Castañeda-Castillo v. Gonzáles

464 F.3d 112 | 1st Cir. | 2006

Lead Opinion

TORRUELLA, Circuit Judge.

Petitioners David Eduardo Castañeda-Castillo (“Castañeda”), his wife, and two daughters¹ ask us to review a decision of the Board of Immigration Appeals (“BIA”) denying their applications for asylum and withholding of removal. The BIA based its decision on a finding that the petitioners were barred from being granted asylum or withholding of removal because Castañeda had assisted or otherwise participated in the persecution of others on the basis of their political opinion. See 8 U.S.C. §§ 1158(b)(2)(A)® and 1231(b)(3)(B)®. After careful review, we grant the petition for review and reverse the decision of the BIA.

I. Facts and Procedural Background²

Castañeda is a native and citizen of Perú and was an officer in the Peruvian army. He and his family entered the United States with tourist visas at Miami, Florida, on August 19, 1991. They overstayed, and Castañeda applied for asylum in January 1993, claiming that he and his family had been persecuted by the terrorist group Sendero Luminoso (“Shining Path”) while they were in Perú.³ Castañeda was interviewed at the Immigration and Naturalization Service’s (“INS”)⁴ asylum office on May 19,1999. The asylum officer referred the application to an Immigration Judge after finding that Castañeda had not met his burden of proof for establishing eligibility for asylum. On July 7, 1999, the INS issued Notices to Appear (“NTA”) to Castañeda and his family, charging them with removability under Section 237(a)(1)(B) of the Immigration and Nationality Act (“INA”), 8 U.S.C. § 1227(a)(1)(B), for remaining in the United States for a time longer than permitted.

Beginning February 16, 2000, Castañeda and his family appeared at twelve different hearings before two different Immigration Judges.⁵ Castañeda was the only person to testify at the hearings.⁶ We glean the following facts from his testimony and the evidence presented at the hearings.

Castañeda joined the Peruvian military in 1979. In 1983, upon graduation from military academy, he was commissioned as a second lieutenant and assigned to a motorized infantry battalion. His original duties included teaching military training, such as the use of weaponry, handling of prisoners, community service, and guarding equipment. After twelve months, Cas-tañeda was transferred to another unit in Tumbes, a region in northern Perú near the border with Ecuador. Castañeda testified that, during this time, the military was fighting the Shining Path. Most of this fighting, however, occurred in the Ayacu-cho region in the Andes, approximately 400 miles to the south of Tumbes. While at Tumbes, Castañeda was a platoon commander and his duties included training and instructing the men in his section.

In January 1985, Castañeda was transferred to Battalion 34, an anti-terrorist battalion. Battalion 34 was located in the Ayacucho region, which is the birthplace of the Shining Path, and was referred to as the “emergency zone” due to the Shining Path’s presence in the region.⁷ The battalion was divided into different bases. Each base was led by a captain or lieutenant. According to Castañeda, each base had around forty men. Castañeda’s responsibilities increased in Battalion 34, as his duties included both training and leading troops out on patrols. Upon his arrival at Battalion 34, Castañeda was assigned to a base in Sacchaeamba, a rural mountainous area. Castañeda was the head of a twenty-man patrol, and went out on his first patrol one day after arriving at Sac-chaeamba.

Castañeda testified that a major whose code name was Wolf instructed him on what areas to patrol. Wolf, who was not at Castañeda’s base, would communicate the patrol area to the commander of the base by radio. The base commander would then relay that information to Cas-tañeda. Castañeda testified that, at times, the patrol route was very specific, while at other times the patrol route was general and it was left up to Castañeda to decide the route. Castañeda also testified that his patrol sometimes operated independently of other patrols, while at other times his patrol’s movements were coordinated with the movements of other patrols.

Castañeda testified that when he was on an independent patrol his base commander would inform him of other patrols’ positions via radio in order to avoid overlapping. He also testified that he sometimes made direct radio contact with other patrols, and that it was important to know the location of other patrols in the area in order to coordinate movement.⁸ When there were several patrols conducting a mission, Castañeda testified that each patrol leader responded to his base commander, who in turn responded to the battalion commander, who coordinated the mission with other base commanders.

According to Castañeda, before going out on any patrol, he was briefed on the mission of the patrol by the base commander. He was given an order that explained the situation on the ground, the possible location of the enemy, what the patrol would be doing on the mission, what route to take, and what equipment to carry. If, while on patrol, one of his men was injured, Castañeda was instructed to radio the base so that the man could be evacuated. He also had flares in the event that radio communication was down. Castañe-da testified that at times he had difficulty making radio contact due to interference caused by the dense terrain.

There were two types of patrols, combat and reconnaissance. Combat patrols engaged the enemy, while reconnaissance patrols gathered information. Most of the patrols were conducted at night in order to avoid detection by the Shining Path. Cas-tañeda testified that if, while on a reconnaissance patrol, he received intelligence that the Shining Path was in the area, he would immediately radio the base commander for instructions on what action to take. He could not engage the enemy without approval from his commander unless the patrol was being attacked or was in some sort of danger. Castañeda testified that, during his time at Sacchaeamba, he went on approximately twenty-five patrols, most of them reconnaissance. He never had any direct contact with the Shining Path.

Sometime in June or July of 1985, Cas-tañeda, while still in Battalion 34, was transferred to a base in Vilcashuan, where he remained the head of a patrol. He testified that the procedures and missions were the same as they had been at the base in Sacchaeamba.

Castañeda further testified that he was involved in a significant operation (the “Operation”) in August 1985, about a month after he arrived at the base in Vilcashuan. The Operation involved four patrols. Two patrols, code-named Links 6 and Links 7, were to enter a village named Lloellapampa in the Accomarca zone⁹ to search for Shining Path members. Links 7 was led by Sub-Lieutenant Telmo Hur-tado (“Hurtado”), and Links 6 was led by Lieutenant Riveri Rondón (“Rondón”). Two other patrols were assigned to block escape routes from the village.¹⁰ Castañe-da led one of these blocking patrols, codenamed Tiger. Castañeda could not remember the name of the leader of the other blocking patrol, code-named Wolf. According to Castañeda, he received his orders a few hours before his patrol left from his base commander, who received them from the command of the division located in the city of Humanaga. Castañe-da testified that he was briefed on the mission of the other patrols before he left. He was told that there were between forty and sixty Shining Path guerrillas in the village. Castañeda’s patrol took vehicles part of the way and went on foot the rest of the way. Castañeda and his men were armed with machine guns and a rocket launcher. They wore Peruvian military uniforms and also had masks on, both to avoid recognition by the Shining Path and to keep their faces warm. One of Castañe-da’s men had a radio to contact their base commander. Castañeda testified that he was near his radio operator during the Operation.

Castañeda’s patrol was the first to arrive at its destination. Once at the assigned location, which was three to five miles away from the village, Castañeda separated his men into three groups. They hid themselves about thirty meters from different sides of the path. Castañe-da testified that he was the one who would have made the decision whether to open fire if anyone came down the path. He testified that he was in radio contact with his base commander, who relayed any information Castañeda provided back to the division headquarters. Castañeda testified that he was not informed when the other patrols were ready to attack and he did not know when the attack began. Cas-tañeda testified that he was only able to communicate with his base; he did not have the frequencies for any of the other patrols and therefore could not hear any of their communications.¹¹ Castañeda also stated that none of the other patrols had his frequency, although each of the two attacking patrols had the other’s frequency.

Apparently, the Wolf blocking patrol got lost and never reached its assigned destination. Castañeda testified that the two other patrols, Links 6 and 7, entered the village and massacred civilians.¹² A subsequent report by a Peruvian Senate Human Rights Commission (“Human Rights Commission”) found that up to sixty-nine civilians were killed, including many women and children. According to Castañeda, he did not find out about the massacre until nearly three weeks after the Operation, when he heard that Hurtado had admitted to executing civilians. He stated that he heard this news on conventional radio. He testified that he did not learn exactly how many civilians were killed because different reports contained different estimates.¹³ According to Castañeda, he and his men stayed in their positions until his base commander told him the Operation was over and ordered him to return to base. Castañeda testified that he and his men never saw anyone come down the path and never fired any shots.¹⁴ Once back at base, he was debriefed.¹⁵

News of the massacre was eventually-reported by various media outlets and a formal investigation followed. Castañeda testified before a Peruvian Senate Human Rights Commission on September 26, 1985. He was represented by a military lawyer at the hearing. The Human Rights Commission concluded that what occurred during the Operation amounted to genocide. It noted that the probable number of people killed in the massacre was sixty-nine, but that it was hard to determine the exact number because many of the bodies had been destroyed by explosions from grenades. It also noted that Castaneda’s patrol, Tiger, was “not involved in any confrontations with fugitive civilians.”¹⁶ The Human Rights Commission found that Links 7, under Hurtado’s command, was the group that carried out the massacre.¹⁷

The massacre was also documented by the State Department’s 1985 Country Report on Human Rights Practices for Perú. The 1985 State Department Report states that “[a]n Army sub-lieutenant and three other officers were responsible for the Ac-comarca massacre of some 25 to 69 peasants. A separate investigation by the Senate Human Rights Commission supported these conclusions.” Castañeda admitted that he was one of the officers referred to in the 1985 State Department Report. The 1985 State Department Report indicated that the men involved in the massacre were charged in the military and civilian systems, and that the Peruvian Supreme Court would decide final jurisdiction.¹⁸

The Human Rights Commission recommended that those involved be tried in civilian — as opposed to military — court. However, the Peruvian Supreme Court decided that the military courts had jurisdiction. Castañeda, along with Hurtado, Rondón, and the officer who led the missing Wolf patrol, was charged in a military court martial in March or April 1986. Castañeda was charged with first degree murder, homicide, and abuse of authority. In connection with the court martial, Cas-tañeda appeared in a military court four or five times. He was represented by a civilian lawyer that he paid for. A judge presided over the proceedings. At the court martial, Castañeda testified in his own defense, answering questions from the judge and the prosecutor. He stated that he did not know if he had the right to examine other witnesses and did not remember if he was told that he had that right, but that no witnesses testified at any of the hearings. When asked by the government what evidence the prosecution presented against him, Castañeda stated that “[t]hey didn’t show me evidence regarding the patrol that I had.” Castañe-das evidence included the presentation of an operation chart that specified the position of each patrol.

