F.3d at 205, and "[d]eference is not due where findings and conclusions are based on inferences or presumptions that are not reasonably grounded in the record." Id. at 202 (citation and internal quotation marks omitted); see also Abdulrahman, 330 F.3d at 597 (stating that "substantial deference" to a finding is to be "afforded ... where it is grounded in evidence in the record"). To this end, it is clear that "[a]dverse credibility determinations based on speculation or conjecture, rather than on evidence in the record, are reversible," Gao, 299 F.3d at 272, and that an IJ must support her adverse credibility findings with "specific[,] cogent reasons." Id. at 276; Abdulrahman, 330 F.3d at 597; see also Secaida-Rosales, 331 F.3d at 307 ("When an IJ rejects an applicant's testimony, the IJ must provide `specific, cogent' reasons for doing so."); He, 328 F.3d at 595 ("[T]he IJ and BIA must offer a `specific, cogent reason for any stated disbelief.'" (quoting Hartooni v. INS, 21 F.3d 336, 342 (9th Cir.1994))).

If the IJ's conclusion is not based on a specific, cogent reason, but, instead, is based on speculation, conjecture, or an otherwise unsupported personal opinion, we will not uphold it because it will not have been supported by such relevant evidence as a reasonable mind would find adequate. In other words, it will not have been supported by substantial evidence. Guided by this examination of our principles of deference regarding an IJ's credibility determination, we conclude that the IJ's determination here fails this test.

The IJ rejected numerous aspects of Dia's testimony, as well as the entire testimony of Dia's expert handwriting witness. For purposes of our analysis, we will divide the testimony rejected by the IJ into three areas: past persecution, the circumstances surrounding Dia's procurement of a passport and visa, and future persecution. Dia insists that the IJ's findings of fact underpinning the adverse credibility determination as to each of these categories of testimony are not supported by evidence in the record, and are not otherwise sufficient to support the conclusions reached by the IJ.20 We must, therefore, examine the IJ's analysis and reasons put forth in order to determine whether she based the adverse credibility determination on substantial evidence. Doing so, we find that the IJ's conclusions do not flow in a reasoned way from the evidence of record and are, at times, arbitrary and conjectural in nature. Repeatedly, we are left wondering how the IJ reached the conclusions she has drawn. Her opinion consists not of the normal drawing of intuitive inferences from a set of facts, but, rather, of a progression of flawed sound bites that gives the impression that she was looking for ways to find fault with Dia's testimony.

Accordingly, we find that the IJ's adverse credibility determination is not supported by substantial evidence in the administrative record, and we will remand for the IJ to either state, or seek, the necessary foundation to augment her opinion.

Before reviewing the specific aspects of the IJ's ruling that we find troubling, however, it is appropriate that we note our agreement with the view expressed by Judge Alito that asylum cases are difficult ones — for us, as well as for immigration judges. And, as Judge Alito points out, some leeway must be given to the administrative arbiters to draw inferences based on common sense and logic as well as on personal experience and background knowledge gained from exposure to certain situations.

However, perhaps because of the difficult nature of these types of cases, and the critical importance of resolving them properly — for the stakes are very high indeed — the soundness of the basis of the decision making, even if experiential or logical in nature, must be apparent. The process of drawing inferences cannot be left to whim, but must withstand scrutiny.21

Here, we are presented with a unique setting in which, as we will catalog, the inferences drawn and conclusions reached are in some instances non sequiturs, and in others, counterintuitive. The flow of the reasoning process appears to break down as the IJ, repeatedly, draws an unreasonable conclusion from a fact susceptible to differing interpretations. Numerous such instances do not, as the dissent suggests, add up to a totality of circumstances that supports a finding that Dia's testimony was not credible. Rather, they are an aggregation of empty rationales that devolve into an unsupported finding of adverse credibility. Moreover, rather than standing our standard "on its head," as the dissent suggests, our appropriate insistence on "substantial evidence" upholds that standard by requiring that there be a sound basis — whether supplied by the record evidence or by background knowledge — to support the IJ's findings.22

Here, the conclusions of the IJ are more puzzling than plausible, more curious than commonsense. Judge Alito suggests that if we refuse to defer to the IJ's reasoning here, we would gut the substantial evidence standard. To the contrary, we suggest that to require sound reasoning breathes life into that standard.

We do not, as Judge Alito implies, conclude that the IJ was bound to find Dia credible. Rather, we recognize the possibility that the IJ's conclusions might ultimately be the correct ones. However, we cannot affirm the IJ's findings and conclusions on the record presented to us, as the reasons she does provide in support of her decision do not logically flow from the facts she considered. Accordingly, we conclude that the best course is to remand for further explanation by the IJ as to the basis for her various conclusions.

We first address the various parts of Dia's testimony regarding past persecution that the IJ rejected. We begin with a statement made by the IJ that foretells many of the errors that also infect other parts of her opinion: the IJ's rejection of Dia's testimony that, in her words, "members of the Guinean police are actively looking for him." The IJ rejected this testimony as not credible for two reasons: "this conclusion ... is, in fact, contrary to the evidence in the Record of Proceedings"; and "this conclusion [was] not supported by any documentation in the Record of Proceedings." On examination, these reasons for rejecting Dia's testimony are patently inadequate.

The IJ's conclusion that Dia's testimony that "members of the Guinean police are actively looking for him ... [was] contrary to the evidence in the Record of Proceedings" is based on a misreading of Dia's testimony. Dia did not testify that the police were after him, but, rather that the military was after him. Even assuming that the IJ concluded that the police in Guinea necessarily would be looking for a man wanted by the military, there is nothing in the record, nor any reasoning set forth by the IJ in her opinion, to support that conclusion. In fact, the Country Report depicts a country where the military, as well as the civilian militias, act independently from the formal government.

Even more troubling is the fact that, considering the testimony as the IJ presented it — i.e., that the police were after Dia — the IJ's rejection of this testimony still is not explained, nor does it have any basis in the record. The IJ stated, rhetorically, that she "question[ed] how [Dia] was able to procure a police stamp if he was actively being sought by the police." But Dia explained how. He testified that a policeman helped him bypass the police and that Sow's friend had procured the police stamp. The IJ dismissed this testimony as "unconvincing," but failed to say why or point to any evidence that contradicted this testimony. Absent a reason such as implausibility or inconsistency based in the record, or that Dia's demeanor in some way led her to question his veracity, the IJ should not have summarily dismissed Dia's testimony on this point.23

We are perplexed by the IJ's rejection of Dia's explanation that a Guinean policeman helped him cross the police border for 300,000 Guinean francs (about $150). The IJ stated that she "question[ed] why this policeman would risk his reputation, not to mention, his life, to assist the respondent, a wanted political opponent, evade detection by the police ... for the equivalent of $150 United States Dollars." This conclusion is not explained, and appears to be pure conjecture. It is not only not based on the record, but, in fact, it contravenes key parts of it. The Country Report confirms that Guinean police extort money from citizens at road blocks and that corruption at road checkpoints is widespread and "systematic." In addition, figures contained in the record show that $150 U.S. is nearly a quarter of the per capita GDP in Guinea for 1999, a sum likely tempting to a policeman in a poor country replete with corruption within its police force.

As for the IJ's reference to a lack of "supporting documentation" in the record that "members of the Guinean police are actively looking for" Dia, the IJ failed to explain what type of "documentation in the Record" she expected or required. We cannot imagine how Dia could have provided documentary support for the fact that the military (or the police as the IJ stated) was after him. At most, an applicant must provide corroborating evidence only when it would be reasonably expected. See In re S-M-J- (Interim Decision), 21 I. & N. Dec. 722, 1997 WL 80984 (BIA 1997). As we have cautioned:

It is obvious that one who escapes persecution in his or her own land will rarely be in a position to bring documentary evidence or other kinds of corroboration to support a subsequent claim for asylum. It is equally obvious that one who flees torture at home will rarely have the foresight or means to do so in a manner that will enhance the chance of prevailing in a subsequent court battle in a foreign land. Common sense establishes that it is escape and flight, not litigation and corroboration, that is foremost in the mind of an alien who comes to these shores fleeing detention, torture and persecution. Accordingly, corroboration is not required to establish credibility. The law allows one seeking refugee status to "prove his persecution claim with his own testimony if it is credible."

Senathirajah v. INS, 157 F.3d 210, 215-16 (3d Cir.1998) (quoting Mosa v. Rogers, 89 F.3d 601, 604 (9th Cir.1996)). Dia was not in a position to corroborate his testimony in this regard. See Qiu v. Ashcroft, 329 F.3d 140, 153-54 (2d Cir.2003) ("Unless the BIA anchors its demands for corroboration to evidence which indicates what the petitioner can reasonably be expected to provide, there is a serious risk that unreasonable demands will inadvertently be made.... What is (subjectively) natural to demand may not ... be (objectively) reasonable.").

In any event, the IJ failed to acknowledge, let alone adhere to, the parameters that we have adopted regarding such corroboration. In Abdulai, we recognized that, under certain circumstances, the BIA may require corroboration, and we found the three-part inquiry that the BIA has developed in this respect to be consistent with the INA. According to this inquiry, we require the following from an IJ: "(1) an identification of the facts for which `it is reasonable to expect corroboration;' (2) an inquiry as to whether the applicant has provided information corroborating the relevant facts; and, if he or she has not, (3) an analysis of whether the applicant has adequately explained his or her failure to do so." Abdulai, 239 F.3d at 554 (quoting In re S-M-J-, 21 I. & N. at 725). Here, the IJ failed to analyze whether Dia adequately explained his failure to present corroborating evidence.