Castañeda stated that he was found innocent of all charges. In support of this, he produced a document from the Supreme Council of Military Justice showing that the Appeals Division affirmed the dismissal of the charges against him.¹⁹ Cas-tañeda provided no other evidence related to his court martial. However, the State Department’s 1999 Country Report on Human Rights Practices for Perú indicates that Castañeda was in fact acquitted.

Hurtado was the only officer convicted by the military tribunal. Castañeda testified that Hurtado was convicted of “participating in the assassination of those village people.” The government’s attorney then asked “[s]o [Hurtado] was convicted of murder?,” to which Castañeda replied, “[y]es.” Castañeda also testified that Hur-tado was incarcerated in a military prison, but that he did not know the length of Hurtado’s sentence. When asked whether Hurtado was ever discharged from the military, Castañeda testified that he thought so, because that was the result required by Perú’s military code, but he could not be sure.²⁰

The government then introduced evidence — in the form of the State Department’s 1995 Country Report on Human Rights Practices for Perú — that Hurtado had reappeared on active duty and had been promoted to captain. When asked if he had heard about this, Castañeda stated “I did not know,” and later testified that it would surprise him to know that Hurtado had been promoted after he was convicted and sentenced.²¹ At this point, the government introduced the 1999 State Department Report, which indicated that Hurta-do had been convicted only for abuse of authority; had been promoted to captain at some point between the massacre and 1993, when he was finally sentenced; and had been released from prison under a general amnesty passed in 1995 by then President Alberto Fujimori and the Peruvian Congress. When asked if Hurtado had been convicted only for abuse of authority, Castañeda replied: “Clearly, I don’t know. What I manifested what I know is that the court found him guilty of that murder and I’m not an expert on the law, but maybe the murder is instead of the disobedience, sorry, the abuse of authority.”

Castañeda testified that he was aware of atrocities committed by the Shining Path against police officers, soldiers, and civilians. He admitted that he was aware that, prior to August 1985, the Peruvian military had at times engaged in extrajudicial killings, torture, and arbitrary detentions. When asked to explain, he stated that he had heard of officers who had detained terrorists or possible terrorists in an unjust manner, and had also heard of detained prisoners being executed. Castañe-da testified that he had not heard about any civilians being detained or killed. According to Castañeda, some of these actions by the Peruvian military occurred in the emergency zone, but it “was not the norm.” At one point, the government presented Castañeda with a portion from the 1985 State Department Report, which stated that “[tjotal deaths related to antiter-rorist operations of the security forces are estimated by human rights groups as being in the thousands for 1980-85.” Cas-tañeda testified that the number sounded incorrect.

Castañeda also testified that the Peruvian military had a good reputation for investigating and prosecuting military members for human rights violations. In an attempt to rebut this testimony, the government introduced evidence that in 2004 the current Peruvian Prime Minister had called for the annulment of previous rulings by the military justice system “that breach human rights, so that there can be in the future no further excuses or pretexts to avoid persecution.” Castañeda denied having any knowledge of these developments, and stated that he did not know whether any annulment of prior rulings would affect his case.

After the court martial hearings concluded, Castañeda returned to duty. He experienced no professional problems as a result of the charges brought against him. In January 1991, he was promoted to the rank of captain. However, due to the publicity surrounding the massacre and the court martial, Castañeda’s name and face had appeared in certain newspapers detailing the allegations against him. As a result, he and his family began to receive death threats and to experience other problems from the Shining Path. In April or May of 1986, Shining Path members set off explosives in front of the home of Cas-tañeda’s parents. Castañeda knew the Shining Path set off the explosives because they left pamphlets with the Shining Path’s seal. In June of 1986, Shining Path members allegedly shot at Castañeda and set off an explosive device while he was waiting for public transportation near his home. On other occasions, the Shining Path left propaganda containing written threats at the homes of both Castañeda and his parents. Castañeda stated in an affidavit that he had been threatened on over twenty occasions. In 1989, Shining Path members allegedly attempted to kidnap one of Castañeda’s daughters from school. Finally, in 1990, one of Castañe-da’s neighbors, who was also in the military and had been receiving threats from the Shining Path for several years, was murdered in his home and in front of his family by the Shining Path.

Following his neighbor’s murder, Cas-tañeda decided to take his family and leave Perú. He obtained tourist visas for himself and his family, retired from the military, and came to the United States.

On October 4, 2004, the Immigration Judge found that Castañeda was barred from applying for asylum and withholding of removal because he had assisted or otherwise participated in the persecution of others on account of their political opinion. In support of this determination, the Immigration Judge first found that Cas-tañeda was not credible. The Immigration Judge also found that, even if Castañeda were credible, he had assisted or otherwise participated in the persecution of others on account of their political opinion.

Castañeda appealed to the BIA, which affirmed the Immigration Judge on September 9, 2005. Writing its own opinion, the BIA affirmed the adverse credibility finding and stated that, even if Castañeda were credible, he had assisted or otherwise participated in the persecution of others. Castañeda has timely appealed to this court. He argues that the BIA’s adverse credibility determination is not supported by substantial evidence,²² that the BIA erred in disregarding the decision by the Supreme Council of Military Justice purporting to affirm the dismissal of charges against him, and that the BIA erred in finding that he had assisted or otherwise participated in the persecution of others on account of their political opinion.

II.

A. Standard of Review and Legal Standard

Under the INA, an alien is ineligible for asylum or withholding of removal if he or she “ordered, incited, assisted, or otherwise participated in the persecution of an individual because of the individual’s race, religion, nationality, membership in a particular social group, or political opinion.” 8 U.S.C. §§ 1158(b)(2)(A)(i), 1231 (b)(3) (B) (i). Under the applicable regulations, if evidence is produced showing that the alien engaged in such persecution, the alien must prove by a preponderance of the evidence that he or she did not engage in such persecution. 8 C.F.R. §§ 1208.13(c)(2)(ii), 1208.16(d)(2).²³

“Where the BIA does not adopt an IJ’s opinion but instead makes an independent, superceding decision, we review the decision of the BIA, and not that of the IJ.” Xu v. Gonzáles, 424 F.3d 45, 48 (1st Cir.2005). “We review factual findings and credibility determinations ... under the deferential substantial evidence standard.” Dhima v. Gonzáles, 416 F.3d 92, 95 (1st Cir.2005). Under this standard, we may not reverse the BIA’s factual findings “ ‘unless any reasonable adjudicator would be compelled to conclude to the contrary.’ ” Kheireddine v. Gonzáles, 427 F.3d 80, 88 (1st Cir.2005) (quoting 8 U.S.C. § 1252(b)(4)(B)). However, “credibility findings resting on analysis of testimony rather than on demeanor may deserve less than usual deference.” Cordero-Trejo v. INS, 40 F.3d 482, 487 (1st Cir.1994) (citation and internal quotation marks omitted). Finally, we review legal questions de novo. Ziu v. Gonzáles, 412 F.3d 202, 204 (1st Cir.2005).

B. Credibility Determination

1. The BIA’s Findings

The BIA gave five different reasons for its affirmation of the Immigration Judge’s adverse credibility determination: (1) that the military tribunals appeared to be fraudulent and Castañeda failed to submit documentary evidence of the proceedings beyond the document that showed his acquittal; (2) that Castañeda gave inconsistent testimony regarding whether or not he was in radio contact with the other patrols during the Operation; (3) that Cas-tañeda’s demeanor during questioning on radio communications indicated a lack of veracity; (4) that Castañeda was evasive regarding the extent of human rights violations by the Peruvian military; and (5) that Castañeda was not forthcoming about what happened to Hurtado. Having carefully reviewed these reasons, along with the record, we believe that the adverse credibility determination was not supported by substantial evidence. We address each reason in turn.

The BIA first faulted Castañeda for not providing documentary evidence of what happened at his military court martial other than the document — which the BIA found was not properly authenticated— from the Clerk’s Office of the Supreme Council of Military Justice showing that the Appeals Division had affirmed the dismissal of the charges against him. The BIA also stated that there was “evidence that the military tribunal was a way in which to grant impunity to the alleged persecutors who took part” in the Operation.

We begin by noting that the document from the Clerk’s Office of the Supreme Council of Military Justice is not the only evidence supporting a finding that Cas-tañeda was acquitted. The 1999 State Department Report specifically states that, except for Hurtado, “all the other defendants” were acquitted of any involvement in the massacre. Given Castañeda’s testimony and this documentary evidence, coupled with the lack of any evidence to the contrary, we conclude that it is uncontro-verted that he was in fact acquitted.²⁴

The Immigration Judge concluded — and the BIA agreed — that the military tribunals were a fraud on the basis of two documents: (1) the 1995 State Department Report, which stated that the military tribunal system often accorded impunity to the perpetrators of human rights abuses; and (2) an article from Latin American Weekly Report, dated August 31, 2004 (the “Article”).