In fact, it appears that the IJ actively discouraged — if she did not indeed prohibit — Dia from presenting such evidence. At one of the hearings, Dia's counsel told the IJ that Dia had been "trying to contact friends or family to substantiate his case. And over the weekend, a friend of his faxed a letter to [counsel's] office, which [counsel] had only obtained on Monday and [] had to get translated." The IJ then marked it for "identification purposes only." At the end of that hearing, the IJ made clear that she was skeptical of Dia's claims and "need[ed] more information." But, when Dia's counsel stated that he understood and noted that he was "somewhat encouraged that over the weekend, [he was] able to make contact with this friend who sent the fax," the IJ remarkably stated: "I'm not interested about the friend. These friends any more.... I'm not going to give any credence whatever any `friend' has to say about it." She further made clear that her only concern after the hearing was the authenticity of Dia's passport and visa, and asked the Government to provide that information. Thus, the IJ expressed a desire for corroboration (i.e., "supporting documentation"), then discouraged Dia from providing it, only to criticize and penalize Dia for not providing it. Such arbitrariness necessarily undermines the IJ's reasoning.

In a recent case involving similar issues, we opined that an IJ's adverse credibility determination does not pass muster under the substantial evidence rubric when it is not supported by an adequate explanation of the IJ's reasoning. In Mulanga, 349 F.3d at 126-27, we encountered and criticized the very same type of analysis that we find problematic here. There, the IJ: 1) faulted the petitioner's failure to provide corroboration in a situation in which that failure seemed quite reasonable; 2) found the petitioner's account — which we found plausible in light of State Department reports — to lack "common sense," without further explanation; and, 3) dissuaded the petitioner from producing evidence, the lack of which was later criticized. Id. at 135-38. There, too, we found a lack of substantial evidence to support the adverse credibility determination, and concluded that the order should be vacated and the matter remanded for further proceedings. Id. at 138.

An even more significant aspect of Dia's testimony rejected by the IJ was his report that twenty-five members of the military went to his home to find him, and, upon realizing that Dia was not there, beat and raped his wife. One reason the IJ gave for discrediting this story was that Dia did not present any evidence that the men who did this were from the military, other than hearsay of his wife. We have two problems with this reasoning. The first is that the IJ again failed to engage in the three-pronged analysis we require when an IJ expects documentary evidence or corroboration, discussed above. The second is that, "[t]hough the hearsay nature of evidence certainly affects the weight it is accorded, it does not affect its admissibility in immigration [removal proceedings]." Kiareldeen v. Ashcroft, 273 F.3d 542, 549 (3d Cir.2001); cf. Ezeagwuna v. Ashcroft, 325 F.3d 396, 406 (3d Cir.2003) (referring to a letter as "multiple hearsay of the most troubling kind"). By matter-of-factly dismissing the evidence as "hearsay," the IJ failed to explain why it should be accorded no weight. We submit that such seemingly reliable hearsay evidence should not be rejected in such a perfunctory manner.

The other reasons that the IJ proffered for rejecting this testimony are similarly based on unsustainable grounds, namely, pure conjecture. The IJ "question[ed] why so many men would go to the respondent's home searching for him," and stated that "[i]t seems, to th[e] Court, [] an unusually large number for that purpose." The IJ also noted that Dia gave no explanation why the men raped his wife and stated that "[i]t seems unlikely to th[e] Court that men who were looking for the respondent would attack his wife." The basis for the IJ's having "questioned" this testimony is not explained, and appears speculative at best. Why would the IJ expect Dia to know why such a large group of men were seeking him or why they raped his wife? Twenty-five was an "unusually large number" — compared to what, or based on what? Without some explanation we are hard-pressed to understand why the IJ would find it so difficult to believe that a group of about twenty-five men would have come to Dia's home. Not only does this account seem consistent with common accounts of the practice of armed groups in war-torn countries, but the record contains documentary evidence confirming that police and civilian militias — groups of roving bands in large numbers — rape and attack suspected rebel supporters, as well as civilians, often burning and looting in the process. On what basis did the IJ determine that these aspects of Dia's account were "unlikely"? Again, we are left wondering.

We also conclude that the IJ unreasonably penalized Dia for presenting "no explanation as to why these men would beat and rape his wife." Dia testified that his wife told him that they were military and that they were looking for him because he supported the rebels. He showed a temporal proximity between Bangora's visit to his home and the raping of his wife. He also testified that Bangora knew of his affiliation with the RPG. So, Dia did, in fact, present some testimony as to why the men came to his home. It seems reasonable for Dia to believe that the militia consisted of government soldiers or was sent by the government. Yet, the IJ rejected this account without explanation. It can hardly be said that Dia had an obligation to investigate further the specific identity of the men who he claims raped his wife and burned his home. How could he have known this? Would not any explanation he might offer be criticized by the IJ as a fabrication, since surely they would not have shared their motivation with him? There is no authority that would require Dia to present direct evidence of the men's motives; rather, circumstantial evidence was sufficient. See Elias-Zacarias, 502 U.S. at 483, 112 S. Ct. 812; Navas v. INS, 217 F.3d 646, 659 n. 18 (9th Cir.2000) ("Where police beat and threaten the spouse of a known dissident, it is logical, in the absence of evidence pointing to another motive, to conclude that they did so because of the spouse's presumed guilt by association."). As a result, the IJ's conclusion in this regard, too, was unsupported either in the record or by a commonsense explanation.

The IJ's disbelief of Dia's testimony that his wife urged him to flee the country without her also lacked foundation in any logical reasoning or any support in the record, and thus seems to demonstrate more speculation and arbitrariness on the part of the IJ. Dia testified that "[his] wife told [him] if he would really like to see her again, next time in the future, that the best thing was for [him] to try to just flee and try to find a place. And [he] decided to leave her with her family members." The IJ stated that she found "it highly unlikely that a woman, who allegedly had just been beaten and raped, would urge her husband to leave her and not suggest that she accompany him out of the country." Why? The IJ also stated that "it appears unlikely that a man whose wife has just been beaten and raped by military personnel would not suggest that they flee the country together." Why? We can think of any number of reasons why Dia's wife might have urged him to leave without her.24 The IJ failed to share any basis for her conclusions and, without proper support, we cannot help but view them as not constituting substantial evidence.25

The second general aspect of Dia's story that the IJ rejected dealt with the circumstances surrounding his procurement of a passport and visa. The reasons put forth as support for the rejection, however, do not, again, satisfy the test for substantial evidence.

First, the IJ expressed incredulity that "a man" — i.e., Sow's friend, the "smuggler" — procured a new passport and a U.S. visa for Dia. Essentially, she concluded that the man did not exist because she found it unbelievable that Dia did not know the identity of a man who helped him leave the country and to whom he "paid a large sum of money." In fact, during the hearing she went so far as to dismiss Sow's friend out of hand as an "imaginary person." But she did so without inquiry into this aspect of Dia's story, and, in her written opinion, she failed to articulate why she considered it so unbelievable. What is more, as mentioned earlier, she inconsistently asked for more information about how Dia got the passport and visa, but stated that Dia could not provide any supporting evidence from any more "friends." As it is presented in her opinion, then, the dismissal of this aspect of Dia's testimony constitutes conjecture as to how the IJ believed Dia should have acted. But it seems highly unreasonable for the IJ, without inquiry into the underlying reasons, to assume that Dia would have, without fail, learned Sow's friend's name, given that Sow was his friend and that it was Sow, not Dia, who dealt with the smuggler.26 Why would someone in Dia's position care about the smuggler's name?

The IJ also inappropriately rejected Dia's testimony that he had never been to Italy. Dia testified that he was told by Sow's friend who procured his passport to say that he was a tomato picker in Italy on his way to Honduras, and that he told this story to the immigration officer at the airport. In rejecting Dia's testimony that he had never been to Italy, the IJ listed a litany of reasons: the U.S. visa in Dia's passport was issued in Italy; the visa and passport are "valid"; Dia's non-immigrant visa contains a signature of "Saidou Dia"; an INS memo to file, derived from an airport interview, states that Dia was a "tomato picker" in Italy; the "tomato picker" lie does not make sense because the fact that he was a tomato picker in Italy was "not beneficial to him as far as his intention to remain in the U.S."; Dia possessed employment documents from Houcon Cargo Shipping Company addressed to "Dia's" address in Italy; Dia's alleged ignorance as to how these documents came to be (and explanation that the smuggler got them for him) was unbelievable; and because she "question[ed] how this [shipping] company would know of an address of the respondent in Italy unless the respondent provided an Italian address to them."

The problems with these "reasons" are manifold. As Dia points out, the testimony and reports regarding the passport and visa bore out the fact that the passport and visa given him were "genuine" and seemingly issued in Italy, but not that they were necessarily legitimately obtained or even obtained by Dia. Dia explained consistently how he came into possession of the "genuine" documents and why he told the INS officials that he was a tomato picker on his way to Honduras. Dia also points out that, as discussed below, he presented an expert witness who gave unrebutted testimony that the handwriting on the documents was not Dia's. As for the letter from the shipping company, a purported address for Dia certainly does not render it authentic. The existence of the address on the letter does not mean that the address exists or that the letter was ever mailed, or even that the cargo company existed. In fact, to conclude that because the documents listed an Italian address, therefore the shipping company actually knew of Dia's address in Italy, strikes us as bizarre. Dia testified that he got the documents from the smuggler and, thus, that they were totally fraudulent. The record contains no evidence contradictory to this story, and the IJ does not explain why it was not plausible.

Basically, the IJ seemed confused by the fact that the documents that Sow's friend procured for Dia could be so good that the documents appeared genuine and that the supporting documentation would support Dia's Italy story. The only evidence in the record that the IJ used to support her conclusion was the INS memo regarding the airport interview, which states that "[h]is last employment appears to be as a Tomato picker while living illegally in Italy." But, if we are to rely on this memo, assuming it was probative, we must do so with care.

First, we are generally skeptical of reliance on reports of airport interviews. In Balasubramanrim v. INS, we stated that the airport interview is usually not "valid grounds upon which to base a finding that an applicant is not credible." 143 F.3d 157, 164 (3d Cir.1998) (citation and internal quotation marks omitted). We noted:

We do not know how the interview was conducted or how the document was prepared. We do not know whether the questions and answers were recorded verbatim, summarized, or paraphrased. We cannot tell from the document the extent to which [the petitioner] had difficulty comprehending the questions, whether questions had to be repeated, or when and how sign language was used. Nor does the document reveal whether [the petitioner's] responses actually correspond to those recorded or whether the examiner recorded some distilled or summary version based on his best estimation of the response.

Id. at 162; see also Mulanga, 349 F.3d at 137-38 (refusing to rely too heavily on the content of an airport interview and noting that "immaterial discrepancies between airport interviews and subsequent testimony should not be used to make adverse credibility determinations"); Zubeda, 333 F.3d at 477 (stating that "[c]aution is required" when considering what weight to give even to an asylum affidavit); Senathirajah, 157 F.3d at 216 (warning against placing too much reliance on an airport interview). Here, too, there does not appear to be any such information. In fact, it appears that much of the memo memorializes information not taken under oath, including the "tomato picker" story, and that the IJ completely disregarded the INS official's expressed lack of certainty, as evidenced by his use of "appears." And, the INS official noted that the information not taken under oath came forth during a conversation in French, a language the official noted Dia was only "able to converse somewhat in," and that Dia "asked for a Malinge interpreter for the sworn statement." AR237. But even more important is the fact that, even if we take the information from the airport interview as accurate, it does not contradict Dia's story, but is, in fact, consistent with it. Dia explained that the smuggler told him a story to present to the immigration officials in the U.S., which he clearly did. This does not prove that Dia had lived in Italy; rather, it bolsters his testimony that this was the story he was told to tell.

In addition, we cannot fathom why the IJ rejected the testimony of Dia's expert witness, Gregory McNally. McNally testified that the signatures on the passport and visa were not Dia's. He came to this conclusion after comparing six of Dia's known signatures with that on Dia's Republic of Guinea passport and with that on Dia's Non-Immigrant Visa Application for the U.S. visa found in the passport. Presumably, the purpose of the testimony was to show that, even though the passport and visa issued in Italy were "authentic," they were not in fact really Dia's because they did not contain Dia's actual signature. Establishing this fact certainly would buttress Dia's testimony that he never went to Italy or obtained the visa and passport himself. This testimony totally undercuts the IJ's reasoning, and it is especially important in that it undermines both the IJ's expressed skepticism that the passport and visa were not valid and her apparent conclusion that Dia must have been a tomato picker in Italy simply because the documents were "genuine." Surely she should have realized that they were "genuine" but were not Dia's. If her experience led her to reject that as improbable, she should have explained her thinking.

The IJ explained her rejection of McNally's testimony that the signatures on the passport and visa were not Dia's, by opining that handwriting analysis is too uncertain to accord it much weight. This outright rejection of McNally's testimony was unfounded. McNally's expertise was unchallenged. McNally was trained by and worked for the U.S. government, has testified as an expert in various courts more than one hundred times, and belongs to two relevant professional societies, one of which has officially certified him an examiner of questioned documents. In his testimony, he clearly concluded that the signatures on the passport and visa were not Dia's, thus lending support to Dia's story. McNally only qualified this conclusion by noting that he preferred to use original documents (some of the documents he had examined were not originals), and by conceding that "anything is possible" with regard to signatures.

The IJ supported her conclusion that handwriting analysis is not probative evidence by referring to United States v. Van Wyk, 83 F. Supp. 2d 515 (D.N.J.2000). However, Van Wyk does not stand for this proposition, but, instead, deals with the admissibility of a forensic stylistics expert's testimony under the Federal Rules of Evidence. Evidence presented in an immigration hearing needs to be "fair," "reliable," and "trustworthy," not necessarily admissible in federal court. Ezeagwuna, 325 F.3d at 405. More importantly, we have found that "[e]xpert testimony as to the similarities in handwriting is generally admissible" in federal court, United States v. McGlory, 968 F.2d 309, 346 (3d Cir.1992), and McNally's curriculum vitae lists dozens of courts in which he has testified as an expert. Therefore, for this reason as well, the chief reason articulated by the IJ for her rejection of Dia's testimony on this count — her conclusion that these were Dia's authentic documents — is not supported by coherent reasoning or by record evidence.27

The final area of Dia's testimony rejected by the IJ pertained to whether Dia would suffer future persecution if returned to Guinea. We emphasize that we address this aspect only to determine any errors in the IJ's credibility determination, not whether Dia should be found eligible for relief from removal. With that said, the IJ's conclusion here, again, lacks the necessary support.

The IJ decried that "[t]here is no evidence that low-ranking persons in the RPG are being arrested and detained," and noted that the only incidence of violence involving Dia during the approximately three years after Dia joined the RPG involved his wife, not Dia. The IJ also expressed doubt about Dia's testimony that the military was pursuing him because it thought that he was helping Alpha Conde escape from jail. She stated that it made little sense that the military thought Dia could enable Conde to escape, although she did not explain this further.

Once again, there are a number of problems with the IJ's fact-finding regarding this aspect of Dia's testimony. First, Dia did present evidence that the Government persecuted low-ranking persons in the RPG. He testified to his own past persecution, the rape of his wife as a means of menacing him, and the beating of local RPG members at the local headquarters, including the beating of Dia's father that resulted in death. See Baballah v. Ashcroft, 335 F.3d 981, 988 (9th Cir.2003) ("Threats and attacks can constitute persecution even where an applicant has not been beaten or physically harmed.... Violence directed against an applicant's family members provides support for a claim of persecution and in some instances is sufficient to establish [a well-founded fear of] persecution."). There are also reports in the record documenting the Guinean government's persecution of RPG members on all levels, including poll watchers like Dia, and a letter from the RPG attesting to the fact that the chief of the neighborhood and military officers attempted to force Dia to fight and that they beat his wife. In fact, Dia points out that the IJ actually states in her opinion that the Guinean government had arrested demonstrators who sought the release of Conde.28

Dia also urges that the IJ placed far too much weight on the fact that neither Dia nor his wife was the victim of any violence for the two years between the time Dia joined the RPG and the alleged rape of his wife and burning of his house, or after that time. He highlights parts of the record that show that the persecution of RPG members "increased significantly" in September of 2000 due to rebel attacks on border towns such as his. He also points out that Amnesty International reported on May 29, 2001, that the government security forces used violence, including torture and rape, "routinely" against members of opposition parties, such as the RPG, and that the security forces "continue to act with impunity." The report stated that "[t]he long-standing pattern of human rights violations by the Guinean security forces is clear." As for the four months after the military came to his home but prior to his flight from the country, Dia notes that he was hiding at Sow's home during that time. Lastly, and perhaps most tellingly, the IJ, again, misstated Dia's testimony. Dia never testified that the military thought that he was helping Conde escape prison. Rather, he testified that the military told his wife that Dia was helping the rebels so that Conde could be released from prison. When an IJ bases her conclusion on an erroneous interpretation of the testimonial and documentary evidence in the record, it undoubtedly is not supported by substantial evidence.

In sum, the IJ's adverse credibility determination was based on a combination of misstatements of Dia's testimony, unreasonably speculative or arbitrary conclusions, inaccurate or insufficiently explained findings of contradictions, and an arbitrary rejection of probative testimony. The "inconsistencies" that the IJ claimed to have found were illusory, and the claimed "implausibility" stemmed solely from conjecture. While we owe deference to the IJ's findings, our focus, and the essence of our review function, must be on the IJ's stated reasons. The reasoning in the IJ's opinion must "bear a legitimate nexus to the finding." Balasubramanrim, 143 F.3d at 162. We are not to invent explanations that may justify the IJ's conclusion. Accordingly, we conclude that the IJ's opinion was not based on substantial evidence; thus, the BIA improperly affirmed the IJ's decision.29

Dia asks that we reverse the BIA and grant the relief he seeks. But we are not finding Dia credible. Rather, we are concluding, as we have repeatedly before, that because of the lack of substantial evidence to support the adverse credibility determination, we will remand in order for the agency to further explain or supplement the record. Gao, 299 F.3d at 279. Moreover, as we have recently said: "We will not assess [Dia's] entitlement to relief based on the record as we have required it to be modified by this opinion because the agency should have the opportunity to do so." Ezeagwuna, 325 F.3d at 411 (citing Ventura, 537 U.S. at 17-18, 123 S. Ct. 353); see also Secaida-Rosales, 331 F.3d at 313 (stating, under similar circumstances, that "[t]he IJ should then reach the questions of asylum and withholding of deportation in light of such evidence, but without regard to its prior adverse credibility determination"); Senathirajah, 157 F.3d at 222 (remanding to BIA with instructions to remand to IJ for decision on asylum and withholding application, but without consideration of erroneous adverse credibility finding reversed on appeal). Instead, we will vacate the BIA order.30

* * * * * *

Accordingly, we will grant the petition for review, vacate the BIA's order summarily affirming the IJ's decision, and remand to the BIA.

ALITO, Circuit Judge, with whom Judges SLOVITER and ROTH, join, concurring in part and dissenting in part.

I join Part I of the opinion of the Court. However, because I believe that the Court's method of analyzing the Immigration Judge's credibility determination is seriously flawed, I would deny the petition for review.