We begin our discussion of the military tribunals by noting that the Supreme Council of Military Justice, along with its lower tribunals, is a duly constituted judicial body in Peru. This military system, which was granted jurisdiction over Cas-tañeda’s case by the Peruvian Supreme Court, acquitted Castañeda.²⁵ In general, American courts will give deference to decisions of foreign tribunals. See, e.g., Finanz AG Zurich v. Banco Económico, 192 F.3d 240, 246 (2d Cir.1999) (discussing deference to determinations made by foreign courts due to principles of international comity); Casey v. Dep’t of State, 980 F.2d 1472, 1477 (D.C.Cir.1992) (discussing deference to a determination of a foreign court in an extradition proceeding); Chiaramonte v. INS, 626 F.2d 1093, 1098 (2d Cir.1980) (discussing the principle of international comity in relation to criminal convictions obtained in a foreign country); Restatement 2d, Conflict of Laws § 98. In the instant case, the Immigration Judge and BIA chose not to give any weight whatsoever to Castañeda’s acquittal. That determination was not supported by substantial evidence.

As for the Article, it reported on a conflict between Perú’s civilian and military justice systems following the Peruvian Supreme Court’s ruling that the military system’s jurisdiction extended only to offenses committed in the furtherance of military duties (and not to human rights violations). It also contained a quote from the Peruvian Prime Minister calling for “the annulment of [military] rulings that breach human rights, so that there can be in the future no further excuses or pretexts to avoid prosecution.” We note that there is absolutely no indication that the Prime Minister considered the ruling in Castañeda’s case to have breached human rights. Further, the Article appears to only talk about events that occurred in the 1990s under the regime of former President Fujimori. The Article does not make any allusion to events that occurred in the 1980s. Finally, although the Peruvian Prime Minister at the time of the Article called for the annulment of certain rulings, there is no evidence whatsoever that any rulings were annulled, much less Castañe-da’s acquittal. In sum, there is no indication that the Article is in any way applicable to Castañeda’s particular case, nor that the Peruvian Prime Minister’s statements were anything more than pure rhetoric. We do not think that the Article was enough to overcome the deference ordinarily due a decision by a duly constituted foreign tribunal. In fact, it is nothing short of anomalous that undue weight is given to a magazine “article”, yet no weight is given to a finding of a duly constituted tribunal.

Regarding the 1995 State Department Report, we note that it does mention that Hurtado had reappeared on active duty after serving some jail time. However, other parts of the record indicate that in 1995, then-President Fujimori and Peril’s Congress passed a law extending amnesty to military members convicted of human rights abuses between 1980 and 1995. Thus, it appears that Hurtado was released from jail because he had been granted amnesty, not because the original trial and conviction were somehow fraudulent. We certainly do not see how a grant of amnesty to Hurtado several years after Castañeda left Perú should affect Castañe-da’s original acquittal, nor do we think that it should overcome the deference we usually give to the judgment of a foreign tribunal. We therefore conclude that the BIA’s finding that Castañeda’s trial was a fraud is not supported by any evidence, much less substantial evidence.

Second, the BIA found that Castañeda “gave inconsistent testimony regarding whether and when he was in radio contact with the other patrols involved” in the Operation. According to the BIA, Cas-tañeda indicated that it was important to have radio contact, but later specifically testified that he was not in contact with other patrols on the day of the massacre. The BIA also rejected Castañeda’s argument that the Immigration Judge had gone beyond the area of its expertise and “ventured into the realm of Peruvian military tactics,” because Castañeda’s own testimony made it “clear that radio contact was important for purposes of coordination as well as efficiency and protection.”

There are several problems with the BIA’s reasoning. First, while Castañeda testified early on that it was important to have radio contact, his testimony indicates that he was referring to radio contact with his base, not other patrols. Castañeda testified that he only sometimes had contact with other patrols. Further, Castañe-da specifically testified about the coordination of patrols for a mission:

Each head of a patrol will respond to the immediate boss, the head of base. The head of base at the attack, they will inform the command of the battalion who can communicate and coordinate directly with the head of patrol and give another specific order or can get contact among heads of patrols.

In other words, Castañeda testified at the outset — before he was asked about whether he had any radio contact with other patrols the day of the Operation — (1) that it was important to have radio contact with his base commander, (2) that in coordinating attacks, the base commander would contact a battalion commander who coordinated the various patrols, and (3) that he only sometimes had direct contact with other patrols. We therefore see no inconsistency in Castañeda’s testimony that, while he had radio contact with his base commander during the Operation, he did not have radio contact with the other patrols on the day in question.

It also appears from the record that the Immigration Judge, and perhaps the BIA, simply found this testimony illogical. Such a finding deserves less deference than other types of credibility findings because it is based on an analysis of the testimony. Cordero-Trejo, 40 F.3d at 487. To the extent that the BIA’s finding rested on this ground, we would agree with Castañe-da’s point that the BIA ventured outside the area of its expertise and into the realm of Peruvian military tactics. As such, we think the BIA’s finding on this point is not supported by substantial evidence.

Third, the BIA accorded weight to the Immigration Judge’s finding that during testimony about radio communication, Cas-tañeda was “blinking his eyes in an unusually rapid rate.” While demeanor evidence is peculiarly within the purview of the trier of fact, we are not persuaded — given the problems we have found with the other aspects of the credibility findings — that rapid blinking alone is enough to find Cas-tañeda not credible.²⁶

Fourth, the BIA affirmed the Immigration Judge’s finding that Castañeda was evasive regarding the extent of human rights violations committed by the Peruvian military. From our review of the record, this finding mischaracterizes Castañe-das testimony. He was not evasive when confronted with statistics from State Department Reports. Castañeda readily admitted that he was aware of instances in which the Peruvian army had engaged in extrajudicial killings, disappearances, torture, and arbitrary detentions. He simply disagreed with the numbers. Castañeda consistently testified that abuses by the Peruvian army were not the norm and that only a minority of the military engaged in such practices. When confronted with the 1985 State Department Report indicating that “[tjotal deaths related to antiterrorist operations of the security forces are estimated by human rights groups as being in the thousands between 1980-1985,” Cas-tañeda took issue with those numbers. We note that, to the extent the government offered the report as evidence that thousands of people had been killed due to abuses by the Peruvian military, Castañe-da appears to have been at least correct. In a portion of the report the government failed to read at the hearing, the State Department noted that, out of these thousands of total deaths, “[pjrecise information is lacking on the number of civilian victims and their identities (terrorist or non-terrorist) as well as the identity of their assailants (terrorists, security forces, or peasant self-defense forces).” In other words, the “thousands” of deaths that resulted from antiterrorist operations included deaths of terrorists, as well as civilians killed by terrorists and other peasant forces. Therefore, it is quite likely that there were not thousands of deaths related to abuses by the Peruvian military, and in that sense, Castañeda was correct.²⁷

Having considered the entire record, we do not think that substantial evidence supports the BIA’s determination that Castaneda was evasive on the issue of human rights abuses by the military. He testified consistently throughout the hearings before the Immigration Judge, and although he disagreed at times with the government’s assertions, this can hardly be considered evasive testimony. We therefore find that the BIA’s finding on this issue is not supported by substantial evidence.

The fifth reason for the BIA’s adverse credibility finding was that Castañeda allegedly was “not forthcoming about the judgment entered against” Hurtado. This, again, is a mischaracterization of Castañe-da’s testimony. There are several aspects of Castañeda’s testimony on the subject of Hurtado, and we deal with each in turn.

a. Whether Hurtado was imprisoned

During cross-examination, the government’s attorney asked Castañeda what happened to Hurtado. Castañeda replied, “[ajfter [Hurtado] admitted [to his involvement in the massacre] and the court found him guilty, he was placed in a military prison.” This statement is borne out by the record; all of the documents indicate that Hurtado was found guilty and placed in prison for a period of time.

b. Hurtado’s conviction

Later on in questioning, the following exchange occurred:

Q. Sir, my question is, what was [Hur-tado] found guilty of?

A. He was found guilty of participating in the assassination of those village people.

Q. So he was convicted of murder?

A. Yes. That’s what I know.

Q. Was he sentenced to prison?

A. Yes. He was recluded in a military prison.

Q. Do you know what his sentence was?

A. I do not know the amount of years.

The latter part of Castañeda’s answer is supported by the record, as the 1999 State Department Report indicates that Hurtado was not finally sentenced until 1993, two years after Castañeda left Perú. It is perfectly plausible that Castañeda did not know the length of Hurtado’s sentence, since Hurtado had not been sentenced when Castañeda and his family left Perú.

Regarding the first part of Castañeda’s response, the government later asked Cas-tañeda if Hurtado was convicted only of abuse of authority. Castañeda stated that “[c]learly, I don’t know. What I manifested what I know is that [t]he court found him guilty of that murder and I’m not an expert on law, but maybe the murder is instead of the disobedience, sorry, the abuse of authority.” On appeal, the government argues that Castañeda testified inconsistently about what crimes Hurtado was convicted of. This is incorrect, for two reasons. First, Castañeda testified that Hurtado was convicted of killing the villagers. He referred to the killings as “that murder” and an “assassination.” It is obvious that, in Castañeda’s untrained legal mind, Hurtado murdered civilians, and this is what he was convicted of. Second, Hurtado was in fact charged with murder, and, at the time of Castañeda’s acquittal, the case against Hurtado was still proceeding. The document confirming Castañeda’s acquittal specifically states that the case would continue against Hur-tado “for crimes against life, body, and health, first degree murder.” Hurtado, however, was not sentenced until 1993, and was sentenced only for abuse of authority. This again occurred after Castañeda had left Perú. We see no reason to find that Castañeda was not credible as to his knowledge and participation in the Operation simply because he thought Hurtado was convicted of murder and not abuse of authority, when (1) Hurtado was originally charged with murder, (2) still faced a murder charge at the time of Castañeda’s acquittal, and (3) was not sentenced until two years after Castañeda had left the country.

c. Whether Hurtado was discharged or promoted

When asked if Hurtado was discharged from the army, Castañeda testified that he thought so, since that was the usual military practice for an officer in such a situation, but did not know for sure. This apparently was not good enough for the government attorney, who commented that Castañeda seemed to not know many details about what happened to Hurtado given the significance the court martial had on his life. Castañeda replied

You’re right. That is a very important — in my life. What I know is that that officer acted outside the rules that are breach of the army. He was found guilty.... I did not try to find out what privileges he might have had and what had happened to his life.