Cases in which aliens seek asylum or withholding of removal based on the likelihood or probability of persecution if they are returned to their own countries are among the most difficult that we face. Much is obviously at stake, but the evidentiary record is very often meager. Indeed, it is common for an Immigration Judge (IJ) to have little other than the testimony of the applicant on which to base the decision, and this presents obvious and serious problems.

On the one hand, it is often unreasonable to demand that asylum-seekers provide corroborating evidence regarding their personal experiences in their home countries. Such incidents are often not memorialized (at least not in any records that are available to anyone outside the home government), and aliens fleeing persecution may be lucky to escape at all and may have no opportunity to secure any documentary evidence that exists or to obtain statements from witnesses who are willing to help. In addition, once asylum-seekers reach this country, they may find it very hard to obtain corroborating evidence. It may be difficult to contact people at home, and persons who might assist may be prevented by the home government from doing so or may be reluctant to help for fear of governmental retribution. Thus, it is often not reasonable to demand corroboration.

On the other hand, however, testimony by asylum-seekers cannot simply be accepted without question. Persons wishing to escape deplorable conditions that fall short of persecution have a strong motive to fabricate tales of persecution, and it must be recognized that such stories are not hard to construct. An asylum-seeker may have heard another person's account of persecution and may substitute himself or herself as the victim. Or an asylum seeker may take an incident in which he or she was actually involved and may exaggerate the conduct of the military or police so as to make it reach the high standard needed to constitute persecution.

As Professor David A. Martin, former general counsel of the INS, has written, the government is rarely able to conduct a field investigation of matters such as applicants' "past political activities, or specific abuses or threats directed against them or their families or friends." David A. Martin, "Reforming Asylum Adjudication: On Navigating the Coast of Bohemia," 138 U. Pa. Law Rev. 1247, 1280 (1990). As a result, "[a]sylum determinations often depend critically on a determination of the credibility of the applicant, for she will usually be the only available witness to the critical adjudicative facts of the case. Because that person has substantial incentives to lie or to embroider the truth (and few disincentives), this makes for a system vulnerable to manipulation." Id. at 1281-82 (footnotes omitted).1 These two factors — the frequent unavailability of corroboration and the ease of fabricating a persecution claim — make determinations regarding the credibility of asylum-seekers critically important.

In deciding how such determinations are to be made in this country, Congress could have taken a variety of different approaches. The approach that it chose, however, was to entrust the responsibility for making these important determinations to the Attorney General, with very limited judicial participation. Specifically, we must accept a credibility determination made by those to whom the Attorney General's authority has been delegated "unless any reasonable adjudicator would be compelled to conclude to the contrary." 8 U.S.C. § 1252(b)(4)(B). This limited role sometimes puts us in the uncomfortable position of deferring to a credibility determination about which we are skeptical. But the statute leaves us no alternative.

In analyzing the IJ's credibility decision in this case, the Court makes three fundamental mistakes. First, the Court fails to recognize that it is entirely proper for a fact finder to take into account "background knowledge" about human behavior in assessing the plausibility of testimony. Second, although the Court pays lip service to the very limited standard of review that we must apply, the Court in effect inverts that standard and refuses to sustain the IJ's credibility finding, not because a reasonable adjudicator could not find that Dia lacked credibility, but because, in the Court's view, a reasonable fact finder could make a contrary finding. Third, the Court errs in failing to take the totality of the circumstances into account in reviewing the IJ's credibility determination. Instead, the Court focuses one by one on specific statements that were made by Dia and asks whether each of those statements is plausible. The Court fails to recognize that a series of statements may, taken together, provide a reasonable basis for finding a witness to be incredible even if each statement standing alone might be credible.

"Background knowledge." In assessing the credibility of testimony, fact finders commonly ask whether the testimony is consistent with their own understanding of how people usually behave. Judge Weinstein uses the term "background knowledge" to describe the "`vast storehouses of commonly-held notions about how people... generally behave,'" and he explains that credibility determinations may be based on such knowledge even though it is not in the record. United States v. Shonubi, 895 F. Supp. 460, 479 (E.D.N.Y.1995) (citation omitted).

Professor Uviller has provided a useful description of the process used by a fact finder in deciding whether testimony comports with such "background knowledge." He writes:

The process by which a juror tests a story for plausibility involves some sort of cerebral matching. A juror juxtaposes a set of recounted actions and events against her experience, imagination, and derived intelligence concerning the behavior and reactions of real and fictitious others: "Is this the sort of thing that I might do in those circumstances?" "Would anyone I know have reacted that way?".... These are the sorts of guides to plausibility on which the juror must rely.

H. Richard Uviller, "Essay: Credence, Character, and the Rules of Evidence: Seeing Through the Liar's Tale," 42 DUKE L.J. 776, 783 (1993).

Professor Uviller also explains the weaknesses of this process:

The trouble, of course, is that frequently the cultural context and customs of the actors in the events recounted by the witnesses are totally alien to the jurors seeking a plausibility match. Neither the jurors nor anyone they are likely to know have had any experiences comparable to those now described from the witness stand by an adolescent drug dealer or a professional underworld hoodlum.

Id. Nevertheless, this process of judging credibility is employed every day in criminal and civil cases, and there is no reason why it cannot also be used in asylum cases.2 In other words, an IJ in an asylum case may properly judge a witness's credibility by comparing that testimony to the IJ's background knowledge about human behavior in general and about the behavior of those seeking entry into the United States.

In this case, however, the Court faults the IJ for employing this very process. As the Court notes, Dia testified that his wife told him that about 25 soldiers had come to their home looking for him and that, when Dia's wife told them that she did not know his whereabouts, the men beat and raped her and burned the house. AR 86. Dia stated that, at his wife's urging, he fled the country alone, leaving his wife and child behind. AR 86. In finding that Dia was not credible, the IJ expressed the belief that a woman in the position of Dia's wife would probably not urge her husband to leave her behind. AR 43. The IJ also expressed the belief that a man in Dia's position would probably "suggest that they flee the country together." AR 43. The IJ thus engaged in precisely the mental process that Professor Uviller described: she compared what Dia's wife allegedly did (urging that her husband flee without her) with what the IJ thought a woman in that position would likely do, and the IJ compared what Dia did (leaving alone) with what she thought a man in that situation was likely to do. Because the IJ concluded that spouses in the position of Dia and his wife would not likely behave in the way that Dia described, the IJ found it unlikely that Dia's story was true.

Although the mental process in which the IJ engaged was routine and entirely permissible, the Court rejects the IJ's approach. The Court criticizes the IJ for failing to explain her reasoning (Maj. Op. 255), but the basis for her conclusion could not be any clearer: the IJ's "background knowledge," "her experience, imagination, and derived intelligence concerning the behavior and reactions of ... others" told her that a woman was not likely to urge her husband to flee alone if she had just been beaten and raped by soldiers who were looking for him. Similarly, the IJ's "background knowledge" told her that a man was not likely to leave his wife and child behind under such circumstances. What more the Court expects the IJ to have said on this point is a puzzle. Does the Court expect the IJ to have cited to empirical evidence about how couples generally behave when the wife has just been beaten and raped by soldiers who are looking for the husband for political reasons?

Reversing the standard of proof. We are required to sustain the IJ's findings "unless any reasonable adjudicator would be compelled to conclude to the contrary." 8 U.S.C. § 1252(b)(4)(B). The Court, however, repeatedly turns this standard on its head and finds that aspects of Dia's testimony should have been found to be credible because a reasonable person might have found them believable.

The Court's evaluation of Dia's testimony about his decision to flee by himself is again instructive. Rejecting the IJ's determination that it was "unlikely" that a couple in the position of Dia and his wife would agree that the husband should escape alone, the Court states: "We can think of any number of reasons why Dia's wife might have urged him to leave without her," such as a belief that Dia "could move more quickly and elude detection more easily if he was traveling alone." Maj. Op. at 255 & n. 23. In other words, the Court inverts the statutory test and refuses to accept the IJ's finding because the Court can imagine grounds for a contrary finding.

The Court performs the same slight of hand in evaluating the "tomato picker" story. Although Dia testified at the asylum hearing that he had never been in Italy (AR 102), the INS memo regarding his interview upon arrival at JFK Airport on March 27, 2001, reported that Dia had apparently been employed as a tomato picker in Italy. AR 237. The memo also reported that Dia was en route to join a ship in Honduras as an engine room technician (AR 237), and there is documentary evidence that supports the account in the memo. Dia had in his possession a valid United States visa that bore his name and photo and that had been issued in Milan. AR 225. He also possessed an "Employment Letter" from Houcon Cargo Systems of Rotterdam that was addressed to him at an address in Brescia, Italy. AR 218-221. The letter stated that Dia had been employed by the company for a two-year period as an engine room technician and instructed that he was to board a ship in Honduras by April 10, 2001. AR 218. When asked about these documents, Dia stated that they had been provided to him by the unnamed man who had procured his passport, but the IJ found that account dubious. AR 44.

Instead of asking whether a reasonable adjudicator could infer from this proof that Dia had been in Italy and had therefore lied at the hearing, the Court postulates all sorts of possible explanations for the evidence pointing to residence in Italy. Dia's statement at the airport might have been misunderstood because of his lack of facility in French, the language in which he spoke at that time; if Dia did say at the airport that he had been employed in Italy, he might have told an untruth because his statement "was not taken under oath" and he was simply parroting what the man who provided the documents had told him to say. Maj. Op. at 257-58. Although the visa was genuine, that did not prove that it had been "necessarily legitimately obtained." Id. at 257. And although the visa bore Dia's photo, it might have been obtained by someone else. Id. Houcon Cargo might not exist. Id. at 257. If it does, the letter might not be authentic. Id. The address in Brescia, Italy, might be fictitious, and even if it is not, the letter might have found its way into Dia's possession without being mailed to that address. Id. The Court concludes:

Dia testified that he got the documents from the smuggler and, thus, that they were fraudulent. The record contains no evidence contradictory to this story.

Id.

In effect, the Court says that the IJ was bound to accept Dia's testimony that he had never been in Italy unless there was conclusive proof to the contrary. If we applied this standard to the findings that are routinely made by judges and juries in federal litigation, few findings could be sustained.

Totality of the circumstances. In judging the credibility of a witness's story, a fact finder is entitled to consider whether the story as a whole has the ring of truth. Suppose that a witness asserts that something a bit out of the ordinary happened. Since unusual things do happen, a fact finder might credit that assertion if the witness does not say anything else that is questionable. But suppose that the witness goes on to assert that a whole series of unusual things happened. At some point, the fact finder may — reasonably — conclude that the witness's testimony as a whole is unbelievable.

Here, the Court rejects the IJ's determination that Dia's testimony was not credible because, taking each of his contested statements one by one, the Court finds each plausible. Thus, the Court fault's the IJ's credibility finding because, in the Court's view, it is plausible that about 25 men would be sent to Dia's home to look for him; it is plausible that Dia would flee alone and leave his wife and child behind even though his wife had been beaten and raped by soldiers looking for him; it is plausible that Dia was able to procure a police stamp even though the authorities were looking for him; it is plausible that a man whose identity Dia did not know was able to obtain for Dia a legitimate passport and legitimate United States visa; it is plausible that even though Dia had never been in Italy, he would say on arriving in this country, that he had been employed in Italy as a tomato picker; and it is plausible that even though Dia had never been in Italy, he would have a United States visa that was stamped as issued in Milan and an Employment Letter addressed to him at a place in Italy. Even if the IJ was bound to view each of these facts as plausible — and I do not think that she was — it does not follow that she was bound, when considering all of these facts together, to find Dia credible.

Viewing the record as a whole, I believe that a reasonable fact finder could find that Dia's testimony was not believable and could deny his application on that basis.

• In light of the previously mentioned documentary evidence and Dia's statement at the airport, it was reasonable for the IJ to question whether Dia told the truth when he testified at the hearing that he had never been in Italy. While this evidence may not prove conclusively that Dia had been in Italy, it certainly provided a reasonable basis for the IJ to infer that he had. And if Dia lied about this point, the IJ could reasonably doubt the truthfulness of the remainder of his testimony.

• It was reasonable for the IJ to question Dia's truthfulness based on Dia's statement that, even though his wife had been raped and beaten by men who were looking for him, he and his wife agreed that he should flee alone. As explained above, the IJ's belief about how a couple would likely react under such circumstances is just the sort of "background knowledge" about human behavior that a fact finder is entitled to consider in evaluating a witness's credibility. Moreover, the particular belief in question (that a couple, under the circumstances, would not likely decide that the husband should run away by himself) is at least reasonable. Since Dia's wife had allegedly been beaten and raped on one occasion when she did not tell Dia's pursuers where he was, is it unreasonable to think that a couple in that position would fear that something similar might recur if the husband left alone and the wife was again questioned?

• It was reasonable for the IJ to question Dia's testimony that about 25 men went to his house looking for him. A reasonable adjudicator could be skeptical that so many men would be sent to look for a single, low-ranking person. Similarly, it was reasonable for the IJ to question whether, as Dia testified, his alleged pursuers would think that a low-ranking person like Dia could help the leader of an opposition political group to escape. And it was reasonable for the IJ to question how Dia was able to obtain a police stamp on his passport if the authorities were searching for him. While each of these statements standing alone may not give rise to a strong inference of mendacity, each statement can be questioned, and when taken together they can reasonably contribute to a finding that Dia was not credible.

Whether I would have believed Dia if I had been given the responsibility to make that determination I cannot say. But viewing all of the evidence in the record and applying the narrow standard of review prescribed by statute, I cannot say that a reasonable adjudicator could not find Dia to be incredible.

STAPLETON, Circuit Judge, dissenting, with whom Judges McKEE, AMBRO, and BECKER join.

An applicant for asylum seeks to avoid removal to a country where he insists he will be the victim of persecution. In recognition of the serious consequences of an erroneous denial of asylum, Congress, in the INA, has given asylum seekers the right to present evidence to an immigration judge ("IJ"), 8 U.S.C. § 1229a (2002),1 the right to an administrative appeal of an adverse decision to the Board of Immigration Appeals ("BIA"), 8 U.S.C. § 1229a(c)(4),2 and the right to judicial review by a Court of Appeals of a final agency order denying asylum and directing removal. 8 U.S.C. § 1252 (2002).3

The Attorney General's streamlining regulations direct a single BIA member to make the final decision on whether a denial of asylum is the "correct" decision, 8 C.F.R. § 1003.1(a)(7)(ii) (2002), and then instructs the Court of Appeals to review the IJ's explanation for his or her disposition4 — an explanation that the BIA member expressly declines to adopt. 8 C.F.R. § 1003.1(a)(7)(iii).5 By thus severing the final decision-maker from the agency's explanation for its denial, the Attorney General effectively says to asylum seekers, "You have a right to appeal to a Court of Appeals, but asylum may be denied and you may be removed for reasons that are not subject to its review." Today, our court sanctions this perversion of judicial review. Chevron does not require us to defer to such a perversion, and I would decline to do so.

The scheme of administrative and judicial review established in the INA for asylum cases is a familiar one. Comparable schemes have been before the Supreme Court on numerous occasions, and its decisions in those cases have articulated fundamental principles of administrative law that govern review under such schemes. See, e.g., SEC v. Chenery Corp., 332 U.S. 194, 196-97, 67 S. Ct. 1575, 91 L. Ed. 1995 (1947); Chevron U.S.A. Inc. v. Natural Res. Def. Council, 467 U.S. 837, 842-44, 104 S. Ct. 2778, 81 L. Ed. 2d 694 (1984); INS v. Ventura, 537 U.S. 12, 17, 123 S. Ct. 353, 154 L. Ed. 2d 272 (2002). Nothing in the INA or its legislative history suggests that Congress intended that review in asylum cases would depart from these well-established principles, and the statute must be applied in light of them. I conclude that the Attorney General's streamlining regulations are inconsistent with the INA scheme as so applied.

In a number of ways, the Attorney General's reading of the statute is more straightforward than that of the Court. The Court begrudgingly "assume[s] [that the Act] contemplates that an alien will have the opportunity for an administrative appeal." Maj. Op. at 237-38. It does so because Congress has expressly mandated that an IJ, upon the entry of an order denying asylum and granting removal, "shall inform the alien of the right to appeal that decision." 8 U.S.C. § 1229a(c)(4). The Attorney General has no doubts about the matter, and his streamlining regulations provide for such an appeal. Those regulations continue to recognize that, in accordance with the requirements of the statute, there can be no final agency order until the BIA has acted or the asylum seeker has foregone his right to appeal to the BIA. 8 U.S.C. § 1101(a)(47)(B) (2002).6 Thus, in any case where an appeal is taken, the final decision of the agency subject to judicial review is that of the BIA. Moreover, the Attorney General understands that, while the statute substantially constrains judicial review of administrative findings of fact,7 it contemplates de novo review by the BIA. Even when streamlining occurs, the single BIA member, before approving "the result reached in the decision under review," must determine that it "was correct." 8 C.F.R. § 1003.1(a)(7)(ii). As we succinctly explained in Abdulai v. Ashcroft, 239 F.3d 542, 548-49 (3d Cir.2001):

Congress has granted [the federal courts] power to review only "final order[s] of removal." Because an alien facing removal may appeal to the BIA as of right, and because the BIA has the power to conduct a de novo review of IJ decisions, there is no "final order" until the BIA acts. Accordingly, we now expressly hold that the "final order" we review is that of the BIA.

See also Kayembe v. Ashcroft, 334 F.3d 231, 234 (3d Cir.2003) ("Our power of review... extends only to the decision of the BIA.... Therefore, only if the BIA expressly adopts or defers to a finding of the IJ, will we review the decision of the IJ."); Awolesi v. Ashcroft, 341 F.3d 227, 231 (3d Cir.2003) ("We review the decision of the BIA, not that of the IJ.").

The Attorney General also recognizes, as he must, that an asylum seeker is entitled under the statute to judicial review of the agency's final decision, and the streamlining regulations so provide. They go on, however, to deprive the Court of Appeals of any basis for reviewing the agency's final decision, to deprive the Court of the benefit of the agency's expertise, and to deprive the asylum seeker of de novo review of the IJ's fact finding. These aspects of the streamlining regulations are at odds with the statute as applied in light of at least two well-established principles of administrative law.

First and foremost, judicial review necessarily requires something to review and, if the agency provides only its result without an explanation of the underlying fact finding and analysis, a court is unable to provide judicial review. See SEC v. Chenery Corp., 332 U.S. 194, 196-97, 67 S. Ct. 1575, 91 L. Ed. 1995 (1947) ("If the administrative action is to be tested by the basis upon which it purports to rest, that basis must be set forth with such clarity as to be understandable.");8 Dunlop v. Bachowski, 421 U.S. 560, 572, 95 S. Ct. 1851, 44 L. Ed. 2d 377 (1975); Guentchev v. INS, 77 F.3d 1036, 1038 (7th Cir.1996) (statement of reasons is the "norm of administrative law"). As we also explained in Abdulai in the specific context of the INA:

[N]othing in the INA specifically requires the Board to explain its decisions. But the availability of judicial review (which is specifically provided in the INA) necessarily contemplates something for us to review.... Because the BIA's failure of explanation makes it impossible for us to review its rationale, we grant the petition for review, vacate the Board's order, and remand the matter....