In other words, Castañeda testified that he knew Hurtado went to prison, thought he was discharged from the army, and did not try to find out what happened to Hurtado afterwards. Instead, like many people would, he tried to resume Ms normal life. We see nothing to indicate that Castañeda was not forthcoming about Hurtado’s fate.²⁸

During cross examination, the government also introduced a 1995 State Department Report indicating that there were reports that Hurtado had reappeared on active duty and been promoted to captain. When asked about this, Castañeda testified “I did not know that.” When asked if it would surprise him to know that Hurta-do was promoted after his conviction and sentence, Castañeda answered that it would. The record supports Castañeda’s testimony. First, the 1999 State Department Report indicates that “the Military Code states that any conviction that entails a sentence of 2 or more years’ imprisonment must result in the officer’s immediate discharge.” Thus, Castañeda was correct in his testimony that the usual practice was for an officer in Hurtado’s situation to be discharged. The 1999 Report also indicates that Hurtado was not released from prison until 1995 under the general amnesty, four years after Castañeda had left Perú. It therefore is entirely consistent with Castañeda’s testimony that he had no idea Hurtado had not been discharged and had in fact been released from prison and promoted, because when Castañeda left Perú, Hurtado’s case was still ongoing and he was in prison. Again, we find no basis for the adverse credibility finding.

We also wish to emphasize several key-pieces of evidence that supported Castañe-da’s story. First, as we noted above, the 1999 State Department report indicates that Castañeda was acquitted of any involvement in the massacre. Second, the report from the Human Rights Commission specifically found that Castañeda’s patrol was “not involved in any confrontations with fugitive civilians.” This is an extremely important piece of evidence because it goes to the heart of the issue: the extent of Castañeda’s participation in the massacre of the villagers. Further, while the Commission report mentioned Cas-tañeda’s name twice, it did so only in reference to the fact that he led one of the non-infringing patrols. The report clearly holds Hurtado to blame for the massacre. It mentions Hurtado’s testimony in depth and also refers to him as “the perpetrator” — as opposed to “one of the perpetrators” — of the massacre. The 1995 and 1999 State Department Reports single out Hurtado and his patrol as the ones responsible for the massacre. For example, the 1999 State Department Report states that “Hurtado commanded the army unit responsible for the 1985 Accomarca massacre.” Neither Castañeda nor any of the other patrol leaders’ names are ever mentioned. In our view the record, when viewed as a whole, indicates that Hurtado and his patrol were responsible for the massacre, and that Castañeda’s patrol, which was three to five miles distant, took no part in it. After a trial by the military court, Hurtado — the man responsible— was convicted²⁹ and the charges against Castañeda were dismissed. The fact that Hurtado was later released due to a general declaration of amnesty by the President and Congress should not affect the legitimacy of the original trial and conviction, and it especially should not be held against Castañeda.

In sum, the evidence indicates that Cas-tañeda credibly testified that (1) he never had any contact with any villagers during the Operation, (2) he was acquitted of any involvement in the massacre by a duly constituted body of the Peruvian military, which was given jurisdiction over his case by the Peruvian Supreme Court, and (3) Hurtado, the leader of the Links 7 patrol, was the one found responsible for the massacre by a military court. This evidence, combined with the problems we have discussed related to the reasons relied upon by the BIA in making its adverse credibility finding, sound the death knell for that finding. Substantial evidence does not support a finding that Castañeda was not credible regarding his involvement in the massacre, and we therefore reverse the BIA’s credibility determination.

2. The “Ordinary Remand Rule”

The dissent, invoking the “ordinary remand rule” argues that we must remand the case to the BIA for further proceedings. We reject the dissent’s reasoning for several reasons which we briefly explain.

In INS v. Ventura, 537 U.S. 12, 123 S.Ct. 353, 154 L.Ed.2d 272 (2002), a Supreme Court case invoked by the dissent, the Court reversed a judgment of the Ninth Circuit. The Ninth Circuit had reversed the BIA based on a “changed circumstances” argument that the government had made to an Immigration Judge, but which the BIA had never addressed. On appeal, the Supreme Court stated that, under the ordinary remand rule, “[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.” Id. at 16, 123 S.Ct. 353 (internal citations and quotation marks omitted). The Court then reversed, stating that the Ninth Circuit had erred by deciding the “changed circumstances” issue “without giving the BIA the opportunity to address the matter in the first instance.” Id. at 17, 123 S.Ct. 353.

The Supreme Court has also recently applied the ordinary remand rule in Gonzales v. Thomas, — U.S. -, 126 S.Ct. 1613, 164 L.Ed.2d 358 (2006). In Thomas, the Immigration Judge denied a petitioner’s asylum claim without considering her alleged fear of persecution based on membership in a particular social group. The BIA then summarily affirmed. The Ninth Circuit reversed, finding that Thomas was in fact persecuted as a member of a social group even though the BIA had not yet addressed the issue. On appeal, the Supreme Court reversed and remanded to the BIA for consideration of the issue, stating that “[t]he agency has not yet considered whether [petitioner’s father-in-law’s] family presents the kind of kinship ties that constitute a particular social group.” Id., 126 S.Ct. at 1615.

While we do not quarrel with the general legal proposition regarding the ordinary remand rule advanced by the dissent and found in Thomas and Ventura, we believe that the facts of this case take it outside of the purview of that rule. In both Thomas and Ventura, the court of appeals considered an issue that the BIA never addressed in the first instance. Here, by contrast, the Immigration Judge and BIA both thoroughly considered the credibility issue based on a fully developed record. In reversing the credibility determination, we are not conducting a de novo review of the record, but rather are reviewing the reasons set forth by the BIA in light of the record. Under the ordinary remand rule, while we could remand the case for further factfinding, we are not required to do so since the BIA has already thoroughly “evaluated] the evidence” and brought “its expertise to bear upon the matter.” Ventura, 537 U.S. at 16, 123 S.Ct. 353.³⁰

We note that this approach is in accord with other circuits that have interpreted Ventura and Thomas. For example, the Seventh Circuit has stated that “we do not agree that Ventura stands for the broad proposition that a court of appeals must remand a case for additional investigation or explanation once an error is identified.” Ghebremedhin v. Ashcroft, 392 F.3d 241, 243 (7th Cir.2004). Several other circuits have adopted a similar line of reasoning. See, e.g., Almaghzar v. Gonzáles, 450 F.3d 415, 423 n. 11 (9th Cir.2006) (“Neither Ventura nor Thomas requires us to remand an issue to the agency when the agency has already considered the issue.”); Zhao v. Gonzáles, 404 F.3d 295, 311 (5th Cir.2005) (stating that the court was not required to remand under the ordinary remand rule because the BIA had already decided the issue).

The dissent also argues that our decision to reverse violates the law of our own circuit. However, the cases cited by the dissent are clearly distinguishable. For example, the dissent relies on Gailius v. INS, 147 F.3d 34 (1st Cir.1998) and Corde-ro-Trejo v. INS, 40 F.3d 482 (1st Cir. 1994). In Gailius, the Immigration Judge did not address whether alleged threats had occurred as the petitioner described because it found that a reasonable person in the petitioner’s situation would not fear persecution on account of such threats. We disagreed with this latter finding and stated that “[i]f the threats are real ... Gailius may well have had a well-founded fear.” Gailius, 147 F.3d at 44. We remanded so that the BIA — which ignored the evidence of the threats based on reasons not having to do with credibility— could explain why it found the threats not credible. One of our main reasons for remand was that “the IJ did not make a determination of the credibility and authenticity of Gailius’ evidence of threats.” Id. at 47. We have a different case here, because the IJ and BIA made specific credibility findings; those findings simply are not supported by substantial evidence. In Cordero-Trejo, the BIA first found the petitioner not credible, then made an alternative holding that the petitioner was ineligible for asylum based on three reasons. Cordero-Trejo, 40 F.3d at 485. We found two of these reasons invalid under the INA, and the third depended on the adverse credibility finding, which we found was not supported by substantial evidence. Id. at 485, 488. We also found that the BIA failed to consider certain general country condition evidence offered by the petitioner. Id. at 485. In reversing the adverse credibility finding, we noted that “it is difficult to ascertain from what sort of evidence the IJ drew his credibility conclusions.” Id. at 491. Given this fact, along with the fact that the Immigration Judge and BIA had failed to address any of the petitioner’s documentary evidence, we remanded. Castañeda’s case is obviously distinguishable because it is clear what evidence the Immigration Judge and BIA relied on in drawing the adverse credibility determination, and the BIA did not simply ignore all of Castañeda’s documentary evidence³¹

C. Whether Castañeda Was a Persecutor

As we noted above, the Immigration Judge also found that, even assuming Castañeda had testified credibly — i.e., even assuming that he and his patrol were several miles away from the village during the Operation, did not encounter a single villager, and did not know what was happening in the village or have any control over what was taking place there — he still “assisted or otherwise participated” in the persecution of others. Specifically, the Immigration Judge concluded that

the objective effect of [Castañeda’s] participation in the Accomarca operation would have been to aid even in some small measure in the confinement of villagers to the combat area and therefore aid even in some small measure with the execution and murder and rape of those unarmed men, women, and children.