Abdulai, 239 F.3d at 555. While we did not cite to Chenery for this proposition in Abdulai, our holdings with respect to judicial review of administrative decisions clearly embrace Chenery's reasoning. See, e.g., N.L.R.B. v. Permanent Label Corp., 657 F.2d 512, 532 (3d Cir.1981) ("Requiring the Board to articulate its reasons for imposing a bargaining order does not represent an unwarranted judicial interference with administrative procedure. Indeed that requirement, as we have pointed out, stems from the Supreme Court's instructions in [N.L.R.B. v. Gissel Packing Co., 395 U.S. 575, 89 S. Ct. 1918, 23 L. Ed. 2d 547 (1969)], and from fundamental rules of administrative law." (citing Chenery, 332 U.S. at 196-97, 67 S. Ct. 1575)).9

Whenever the streamlining regulations are invoked, they deprive a reviewing court, by their express terms, of any basis for knowing either the findings of fact or the rationale upon which the BIA's denial of asylum rests and they thus make judicial review impossible. 8 C.F.R. § 1003.1(a)(7)(iii) requires that the BIA's order expressly state "the Board affirms, without opinion, the result of the decision below." This means that the BIA accepts only "the result" of the IJ's deliberations. Indeed, the regulations, in addition to foreclosing the BIA from giving "further explanation or reasoning," go on to state that the effect of the BIA's order is to "approve... the result reached in the decision below" and to disavow any implication that the approval extends any further. The BIA's order thus does not only fail to adopt the findings and reasoning of the IJ, it expressly disavows endorsing those findings and that analysis. See 8 C.F.R. § 1003.1(a)(7)(iii) (requiring that the order to state that it "does not necessarily imply approval of all of the reasoning of [the IJ's decision]"). The net and necessary result is that the reviewing court has no findings or reasoning of the final decision-maker to review. This result has already been addressed in Chenery, 332 U.S. at 196-97, 67 S. Ct. 1575, and Abdulai, 239 F.3d at 555, and it is rendered no less defective by virtue of being authorized by the Attorney General's streamlining regulations.

In short, judicial review, by definition, necessarily involves an explanation for the agency's final decision. Congress has spoken directly to the issue of whether asylum seekers will have a right to judicial review. Accordingly, the situation before us is simply not one in which Chevron deference is due. See Chevron, 467 U.S. at 842-843, 104 S. Ct. 2778 ("First, always, is the question whether Congress has directly spoken to the precise question at issue. If the intent of Congress is clear, that is the end of the matter; for the court, as well as the agency, must give effect to the unambiguously expressed intent of Congress."). The Attorney General may not deprive asylum seekers of that right by calling something judicial review even though relief may be denied for undisclosed reasons not subject to judicial review.

It does not help that a summary affirmance signifies under the regulations that the BIA views "any errors in the decisions of the Immigration Judge [as] harmless or non-material." In any case where the BIA disagrees with the IJ in whole or in part, but has an alternative and independent basis for denial, the error of the IJ will be harmless or non-material, yet that alternative and independent basis will never be disclosed and, accordingly, will never be exposed to judicial review. It necessarily follows that an alien may be denied asylum and be put in serious jeopardy for a reason that is never exposed to judicial review.

Nor does it help that the regulations instruct the Court of Appeals to review the fact finding and analysis of the IJ. See Executive Office of Immigration Review: Board of Immigration Appeals Streamlining, 64 Fed. Reg. at 56,138 (Oct. 18, 1999) ("The decision rendered below will be the final agency decision for judicial review purposes.... [T]he Immigration Judge's decision becomes the decision reviewed."). In any case in which the BIA disagrees with the IJ in whole or in part, but has an alternative and independent basis for denying asylum, review of the IJ's fact finding and analysis is a meaningless exercise. The Court's suggestion that it makes no difference whether the explanation is provided by the final decision-maker or the IJ is reminiscent of Alice's Wonderland. The difference is between the asylum seeker's having judicial review of the reason for his removal and his having no such review.10

The Attorney General's perversion of the judicial review mandated by Congress is strikingly illustrated by the record that was recently before us in Ezeagwuna v. Ashcroft, 301 F.3d 116 (3d Cir.2002), a case that was not, but could have been, affirmed without opinion under the streamlining regulations. There, as here, the IJ concluded that the alien was not credible and, accordingly, had not carried his burden of persuasion. Id. at 123. The IJ reached this conclusion because: (1) he found it implausible that the alien had been abused in the manner she claimed; and (2) the alien appeared to be giving testimony she had rehearsed. Id. at 123-24. On appeal, the BIA found that the record did not support the reasons given for the IJ's credibility decision. Nevertheless, it also found that the alien lacked credibility and, accordingly, that the IJ had reached the correct result. The BIA reached its conclusion on the credibility issue for a reason different from those of the IJ: the alien had impeached her own credibility by submitting fraudulent documents in support of her asylum application. The BIA accordingly ordered removal. Id. at 124-25.

On appeal, we held that the evidence the BIA relied upon in concluding that the documents were fraudulent was unreliable and untrustworthy and that the BIA's reliance upon it had violated the alien's right to Due Process. Id. at 130.

Cases like Ezeagwuna can, consistent with the streamlining regulations, be decided without an opinion. A single BIA member could well have determined in good faith that the errors of the IJ were harmless or immaterial because the BIA member reached the same conclusion. He could also have concluded in good faith that there was no genuine dispute regarding the applicable law and that the case did not involve the application of precedent to a novel fact situation. Had a single BIA member made these determinations and decided the case without opinion under the streamlining regulations, we, as the reviewing court, (1) would have been unaware of the basis for the agency's final decision and, accordingly, would not have had the opportunity to vindicate the violation of the alien's constitutional rights; and (2) would have been unaware that, in the expert eyes of the BIA exercising de novo review, the fact finding done by the IJ was unacceptable and, exercising restricted review, we might well have allowed that fact finding to stand. Fortunately for the asylum seeker in Ezeagwuna, her case was not streamlined and her rights were vindicated. Had the streamlining decision gone the other way, however, she might well have been sent home to persecution without any judicial review of the basis of the decision to deny asylum.

This case may well be one like Ezeagwuna. The BIA member could well have viewed the grounds for the IJ's credibility determination as impermissibly speculative and, accordingly, may well have relied on alternative, undisclosed grounds. If so, Dia will have been removed for reasons that were unexposed to judicial review.

The streamlining regulations, and the Court's sanctioning of them, have further consequences that conflict with the statutory scheme viewed in light of a second well-established principle of administrative law. A reviewing court must defer to, and must insist upon the benefit of, the application of the agency's expertise in the context of the matter before it. "[A] reviewing court, in dealing with a determination or judgment which an administrative agency alone is authorized to make, must judge the propriety of such action solely by the grounds invoked by the agency. If those grounds are inadequate or improper [or unknown] the Court is powerless to affirm the administrative action by substituting what it considers to be a more adequate or proper basis." Chenery, 332 U.S. at 196, 67 S. Ct. 1575. The absence of an explanation from the agency decision-maker not only precludes judicial review, it also deprives the Court of the benefit of the agency's expertise and impairs the ability of the Court to confine itself to its intended role.

In a recent asylum case, the Court of Appeals for the Ninth Circuit overturned an order of removal based on an argument for asylum that had been addressed by the IJ but not by the BIA. The argument turned on whether conditions in Guatemala had improved to the point that no realistic threat of persecution currently existed. The Supreme Court held that the Court of Appeals had violated "well-established principles of administrative law" by proceeding "without giving the BIA the opportunity to address the [disputed] matter in the light of its own expertise." INS v. Ventura, 537 U.S. 12, 17, 123 S. Ct. 353, 154 L. Ed. 2d 272 (2002). The Court explained why the Ninth Circuit Court of Appeals's action "seriously disregarded the agency's legally mandated role":

Within broad limits the law entrusts the agency to make the basic asylum eligibility decision here in question. In such circumstances a "judicial judgment cannot be made to do service for an administrative judgment." Nor can an "appellate court ... intrude upon the domain which Congress has exclusively entrusted to an administrative agency." A court of appeals "is not generally empowered to conduct a de novo inquiry into the matter being reviewed and to reach its own conclusions based on such an inquiry." Rather, "the proper course, except in rare circumstances, is to remand to the agency for additional investigation or explanation."

INS v. Ventura, 537 U.S. at 16, 123 S. Ct. 353 (citations omitted).

Just as Congress did not intend the Ninth Circuit Court of Appeals to perform the BIA's "legally mandated role," neither did it intend for our Court to address in the first instance issues like whether Guinea is a "country where the military, as well as civilian militias, act independently of the federal government." Maj. Op. at 252. While our Court has not usurped the role of the BIA as did the Ventura court, it has permitted the agency to absolve itself, by regulation, of the responsibility for fulfilling the BIA's legally mandated role. Given the Congressional mandate, our Court should not allow this abdication of responsibility.

The Attorney General's desire to streamline the appeals process is understandable. Moreover, I agree with the Court that he has been given by Congress the authority to simplify and expedite the process before the agency. I perceive no reason, for example, why he cannot have a single BIA member decide a category of appeals like the one identified in the streamlining regulation. Moreover, a single member can clearly decide appeals by adopting the opinion of the IJ or by adopting it with specified exceptions. What the Attorney General may not do consistent with the INA and well-established principles of administrative law is to deprive the reviewing court of the ability to provide the judicial review mandated by Congress. More specifically, he cannot foreclose the BIA from explaining its decision in some way.