The BIA, citing to the above portion of the Immigration Judge’s opinion as well as In re Rodríguez-Majano, 19 I. & N. Dec. 811 (BIA 1988), affirmed this finding, stating that

we find no error in the Immigration Judge’s conclusion that [Castañeda] failed to meet his burden of proving that he did not assist or otherwise participate in the persecution of villagers in Acco-marca based on their political opinion; even if [Castañeda] testified credibly that he did not harm or execute anyone on August 8, 1985, there is substantial evidence to support the Immigration Judge’s conclusion that [Castañeda] did aid in the persecution of others by helping to confine them, inasmuch as his patrol was charged with blocking an escape route for the villagers and people located in the emergency zone.³²

An alien is ineligible for asylum or withholding of removal if he or she “ordered, incited, assisted, or otherwise participated in the persecution of an individual because of the individuars race, religion, nationality, membership in a particular social group, or political opinion.” 8 U.S.C. §§ 1158(b)(2)(A)®, 1231(b)(3)(B)®. In the instant case, it is undisputed that Castañe-da did not order or incite the persecution of anyone. It is also undisputed that the villagers massacred during the Operation were killed because of their political opinion, i.e., their support of or suspected involvement with the Shining Path. Therefore, the only issue is whether Castañeda “assisted or otherwise participated in” the persecution of the villagers.

Castañeda’s main argument is that he could not have “assisted or otherwise participated in” persecution because the massacre took place without his control and awareness, and because he had no connection to the massacre beyond being a member of the military at the time it occurred.³³ The government, following the reasoning of the BIA, argues that Castaneda’s actions in blocking an escape path furthered the massacre and therefore demands a finding that he assisted or otherwise participated in persecution. After careful review of the record and various cases interpreting the phrase “assisted or otherwise participated in” the persecution of others, we are compelled to reverse the BIA’s decision in this respect.

We begin our discussion by looking to the Supreme Court’s decision in Fedorenko v. United States, 449 U.S. 490, 101 S.Ct. 737, 66 L.Ed.2d 686 (1981).³⁴ In Fedoren-ko, the Court interpreted a provision of the Displaced Persons Act of 1948, Pub.L. No. 80-774, 62 Stat. 1009 (“DPA”), which excluded from the definition of “displaced persons” individuals who had “assisted the enemy in persecution of civilians] or had ‘voluntarily assisted the enemy forces ... in their operations ... ’ ” Fedorenko, 449 U.S. at 495, 101 S.Ct. 737 (internal quotation marks omitted) (alteration in original).

Fedorenko was admitted to the United States as a displaced person following World War II after stating in his application that he had been a farmer in Poland from 1937 until March 1942, then had been deported to Germany and forced to work in a factory until the end of the war. In reality, Fedorenko was a Russian soldier who had been captured by the German army and forced to serve as a guard at the Nazi concentration camp in Treblinka, Poland, where several hundred thousand Jewish civilians were murdered. When the falsehoods on his application were discovered, the government filed an action to have Fedorenko’s citizenship revoked. At the resulting trial, Fedorenko admitted to his service at Treblinka and admitted that he knew that thousands of Jewish civilians were being murdered there, but claimed that his service as a guard was involuntary and that he did not have any personal involvement in the atrocities committed. Fedorenko also admitted that he had “shot in the general direction of escaping inmates” during a prison uprising, id. at 500, 101 S.Ct. 737, “that the Russian armed guards significantly outnumbered the German soldiers at the camp, that he was paid a stipend and received a good service stripe from the Germans, and that he was allowed to leave the camp regularly but never tried to escape.” Id.

The Court found that Fedorenko had assisted in the persecution of civilians. In doing so, the Court rejected Fedorenko’s argument that he did not assist in any persecution because his actions were involuntary. It noted that the DPA incorporated the definition of “displaced person” contained in the Constitution of the International Refugee Organization of the United Nations (“IRO Constitution”). Id. at 496 n. 3, 101 S.Ct. 737. While section 2(a) of the IRO Constitution provided that persons who “assisted the enemy in persecuting civil populations” were excluded from the definition of “displaced persons,” section 2(b) of the IRO Constitution excluded only those persons who “voluntarily assisted the enemy forces ... in their operations.” Id. at 496 n. 4, 101 S.Ct. 737. The Court found that “[ujnder traditional principles of statutory construction, the deliberate omission of the word ‘voluntary’ from § 2(a) compels the conclusion that the statute made all those who assisted in the persecution of civilians ineligible for visas.” Id. at 512, 101 S.Ct. 737 (emphasis in original).

After rejecting Fedorenko’s involuntariness defense, the Court found that his actions constituted assistance in persecution. In an oft-cited footnote, the Court described a continuum of conduct along which an individual’s actions should be considered in determining whether he or she assisted in persecution:

Thus, an individual who did no more than cut the hair of female inmates before they were executed cannot be found to have assisted in the persecution of civilians. On the other hand, there can be no question that a guard who was issued a uniform and armed with a rifle and a pistol, who was paid a stipend and was regularly allowed to leave the concentration camp to visit a nearby village, and who admitted to shooting at escaping inmates on orders from the commandant of the camp, fits within the statutory language about persons who assisted in the persecution of civilians. Other cases may present more difficult line-drawing problems but we need decide only this case.

Id. at 512 n. 34., 101 S.Ct. 737

In cases involving “more difficult line-drawing problems,” courts have relied on Fedorenko’s “continuum of conduct” footnote and BIA precedent to determine whether an alien “assisted or otherwise participated in” persecution. Following Fedorenko, the BIA has held that

The participation or assistance of an alien in persecution need not be of his own volition to bar him from relief. However, mere membership in an organization, even one which engages in persecution, is not sufficient to bar one from relief, but only if one’s action or inaction furthers that persecution in some way. It is the objective effect of an alien’s actions which is controlling.

Rodríguez-Majano, 19 I. & N. Dec. at 814-15; see also Miranda Alvarado v. Gonzales, 449 F.3d 915, 920 (9th Cir.2006); Higuit v. Gonzáles, 433 F.3d 417, 421 (4th Cir.2006) (distinguishing between “ ‘genuine assistance in persecution and inconsequential association with persecutors’ ”) (quoting Singh v. Gonzáles, 417 F.3d 736, 739 (7th Cir.2005)).

In determining the objective effects of an alien’s actions, a court must examine the totality of the alien’s relevant conduct. See Miranda Alvarado, 449 F.3d 915, 918; Xie v. INS, 434 F.3d 136, 142-43 (2d Cir.2006) (stating that the court “focused on the nature of [the alien’s] conduct as a whole”); In Re A-H, 23 I. & N. Dec. 774, 785 (A.G.2005) (citing with approval Hernández v. Reno, 258 F.3d 806, 814 (8th Cir.2001) (stating that “[t]he Board should have analyzed all the pertinent evidence related to [the alien’s] conduct”)). Further, some degree of individual culpability is necessary in order for an alien to be held responsible for assisting or participating in persecution. See Miranda Alvarado, 449 F.3d 915, 918 (stating that “determining whether a petitioner ‘assisted in persecution’ requires a particularized evaluation of both personal involvement and purposeful assistance in order to ascertain culpability”); Higuit, 433 F.3d at 421 (finding that “there can be no dispute over Higuit’s personal culpability in this case”); Hernández, 258 F.3d at 814 (stating that a court should “determine whether the individual should be held personally culpable for his conduct”). Moreover, several courts have explicitly required that a petitioner’s actions have direct (as opposed to tangential) consequences for the victims or further the persecution in some material way. See Miranda Alvarado, 449 F.3d 915, 918 (asking whether the petitioner’s actions furthered the persecution or were tangential to it, and finding that petitioner was barred because, without his actions, the events leading to the persecution could not have proceeded); Xie, 434 F.3d at 143 (stating that the court would find assistance in persecution when the alien’s conduct “was active and had direct consequences for the victims,” but would not find assistance in persecution where the alien’s conduct was tangential and passive). Some courts have also considered the repetitive nature of the alien’s actions. See, e.g., Miranda Alvarado, 449 F.3d 915, 919; Singh, 417 F.3d at 740 (in discounting petitioner’s “protestations of case-specific ignorance” of persecutory acts, noting “the repetition of [petitioner’s] conduct over an extended period”).³⁵

Turning now to Castañeda’s particular case, we conclude that substantial evidence does not support the BLA’s determination that he “assisted or otherwise participated in” persecution. Taking Castañeda’s testimony as true, we are presented with the following set of facts: (1) the uncontradict-ed evidence is that the purpose of the Operation was directed at the Shining Path guerrillas in the village, not civilians; ³⁶ (2) Castañeda and his men were hidden in the jungle three to five miles from the village and about thirty meters from the path they were guarding; (3) during the time they were there, they did not see or hear anyone, nor did they fire any shots; (4) they never entered the village and never directly or indirectly participated in the killings of the villagers, which was carried out by a rogue patrol led by Hurtado in a previously unplanned manner; (5) Castañeda and his men lacked any control or authority over this rogue patrol; and (6) they did not know what occurred in the village during the Operation.

Given these facts, and considering the totality of relevant conduct, we are compelled to conclude that Castañeda did not assist or otherwise participate in the persecution of the villagers. First, Castañe-da’s personal involvement in the murder of the villagers was non-existent. Compare Miranda Alvarado, 441 F.3d at 762 (noting that petitioner, who served as an interpreter for the Peruvian army while officer tortured suspected Shining Path member, was personally involved in the persecution because he was present and active during the interrogation); Singh, 417 F.3d at 740 (finding that the petitioner had assisted or otherwise participated in persecution where the petitioner “took innocent Sikhs into custody ... and transported them to the police station, where he knew they would be subjected to unjustified physical abuse[,]” and “participated in raids on the homes of innocent Sikh families”). Second, Castaneda’s conduct — which amounted to hiding in the jungle for several hours three to five miles away from the scene of the massacre — can hardly be characterized as “active and halving] direct consequences for the victims” of the persecution, Xie, 434 F.3d at 143, especially since no one ever came down the path Castañe-da’s patrol was watching. Instead, Cas-tañedas actions are better characterized as irrelevant “to the acts of oppression and passive in nature.” Id. In a similar vein, Castañeda’s actions simply had no “objective effect” on the persecution because Castañeda and his men never saw anyone or fired any shots.³⁷ In other words, the massacre of the villagers during the Operation would have occurred regardless of whether Castañeda ever reached his assigned position,³⁸ and Castañeda did nothing to facilitate the massacre in any way because his patrol never encountered anyone.³⁹ Cf. Miranda Alvarado, 449 F.3d 915, 927 (finding that, without the petitioner’s actions, the persecution would not have gone forward). Add to these facts Castaheda’s lack of awareness of what was happening in the village, and we simply do not see how his actions rise to the level of culpability necessary to find that he assisted or otherwise participated in the persecution of the villagers. We therefore find that substantial evidence does not support the BIA’s determination in this regard.

D. The Dissent

This court has never condoned, nor will it ever condone, the massacre of innocent civilians by the military in the course of its war-related activities. See United States v. Zajanckauskas, 441 F.3d 32 (1st Cir.2006). However, we also cannot tolerate a finding that amounts to a conclusion that someone is guilty of such conduct by remote association. The findings of the Immigration Judge and the BIA amount to just that: a conclusion of guilt by remote association. These findings fly against the evidence that is relevant to a determination of whether Castañeda was credible and to whether he was in any way responsible for the atrocities committed by a different Peruvian army unit, operating independently of his command, several miles away from his location, with nothing but mountains and jungle in between.

Because the findings below — which we have already addressed as unsupported or unsupportable — are clothed in the presumably impervious mantle of credibility findings, and considering the assertions in the dissent that follows, it is appropriate that we conclude by re-emphasizing certain uncontested evidence which goes both to the issue of Castaneda’s credibility and to the question of his responsibility for the heinous acts committed beyond his control or participation.

First, it is undisputed that the patrol which committed the massacre of the villagers was several miles away and acted independently from Castañeda, whose assigned duty was to watch a trail to intercept Shining Path guerrillas, not civilians. The dissent’s phrasing of the facts, that “army units fired and threw grenades into the houses and killed many women and children,” and that Castañeda “headed one of four army units involved in the military operation during which this massacre occurred” is unfortunate because it gives the impression that the military operation was one that included, as a pre-planned goal, the premeditated killing of civilians. In fact, there is absolutely no evidence to this effect. This unwarranted implication in the quoted portion of the dissent is reenforced by the phrasing that follows: “His patrol was assigned to and did block a possible escape route from the village, but there is no evidence that he was in the village during the massacre or that he personally killed anyone.” This quote omits the rather important fact that Cas-tañeda was several miles away from the village, that there is no evidence that he ever entered the village before, during or after the massacre, or that anyone, civilian or otherwise, ever went down the trail he was guarding.

Second, as we just noted, it is undisputed that nothing happened on Castañeda’s watch. It is undisputed that no one came down the trail — no Shining Path guerrillas, no civilians, no one. Thus, neither Cas-tañeda nor anyone under his command engaged in any action against innocent civilians.

Third, there is not a scintilla of evidence that Castañeda was aware, either before or during the massacre, that these acts would be committed or were being committed by others over which he had no control and were in fact not his subordinates. The issue regarding radio communication during the Operation is simply another red herring, for several reasons. First, there is no affirmative evidence of any communication between Castañeda and the other patrols during the Operation. Second, radio silence during operations is not an unheard of military tactic. In any event, the only evidence is the assertion by the only person there, Castañeda, that there was no communication between the patrols. The conclusions to the contrary, used by the Immigration Judge and the BIA as a basis for their credibility conclusions, are mere speculation unsupported by evidence, especially given Castañeda’s prior testimony about communication up the chain of command. Third, the issue regarding radio communication is an irrelevant distraction from the main issue: Castañeda’s culpability for the actions of Hurtado’s rogue patrol, over which it is undisputed that Castañeda lacked control or command.

Finally, Castañeda was acquitted by a Peruvian military court for his involvement in the Operation. This is the same military court that found the captain in charge of the rogue patrol, Hurtado, guilty of perpetrating the massacre against the civilians. Notwithstanding these undisputed facts, and that such tribunals are the ones that normally deal with crimes allegedly committed by the military in the course of military operations, the Immigration Judge and the BIA latched on to irrelevant issues to detract from the most central, irrefutable fact: that the only adjudicative body to pass upon this unfortunate incident placed the blame for these actions on someone other than Castañeda.⁴⁰

III.

Because we find that substantial evidence does not support the BIA’s determinations that Castañeda was not credible and assisted or otherwise participated in persecution, we grant the petition for review, reverse the decision of the BIA, and remand the case so that an Immigration Judge may consider the merits of Castañe-da’s asylum and withholding of removal applications.

Castañeda’s asylum request merits serious, unbiased consideration. As we pointed out earlier in this opinion, the Shining Path have been catalogued by our government as a ruthless terrorist organization. Castañeda’s unchallenged testimony is that Shining Path members threatened him on numerous occasions and attempted to murder him in 1986. Furthermore, and in keeping with its record of indiscriminate violence, see, e.g., Philip Bennett, Pol Pot in Peril, New Republic, January 28, 1985, at 16, Shining Path members tried to kidnap Castañeda’s daughter in 1989. The group also murdered Castañeda’s neighbor and military colleague in front of the man’s family in 1990. Thereafter, Castañeda sought refuge for himself and his family from these terrorist thugs by leaving his native country.

The BIA’s decision cannot be sustained on appeal.

Reversed and remanded for actions consistent with this opinion.

1 . Castañeda’s wife and daughters were included in his asylum application.

2 . The dissent claims that our discussion of the facts is "materially incomplete” and proceeds to submit its own version of the facts. Although it is certainly entitled to do so, we have some basic disagreements with this rendition both by reason of its content and its emphasis. With due respect, we believe that certain parts of its recitation are either irrelevant or taken out of context and clearly the subject of explanation. Perhaps a more egregious problem is presented by the inclusion of certain additional details, particularly those contained in page 66, line 11, through page 67, line 4, of the dissent, which recount the actions of persons for whom Castañeda was not responsible and who were without his command or authority, and additionally, for which Castañeda was absolved of responsibility. The inclusion of this information in the dissent's recount is thus not only unnecessary and irrelevant to a discussion of the issues before us, but also is unduly inflammatory and distracts from the proper focus of the legal issues raised by this appeal. The offending facts, although part of the decisions below, and undoubtedly an improper factor in the erroneous outcome reached, add nothing material to the appeal before us. Throughout this opinion we will, in footnotes and elsewhere, explain where we have disagreement with the dissent’s recitation of these and other matters.

3 . The Shining Path is a Maoist guerrilla group that "is among the world’s most ruthless guerrilla organizations” and "[ejngages in particularly brutal forms of terrorism, including the indiscriminate use of bombs.” The Institute for Counter-Terrorism, available at http://www.ict.org.il/inter — ter/orgdet.cfm? orgid=40 (last visited on August 2, 2006). The State Department has designated the Shining Path as a foreign terrorist organization. See United States Dept, of State, Country Reports on Terrorism, 2004 (April 2005), available at http://www.state.gOv/s/ct/rls/ 45394htm (describing the Shining Path as "one of the most ruthless terrorist groups in the Western Hemisphere” and noting that ”[a]pproximately 30,000 persons have died since Shining Path took up arms in 1980”). The Shining Path has perpetrated mass executions and committed "numerous massacres of entire families including young children and the elderly.” Anthony James Joes, Resisting Rebellion: The History and Politics of Counterinsurgency, 117 (2004).

4 . In March 2003, the relevant functions of the INS were transferred into the new Department of Homeland Security and reorganized into the Bureau of Immigration and Customs Enforcement ("BICE”). For simplicity, we refer to the agency throughout this opinion as the INS.

5 . The first Immigration Judge left the bench after Castañeda had already presented his entire testimony but before making a decision. The second Immigration Judge decided to hear Castañeda’s testimony again "in order to make [my] own determination of not only the lead respondent's credibility, but the facts before [me].”

6 .Castañeda testified in Spanish. His testimony was translated into English by an interpreter. During the government’s cross-examination of Castañeda, the interpreter participated in the hearing via telephone and was not actually in the courtroom. This caused problems during the testimony, as the interpreter often had to ask Castañeda to repeat his answers or the government to repeat its questions so that she could interpret them and often had to clarify or correct the original interpretations. We think it is important to note this fact because, in our view, some of the supposed inconsistencies or evasiveness mentioned by the dissent were more likely than not simple problems in translation. For example, when asked whether he could contact other patrols by radio during a military operation on August 5, 1985, Castañeda originally replied "No. I did not on my radio.” The Immigration Judge then stated "That's not what I asked” and asked the question again, to which Cas-tañeda replied "No. I could not.”