Contrary to the Court's suggestion, the Attorney General's streamlining scheme is not at all analogous to the affirmance of District Court judgments by court of appeals without an opinion. The BIA's order is materially different from a judgment order of a court of appeals that may be reviewed by higher judicial authority. While an opinion of a court of appeals may be helpful, it is not essential to the intended operation of the federal judicial system. A court of appeals possesses no special expertise that the Supreme Court lacks and both courts review a court's judgment using the same standards of review. The INA, on the other hand, contemplates an administrative review by an entity which has special, relevant expertise and which will exercise de novo review of the IJ's findings of fact as well as her conclusions of law. Both factors are very important in cases like the one before us, and the Congressional scheme cannot function as intended if the BIA does not tell us what facts it found and what reasoning it relied upon.

I would grant the petition for review and remand to the BIA with instructions to provide an explanation of the grounds for its decision.

McKEE, Circuit Judge, concurring in part and dissenting in part.

I join Part II of the majority opinion because I agree that we must grant the Petition for Review based upon the many problems with the Immigration Judge's adverse credibility ruling that the majority explains. However, I must respectfully dissent from Part I of the majority opinion upholding the streamlining regulations. In my view, those regulations should be invalidated for all of the reasons so ably explained in Judge Stapleton's thoughtful dissent.

I write separately because I am troubled by the majority's suggestion that the IJ's flawed and unsupportable credibility ruling could somehow have been saved if it were based upon Dia's demeanor rather than the substance of his testimony. The majority notes that "the IJ did not rely on her personal observations of Dia's demeanor or any other observations to which we must accord an even greater degree of deference." Majority Opinion at 252 n.23.

It is of course true that the law has traditionally recognized a relationship between demeanor and credibility. The Supreme Court has even noted that the opportunity to observe a witness's demeanor is embodied in the Confrontation Clause of the Sixth Amendment. See California v. Green, 399 U.S. 149, 158, 90 S. Ct. 1930, 26 L. Ed. 2d 489 (1970) (stating that confrontation "permits the jury that is to decide the defendant's fate to observe the demeanor of the witness in making his statement, thus aiding the jury in assessing his credibility").1 However, this principle has evolved in the context of proceedings where the fact finder and witnesses share a common culture. Fact finders who are unfamiliar with the mannerisms and subtleties of a witness's cultural tradition have no advantage in assessing credibility based upon demeanor. Moreover, to the extent that the customs of a witness's native land differ from the fact finder's, the fact finder may be at a substantial disadvantage because he/she may misinterpret subliminal clues that mean one thing in the fact finder's culture, but something entirely different in the witness's.

Courts have addressed the extent to which ignorance about an alien's native land can shape conclusions. See Senathirajah v. INS, 157 F.3d 210, 220-21 (3d Cir.1998) (finding that the IJ's unsupported assumptions about the Tamil Tigers group and the Sri Lankan government were not an appropriate basis for her factual findings). However, courts have not been as willing to recognize that unfamiliarity with a witness's cultural experience may similarly color the intangible aspects of credibility determinations. Moreover, case law demonstrates that even experienced IJs who are accustomed to evaluating the testimony of aliens are not immune from allowing their conclusions to be colored by such cultural bias.

In Chouchkov v. INS, the Court of Appeals for the Ninth Circuit cautioned: "It must be stressed that what sounds peculiar in one country may be the norm in another." 220 F.3d 1077, 1083 n. 15 (9th Cir.2000). In doing so, the court cited Perez-Alvarez v. INS, 857 F.2d 23, 24 (1st Cir.1988). There, the Court of Appeals for the First Circuit incorporated the comments of the dissenting member of the BIA into the court's opinion. In his dissent from the decision of the BIA, Board Member Heilman had proclaimed:

[T]he evidence was cut off on the apparent assumption that [evidence of] a 10-year-old membership in a union was too old or too stale to constitute a ground for persecution. Perhaps this is so, but there is nothing in the record to sustain the immigration judge's assumption in this regard, except perhaps his general perception of life or political conditions in El Salvador which may or may not be grounded in fact.

As a general rule, in considering claims of persecution I think it highly advisable to avoid assumptions regarding the way other societies operate. Time and again this Board has considered appeals in which assumptions of this nature have been proven to be totally wrong, once the applicant has been given a full hearing.

Id. (emphasis added).

In Cordero-Trejo v. INS, 40 F.3d 482, 490 (1st Cir.1994), the IJ based an adverse credibility determination in part upon the fact that petitioner's wife had signed her full name on several letters to petitioner and she had addressed the letters using petitioner's formal name rather than using a more familiar reference. The IJ believed that was suspicious based upon his assumption that "one would normally expect the spouse to use the more familiar form" when addressing letters to her husband. Id. (internal quotation marks omitted). On appeal, the court rejected the IJ's skepticism because "there was [no] evidence in the record to suggest that signing a letter to a spouse residing in a foreign country by using one's full name is contrary to the common practice of someone of [the petitioner's wife's] cultural background." Id.

The cultural bias at the heart of the adverse ruling of the IJ in Barapind v. Rogers, No. 96-55541, 1997 WL 267881 (9th Cir. May 15, 1997) (reported at 114 F.3d 1193 as an unpublished summary affirmance), furnishes an even more dramatic example of the dangers of assessing credibility across a cultural divide and also illustrates the danger of placing too much emphasis on demeanor without elaboration or explanation.2 There, the IJ rejected an alien's testimony based in part upon the IJ's belief that the alien's "stoic" demeanor while testifying was inconsistent with having been subjected to the kind of gruesome torture he testified about in support of his asylum claim. The IJ thus concluded that the alien's "`stoic' demeanor as he testified about torture by the Indian police was a sign that he was lying." Id. at *2. On appeal, the court recognized that the alien petitioner's demeanor reflected his cultural tradition. The court explained that "stoic acceptance of misfortune is expected from persons of constancy and courage," and Sikhs had "long enjoyed the reputation of being `unsurpassed' as soldiers." Id. (quoting LEPEL HENRY GRIFFIN, RANJIT SINGH 36-37 (1892)).

The IJ also doubted that the alien was 29 years old as he testified. Based only upon her personal observation of the alien during his testimony, the IJ "thought he looked 40." Id. The court quickly rejected the purported age discrepancy as a basis for concluding that the alien was not credible. The court stated: "We see no basis for the IJ to have thought her own sizing up of the physical appearance of an alien gave her a superior insight into the age of the Sikh before her; still less do we see how her hunch showed that Barapind was lying." Id. The court dismissed the IJ's conclusion that the alien lied about his age as nothing more than a "hunch" improperly based upon "personal conjecture." Id. at *2, *3. The court concluded, "[a]gain, the inference by the IJ seems to reflect her own cultural bias." Id. at *2. Indeed, given the alien's testimony that he was subjected to torture that included applying electric shock to various "parts of his body," id., it would have been surprising if he had not appeared to be older than he was.

These cases primarily exemplify tangible manifestations of bias. However, resting factual conclusions upon unexplained and unarticulated demeanor poses an even greater risk of biased fact finding that can deny a petitioner due process of law.3

For example, eye contact plays a central role in evaluating the credibility of a witness in our own culture. The central issue in Morales v. Artuz, 281 F.3d 55 (2d Cir.2002), was whether the defendant's constitutional right of confrontation had been violated by the trial judge allowing a key defense witness to testify while wearing sunglasses that were so dark that the jury could not see her eyes. In writing for a unanimous panel, Judge Newman thoughtfully outlined the importance of the role that eye contact has traditionally been afforded in this society.4 The analysis began by explaining that the sunglasses created no obstacle to the right of confrontation insofar as the Confrontation Clause seeks to guarantee cross-examination, but conceivably infringed on the right of confrontation to the extent the Confrontation Clause "assures an opportunity for the defendant, especially jurors to see the witness's eyes in order to consider her demeanor as an aid to assessing her credibility...." Id. at 60. The court cited several cases in noting that "`eye contact' [is] among [the] factors aiding the fact-finder in assessing a witness's credibility." Id. (quoting Churchill v. Waters, 977 F.2d 1114, 1124 (7th Cir.1992)). Similarly, in Coy v. Iowa, 487 U.S. 1012, 1019, 108 S. Ct. 2798, 101 L. Ed. 2d 857 (1988), the Court stated that the trier of fact could "draw its own conclusions" from a witness who looked away from the defendant while testifying.

Even assuming arguendo the presumed relationship between such demeanor and credibility in the usual context, I submit that the relationship is often non-existent when the fact finder and witness are from different cultures. Thus, while the failure to look someone in the eye while speaking is usually interpreted as an indication of deception by people in Western cultures, avoiding eye contact has a very different meaning in some other cultures. See Joanna Ruppel, The Need for a Benefit of the Doubt Standard in Credibility Evaluation of Asylum Applicants, 23 COLUM. HUM. RTS. L. REV. 1, 12-13, 13 n.44 (1992) (quoting panelist Ira J. Kurzban, Esquire, as saying the assumptions made about the relationship between eye contact and credibility can be "the product of culture and not credibility" in Annual Judicial Conferences, Second Judicial Circuit of the United States, 115 F.R.D. 349, 440 (Sept. 4, 1986)). For example, in certain Asian cultures, avoiding eye contact is a sign of respect, and direct eye contact is considered inappropriate in traditional Navajo society. See Paul R. Tremblay, Interviewing and Counseling Across Cultures: Heuristics and Biases, 9 CLINICAL L. REV. 373, 394 (2002). A witness from a culture where it is disrespectful to "look someone in the eye" would naturally be expected to testify in a manner that reflected the solemnity and respect inherent in all judicial proceedings, including proceedings before an immigration court. It would be very unlikely that such a witness would maintain eye contact while answering questions out of respect for the interrogator, the judge, and the proceedings. Yet, this very manifestation of respect may cause the fact finder to conclude that such a witness is not credible and therefore view all of his/her testimony with a jaundiced eye.5 When this happens, "inconsistencies" that ought to convey nothing more than cultural differences or the fragile imperfections of memory can assume unwarranted importance.