7 . According to the State Department's 1985 Country Reports on Human Rights Practices in Perú, the emergency zone was an area in which there was a declaration of a state of emergency and the armed forces were given full civil and military authority.

8 . The dissent states that Castañeda said it was " 'very important to coordinate the movement.’ Castañeda said his patrol would 'keep contact with the base by radio,' and at least sometimes would make direct radio contact with other patrols in the area.” One possible implication of the dissent’s phrasing is that Castañeda was stating that radio contact with other patrols was important to coordinate the movement. However, we wish to be clear that Castañeda was not testifying that radio contact with other patrols was important to coordinate movement. Rather, his statement that it was "important to coordinate the movement” was in response to the question "Was it important to know the location of otherfs] in your area?” From the full context of the statement, it is clear that Castañeda was simply stating that it was important to know the location of other patrols to coordinate movement, not to have actual radio contact with them. Castañeda had already testified that he kept in radio contact with his base and that his base commander would let him know where other patrols were, although he also sometimes had direct radio contact with the other patrols.

9 . This zone was located within the Ayacucho region, i.e., the emergency zone.

10 . Apparently, Lloellapampa is located between two mountains, and there were very few paths into or out of the village.

11 . The dissent states that, when asked if he was informed what frequencies the two attack patrols were using, Castañeda replied "That’s correct.” However, the dissent leaves out the rest of Castañeda’s reply. The full quote is "That's correct, but both patrols entered from the same place and where those patrols are directed, four patrols are directed from the same head of unit, but these orders came from the division. I was directly directed by the head of commander of base.” From this full quote, it is clear that Castañeda misunderstood the question, and was explaining that the four patrols were directed by a division head, and that his orders during the Operation came directly from the head commander of his case. A few questions later, Castañeda explicitly clarified that he could not contact the other patrols because he did not have their frequency.

12 . The dissent notes that, when asked what happened in the village, Castañeda originally stated that ”[o]ne of the patrols ... committed excesses.” The implication behind the inclusion of this fact is that Castañeda was being evasive. This is not the case. From the transcript, it appears that, after Castañeda testified about "excesses,” the Immigration Judge interjected to ask the interpreter what she had said. The government then asked Castañeda if he could be more specific, and Castañeda immediately replied that the patrol had executed civilians. We do not think that this sequence of events indicates any evasiveness on Castañeda’s part.

13 . The dissent argues that Castañeda’s testimony shows that he was evasive regarding his knowledge of the number of civilians killed in Llocllapampa. This is incorrect. The Human Rights Commission noted that there were three different reports regarding the number of civilians killed: twenty-five, forty, and sixty-nine. It also noted that the exact number could not be determined. To this day the exact number of civilians killed is unknown. Castañeda was therefore entirely correct in stating that he never learned the exact number of people killed.

14 . This testimony is supported by other aspects of the record, which we discuss infra.

15 . The dissent makes much of the fact that Castañeda was debriefed, arguing that the debriefing means it is unlikely that he did not find out about the massacre until three weeks later, as he claimed. We have two responses. First, Castañeda stated that the "debriefing” simply meant that he gave a report to his base commander. He gave no indication that he met with the other patrols or discussed what the other patrols did, and there is no evidence in the record to controvert that testimony. Second, it would make sense that those who slaughtered the civilians would not tell other soldiers who were not involved. One would expect that those who participate in the murder of innocent civilians would not openly discuss what they have done with others who did not engage in such atrocities.

16 . As previously noted, we are concerned that the dissent's description of the murders of the villagers is inflammatory, and we therefore want to emphasize that it is undisputed that Castañeda was not involved in the killing of any civilians. He was also later absolved of responsibility by a military court, a fact which we discuss in more detail below.

The dissent is equally inflammatory in speculating, in footnote 43, that "if Hurtado was acting on orders, rather than in the heat of the moment, it is much more likely that Cas-tañeda would have known about the massacre before it happened”. Such speculation loses sight of the facts of the case: there is no evidence that Hurtado was acting on orders to commit the massacre, and there is no evidence that Castañeda would have known of these hypothetical orders had they ever been given.

17 .The dissent notes that the Human Rights Commission stated that Hurtado "is only a piece of the larger picture and it is necessary to study whether he acted on virtue of expressed verbal orders.” One possible implication of this isolated fact is that Castañeda was perhaps part of the "larger picture.” However, taking the statement within the context of the report, it is clear that the Commission was concerned that persons with a higher military rank had ordered Hurtado to kill the civilians. In particular, elsewhere in tire report the Commission refers to possible "verbal orders” given to Hurtado by a general named Wilfredo Mori Orzo. Nothing in the report indicates Castañeda had any orders beyond those to watch the path for any Shining Path guerrillas or that Castañeda was somehow part of the "larger picture.”

A more egregious misconstruction of the facts of the case is the dissent's statement that "[i]t is disputed whether [Castañeda] knew about the massacre before or while it took place.” Again, there is no evidence that the massacre was planned or that Castañeda had knowledge of these plans if they ever existed. The IJ did dispute whether Castañeda learned, via radio contact, of the massacre as it took place. However, this dispute is irrelevant. Castañeda did not have authority over Hurta-do and he was located 3 to 5 miles outside of Llocllapampa. Had Castañeda learned of the massacre as it transpired, he would not have been able to either assist or stop it.

18 .In stating that “three officers [including Castañeda] were responsible” for the massacres, the report was drawing from initial reports by the Peruvian military and the Human Rights Commission. It should not be taken as a definitive finding that Castañeda was responsible for the massacre, especially since it was released in February 1986 and thus does not take into account Castañeda's subsequent acquittal by the Peruvian military court system following a formal proceeding. We discuss the acquittal below.

19 . From the record, it appears that the Peruvian military justice system has three levels: a trial level, an appeals level, and the Supreme Council of Military Justice as the highest level.

20 . The record indicates that Hurtado's case was still pending when Castañeda left Perú in 1991. A final disposition of Hurtado’s case did not occur until 1993.

21 . Castañeda did testify that, after he came to the United States, he had heard rumors that Hurtado had been promoted to captain, but he did not know if they were true.

22 . Castañeda’s petition was filed after the passage of the Real ID Act of 2005, Pub.L. 109-13, 119 Stat. 302. The Real ID Act "alters, among other things, the standards governing credibility determinations and the need for corroboration of testimony in asylum cases.” Dhima v. Gonzáles, 416 F.3d 92, 95 n. 3 (1st Cir.2005). However, because Cas-tañeda’s asylum application was filed well before the effective date of these provisions of the Act, the provisions are inapplicable here.

23 . The regulations dealing with asylum applications filed on or after April 1, 1997 do not explicitly contain this burden shifting provision. See 8 C.F.R. § 1208.13(c)(1). However, because Castañeda’s application was filed in 1993, there is no issue regarding what the burden should have been before the Immigration Judge.

24 . The BIA also stated that “the respondent’s evidence of his dismissal is not necessarily the legal equivalent of a criminal finding that he was innocent.” This statement is belied by the record, as the 1999 State Department Report specifically notes that all the defendants (including Castañeda) except Hurtado were “acquitted.” To the extent the BIA was requiring an actual finding of “innocent,” in our view such a requirement would be unreasonable as it goes beyond what even our own justice system offers. Under our own justice system, defendants who are "acquitted” are found “not guilty,” they are not found “innocent.”

The dissent argues that the BIA did not require an actual finding of innocence, but simply required Castañeda to show his innocence through court documents or otherwise. We disagree. The BIA specifically discounted Castañeda’s evidence because it was not “the legal equivalent of a criminal finding that he was innocent.” (emphasis supplied). Our point is that, under our own justice system, it would have been impossible for Castañeda to provide such a finding, and it was therefore unreasonable to discount the evidence of his acquittal simply because it is not a criminal finding of innocence.

25 . There is some indication that the Immigration Judge, and perhaps the BIA, faulted Cas-tañeda for not providing evidence of an acquittal from the civilian court system of Perú. We wish to note that, to the extent that the decision rested on these grounds, it was erroneous. The highest body of the civilian court system, the Supreme Court of Perú, determined that the military system had jurisdiction over Castañeda’s case. There is therefore no doubt that the case was properly before the military courts, and there is simply no reason to fault Castañeda for not providing evidence of acquittal in the civilian system.

26 . The dissent seizes on the both the Immigration Judge's and BIA's use of the word "demeanor,” noting in several places that Castañeda was found not credible based on "observations” of his demeanor. The implication of the dissent’s emphasis on demeanor evidence is that the Immigration Judge and BIA based the adverse credibility finding on numerous problems with Castaneda’s demeanor. For example, the dissent notes that the Immigration Judge stated that he had "carefully observed [Castañeda’s] demeanor ... and found him vague, evasive, and non-responsive.” However, as the dissent also notes, a hearing officer who makes an adverse credibility determination must support the determination with specific findings in order to be accorded deference. Aguilar-Solís v. INS, 168 F.3d 565, 570 (1st Cir.1999). The only specific finding that the Immigration Judge made regarding Castaneda’s demeanor-as opposed to the content of his testimony — was that during testimony about radio communication, Castañeda was "blinking his eyes in an unusually rapid rate.” We readily agree that demeanor evidence is peculiarly within the purview of the trier of fact. However, absent other evidence of a lack of credibility, we cannot affirm an adverse credibility finding based solely on a person’s "blinking” during a portion of their testimony. Such ambiguous indicia are simply too thin to support such a conclusion.