Once a fact finder begins to doubt the veracity of a witness, it will be exceedingly difficult for even the most compelling witness to offer testimony sufficient to sustain his/her burden under the immigration laws. We have recognized that aliens often have to flee their native land with precious little documentation or corroboration. Senathirajah, 157 F.3d at 216 ("[O]ne who flees torture at home will rarely have the foresight or means to do so in a manner that will enhance the chance of prevailing in a subsequent court battle in a foreign land."). The alien trying to qualify as a "refugee" or for relief under the Convention Against Torture will therefore usually have precious little other than his/her own testimony to take before an IJ. See Matter of Mogharrabi, 19 I. & N. Dec. 439, 445, 1987 WL 108943 (B.I.A.1987) ("The alien's own testimony may in some cases be the only evidence available [to support his or her claims], and it can suffice where the testimony is believable, consistent, and sufficiently detailed to provide a plausible and coherent account of the basis for his fear.").

Moreover, cross-cultural misunderstandings about the veracity of petitioners' testimony can be exacerbated by difficulty understanding the procedure and structure of immigration proceedings. The proceedings are conducted in English, and petitioners are generally not provided with simultaneous translation. See Anker Study, supra note 3, at 505-06. In addition, the structure of the hearings is not transparent to petitioners. An empirical study of U.S. immigration court hearings and decisions found that "the simultaneously ambiguous and rigid structure of the hearing and the judges [sic] perceived need to control and limit the scope of the hearing, in many instances made it difficult for [petitioners] to communicate intelligibly the essential facts that formed the basis of their claims." Id. at 515.

Furthermore, petitioners often fear government officials because of past persecution in their native country. Cf. Balasubramanrim v. INS, 143 F.3d 157, 163 (3d Cir.1998) ("[A]n arriving alien who has suffered abuse during interrogation sessions by government officials in his home country may be reluctant to reveal such information during the first meeting with government officials in this country."); Senathirajah, 157 F.3d at 218 (stating that a petitioner may be reluctant to disclose the breadth of his suffering in his home country to a government official upon arriving in the United States). This may only exacerbate the difficulties of articulating the basis of a valid claim during immigration proceedings even if the petitioner does not exhibit the kind of demeanor that will suggest deception.6

It can not be overstated that "[c]aution is required because of the numerous factors that might make it difficult for an alien to articulate his/her circumstances with the degree of consistency one might expect from someone who is neither burdened with the language difficulties, nor haunted by the traumatic memories, that may hamper communication" between a government agent and a petitioner. Zubeda v. Ashcroft, 333 F.3d at 476 (3d Cir.2003).

The majority's thoughtful rejection of the IJ's adverse credibility determination here is yet another example of how even experienced IJs can place too much reliance on their own experiences in evaluating the testimony of petitioners from very different cultures. Although the IJ's credibility determination here does not rest upon Dia's demeanor, it is no less important to note that the IJ did not properly allow for differences between Dia's circumstances and the IJ's own in evaluating Dia's credibility. Rather, she failed to demonstrate any awareness of the context in which Dia's claim arose.

For example, the majority notes that the IJ rejected Dia's testimony about giving a $150 bribe to a Guinean police officer because she believed that the amount of the bribe was too small given the level of risk the police officer allegedly undertook on Dia's behalf. As my colleagues point out however, the amount of the bribe Dia said he gave is about a quarter of the average annual per capita income in Dia's country. See Maj. Op. at 252-53. Accordingly, $150 was a very substantial sum indeed.7 Moreover, the IJ's rationale assumes that the police officer actually exposed himself to a substantial risk in accepting the bribe. A more neutral assessment of this testimony would readily have lead to the realization that bribery is a way of doing business in some countries and that in such countries it is highly doubtful that any risk attends a police officer taking a bribe. The IJ's failure to realize this certainly ought to give us pause before assuming that IJs necessarily possess the kind of expertise in evaluating testimony of aliens that would insulate their conclusions from the bias I am concerned about. Such bias is only masked, not eliminated, if we uphold an adverse credibility ruling simply because we are told it rests upon the alien's demeanor with no further explanation.

I am, of course, aware of our limited standard of review when we adjudicate appeals in immigration cases. See INS v. Elias-Zacarias, 502 U.S. 478, 483-84, 112 S. Ct. 812, 117 L. Ed. 2d 38 (1992). We have wrestled with that narrow scope of review on more than one occasion when troubled by credibility rulings that appeared both unfair and unfounded. For example, we concluded that we were forced to affirm the IJ's ruling in Abdulrahman v. Ashcroft, 330 F.3d 587 (3d Cir.2003), even though the panel was very troubled by the IJ's findings. There, the IJ based her rejection of Abdulrahman's testimony on such "inconsistencies" as his assertion that he relied upon traditional herbal medicine administered by family members rather than visiting a hospital following alleged torture at the hands of government officials. 330 F.3d at 594. Given our standard of review, we were constrained to affirm and we rejected the alien's claim of bias. However, in doing so, we noted:

it must be added that there were places where the IJ did go beyond the bounds of propriety to make some additional and problematic generalized assertions of her own. While as discussed below we are understandably troubled by some of those comments, in the context of the record as a whole there is insufficient evidence to conclude that the overall proceedings were biased in violation of Abdulrahman's right to due process.

330 F.3d at 595. In his concurring opinion, Judge Becker commented on some of the more troubling aspects of the IJ's analysis. He explained:

The opinion of the Immigration Judge (IJ) is laden with statements such as the following, which I find troubling in terms of their viability as credibility judgments:

(1) "The respondent testified that he was treated with herbs, by his grandmother and mother, and told the Court these are the way things are done in Sudan, people do not go to the hospital as they do here in the Western World. Again, that is not the case, all countries all [sic] have hospitals and doctors, however, he wish [sic] to provide this false information regarding the medical institution about his country, so be it."

However, based upon available information about the Sudan, the Respondent's contention seems reasonable. At all events, the basis for the IJ's conclusion seems far from clear; rather, it seems quite tenuous.

Id. at 599-600.

The Immigration Judge's statements barely cross the line into the realm of fact finding, although Judge Shadur [author of the opinion] is correct that, in view of our extremely narrow standard of review, we are constrained to view them as so doing. While I join in Judge Shadur's opinion, I write separately to highlight these statements and to express my extreme discomfiture with them, as they border on the cavalier. Indeed, in my view, they come extremely close to constituting reversible error.

Id. at 600. Judge Becker's concurring opinion was joined by the entire panel.

However, even given our narrow scope of review, we still require that specific reasons be given for adverse credibility determinations so that we can review the BIA's decision. In Balasubramanrim, where the BIA made its own credibility ruling based upon the record before the IJ, we stated: "[t]he Board should give specific reasons for its determination that a witness is not credible. We must evaluate those reasons to determine whether they are valid grounds upon which to base a finding that the applicant is not credible. The reasons must bear a legitimate nexus to the finding." 143 F.3d at 162 (internal citations and quotation marks omitted). See also Mulanga v. Ashcroft, 349 F.3d 123, 131-32 (3d Cir.2003) ("Adverse credibility determinations are ... reviewed for substantial evidence.").

We should require nothing less than "specific reasons" for rejecting an alien's credibility when that assessment is based upon the alien's demeanor while testifying. Unless we require the IJ to explain those aspects of a witness's demeanor that undermine credibility, such as eye contact or similar factors that may be culturally determined, we simply can not afford the meaningful review the law requires.

In In re B—, 21 I & N Dec. 66, 1995 WL 326740 (BIA 1995) (Interim Decision), the BIA rejected the IJ's negative credibility assessment where that conclusion rested largely upon the alien's "tendency during his testimony to look down at the table or at the wall behind the interpreter instead of at the Immigration Judge" while testifying. 21 I & N Dec. at 70. The BIA explained its reasons for rejecting the IJ's negative assessment of that demeanor as follows:

Although, of course, we have not been able to observe this behavior by the applicant, we do not find that it necessarily indicates deception. Instead, it may indicate the applicant's concentration on the questions being asked of him through the interpreter. We note that the applicant seems to have been listening carefully, as the transcript contains about half a dozen instances where the applicant requested clarification of a question before he answered. These requests for clarification appear to have been conscientious attempts to provide the information sought by the questioner rather than attempts to evade answering.

Id. at 71.

Although the BIA's concern with the IJ's reliance on demeanor in In re B— does not implicate the cultural bias I am concerned about here, the BIA's opinion clearly shows the danger of placing too much reliance upon one person's interpretation of a witness's demeanor. It also demonstrates why we must not be content with allowing credibility determinations to rest upon "demeanor" with no further explanation by the IJ.

Requiring the fact finder to specify, and thereby think about and identify, those aspects of an alien's demeanor that are troubling will also enhance the quality of the entire process by affording IJs an opportunity to reflect upon perceptions that may simply reflect differing customs. Suggesting that adverse credibility rulings will be affirmed whenever they rest upon an alien's demeanor, with no further explanation or elaboration, will substantially undermine the process and open the door to no small amount of mischief.

I obviously I do not intend to suggest that all claims for relief under the immigration laws are valid or that petitioners do not sometimes fabricate testimony in order to avoid removal. Similarly, I do not minimize the difficulty of distinguishing valid claims from invalid ones at times. However, those difficulties are not resolved by unjustifiable deference to an IJ's unexplained interpretation of a witness's demeanor. Rather, they are exacerbated. Accordingly, for all of the reasons I have noted, I take this opportunity to express my concern with the import of footnote 23 in the majority opinion even though I join that portion of my colleague's analysis.8