27 . Further, the 1985 Report notes elsewhere that "[b]eginning in 1983, some members of the security forces responded to [the Shining Path] with violence of their own, engaging in extrajudicial killings, disappearances, torture, and arbitrary detentions. Most of these violations occurred in 1983 and 1984 in the Ayacu-cho Emergency Zone.... During 1985, reports of such abuses by security forces dropped sharply....” In 1983 and 1984, Cas-tañeda was serving in Tumbes, around 400 miles away from the emergency zone. It is therefore likely that he heard only occasional reports of abuses occurring in the emergency zone. Further, as indicated by the State Department, by the time he was transferred into the emergency zone in 1985, the number of abuses decreased sharply. This further bolsters his testimony that he only rarely heard of abuses by the military.

28 . We also note that Castañeda's knowledge, or lack thereof, regarding Hurtado is a red herring. Castañeda was denied asylum because of what he allegedly did, not because of what Hurtado did. As we previously noted, Castañeda's case is based on his disassociation from Hurtado’s actions, actions over which Castañeda had no control or knowledge.

29 . The fact that there was a conviction in the case should further boost Castañeda's credi-bilily.

30 . The dissent's analysis would require us to remand whenever we found an error in the BIA's factual decisions, even if the BIA had clearly decided the issue based on a fully developed record. This would transform the "ordinary remand" rule into an "automatic remand” rule, something we do not believe is required by either the Supreme Court or our own precedents.

31 . The dissent also cites Seavey v. Barnhart, 276 F.3d 1 (1st Cir.2001). However, that case actually supports our decision to reverse. In Seavy, which involved an agency's denial of social security benefits, we remanded to the agency for further consideration. However, we noted that “if the evidence and law compelled one conclusion or the other, then [we] could order an award of benefits or affirm a denial of benefits.” Id. at 11. Here, under substantial evidence review, we may reverse the decision of the BIA only if we are compelled by the evidence. In deciding that the BIA’s determination is not supported by substantial evidence, we are in reality stating that the evidence compels a contrary conclusion. Thus, under the language of Seavy, we may simply reverse the agency's decision.

32 . The dissent argues that this portion of the BIA's opinion was not an alternate holding and that we should remand to the BIA so that it can address in the first instance whether Castañeda's actions amounted to assistance or participation in persecution. In essence, the dissent believes that the BIA did not assume the credibility of all of Castañeda's testimony and that the quoted portion of the BIA's opinion does not adequately explain the BIA’s reasoning. It bases this argument in part on the BIA's statement that “even if [Castañeda] testified credibly that he did not harm or execute anyone on August 8, 1985,” (emphasis supplied), he assisted or otherwise participated in persecution.

We disagree with the dissent's characterization of the ruling below. The context of the BIA’s statement, especially its citation to, and affirmation of, a portion of the Immigration Judge's decision where the Immigration Judge explicitly makes an alternate holding, as well as its citation to In re Rodríguez-Majano, one of the BIA's seminal opinions on the issue of whether an alien assisted or participated in persecution, leaves us with no doubt the BIA was making an alternate holding. This alternate holding was that, even if Castañeda were credible, his actions constituted assistance or participation in persecution. Further, although brief, the BIA’s reasoning is crystal clear: it believed that the act of blocking a path during the Operation had the objective effect of assisting in the massacre by helping to confine the villagers. Given these considerations, we are convinced that this alternate holding is ripe for review. See Xu v. Gonzáles, 424 F.3d 45, 49 (1st Cir.2005) (noting that “[w]hen considering ... the clarity of an administrative decision ..., we are not blind to the context in which the decision is made or oblivious of the record on which it is based"). The fact that the BIA has already decided the issue also takes this case outside of the dissent’s arguments about the ordinary remand rule, which we have already addressed above.

33 . Castañeda also briefly argues that his actions were directed toward the defense of the government, and therefore any harm that resulted from such behavior would not amount to persecution. See In re Rodríguez-Majano, 19 I. & N. Dec. 811, 815 (BIA 1988) (stating that “harm which may result incidentally from behavior directed at ... the defense of [a] government against an opponent” does not amount to persecution.). However, the massacre of the villagers during the Operation went well beyond that which was incidental to the defense of the Peruvian government. See, e.g., Miranda Alvarado v. Gonzáles, 449 F.3d 915, 927 (9th Cir.2006) (stating that "torturing individuals selected for their affiliation with an opposition group is not inherent in armed conflict, any more than is 'ethnic cleansing’ ”). Therefore, if Castañeda assisted or otherwise participated in the massacre, he should be barred as a persecutor.

34 . This is the course taken by most of the circuit courts that have considered issues similar to the one before us today. See, e.g., Miranda Alvarado, 449 F.3d 915, 924-25; Xie v. INS, 434 F.3d 136, 140-42 (2d Cir.2006); Hernández v. Reno, 258 F.3d 806, 812-13 (8th Cir.2001).

35 . We note that the courts have been in less agreement regarding the subjective aspects of an alien's actions. In Fedorenko, the Supreme Court held that, in the context of the DPA, the voluntariness of an alien's actions was irrelevant. 449 U.S. at 512-13, 101 S.Ct. 737. Some courts have taken Fedorenko’s holding and applied it to the INA. See Xie, 434 F.3d at 141-42 (stating that "we find it unlikely that the phrase 'assisted in persecution’ implicitly includes a voluntariness requirement” in the INA but not the DPA and noting that, in prior cases, the court had "deemed irrelevant [the alien's] personal motivation or intent”); Singh, 417 F.3d at 740 (stating that an alien's personal motivation for his or her actions is irrelevant); Bah v. Ashcroft, 341 F.3d 348, 351 (5th Cir.2003) (same); Rodríguez-Majano, 19 I. & N. Dec. at 815 ("It is the objective effect of an alien’s actions which is controlling.”). Other courts appear to be more willing to look into the alien’s intent. See Miranda Alvarado, 449 F.3d 915, 925 (stating that a court should evaluate an alien's “purposeful assistance”); Hernández, 258 F.3d at 814 (faulting the BIA for not considering the alien's testimony that his actions were involuntaiy and motivated by fear for his own life). We need not decide that issue today because, even without considering the motivation or intent behind Castañeda's actions, we do not think his actions had any objective effect on the persecution of the villagers.

36 . We again emphasize that títere is no evidence in the record to support the dissent's intimation, see footnote 43, to the effect that the massacre was pre-planned or that Cas-tañeda had prior knowledge that it would take place.

37 .The dissent, citing Naujalis v. INS, 240 F.3d 642 (7th Cir.2001), argues that our reading of the "objective effect” line of cases is erroneous. We do not agree with the holding in Naujalis. Furthermore, the Seventh Circuit’s opinion in Singh v. Gonzáles, 417 F.3d 736 (7th Cir.2005), has recently clarified the holding in Naujalis. In Singh, the alien was a citizen of India who worked for the police department in India's Pujab. The court noted that, in the context of Nazi guards and the DPA (which is what Naujalis involved), "membership in the ranks of Nazi guards is sufficient to constitute assistance in prohibited persecution.” Singh, 417 F.3d at 739. It then went on to state that the DPA line of cases was not fully compatible with the INA in general and Singh’s case in particular because "[ujnlike Nazi concentration camps, whose complete existence was premised upon the persecution of innocent civilians, local Punjabi police departments served traditional, legitimate law enforcement purposes and did not exclusively engage in the persecution of innocent [people].” Id. Therefore, the court stated, in Singh’s case "a distinction must be made between genuine assistance in persecution and inconsequential association with persecutors.” Id. The court then went on to conduct an analysis similar to the one we have conducted. Id. at 739-741.

In sum, the dissent's reliance on Naujalis is misplaced because Naujalis involved a special rule applicable "exclusively ... to Nazis from the period of 1933 to 1945.” Id. at 739. Had Naujalis been the member of an organization which did not exclusively engage in persecution — as is the case with Castañeda and his membership in the Peruvian army — the Seventh Circuit’s analysis would have been entirely different. Given the Seventh Circuit’s explanation of Naujalis in Singh, along with the other cases that have construed the relevant provisions of the INA, as opposed to the DPA, we are convinced that our approach is the correct one.

38 . As evidenced by the facts that the massacre occurred even though one of the blocking controls never reached its position and that no persons came down the path Castañeda's men were watching.

39 . We find that what might have occurred had a fleeing villager come down the path to be absolutely irrelevant and speculative, since Castañeda was there to intercept fleeing members of the Shining Path, not civilians. If anything, the cases make clear that we are to concern ourselves only with what a petitioner actually did, not what he might have done. See, e.g., Rodríguez-Majano, 19 I. & N. Dec. at 814-15. Castañeda credibly testified — and this testimony is supported by the Human Rights Commission — that no one came down the path.

40 . Other non-adjudicative bodies, such as the Human Rights Commission, also placed the blame elsewhere.

---

Dissenting Opinion

LYNCH, Circuit Judge,

dissenting.

I respectfully dissent and do so with regret. But the majority opinion is contrary to decisions of the Supreme Court and this court, and the errors are of exceptional importance and potentially of national significance. Cf. Fed. R.App. P. 35.

I.

I summarize my view of the case and the reasons I am compelled to dissent before going on to a fuller explanation.

According to an investigation by a Commission of the Senate of Perú, in August 1985, some sixty-nine innocent villagers in Llocllapampa, Perú were massacred by units of the Peruvian army in its war with Shining Path guerrillas. The village’s civilian inhabitants were herded into two homes; the army units fired and threw grenades into the houses and killed many women and children in the slaughter. The investigation concluded the operation amounted to genocide.

Lieutenant Castañeda, petitioner here, headed one of four army units involved in the military operation during which this massacre occurred. It is disputed whether he knew about the massacre before or while it took place. His patrol was assigned to and did block a possible escape route from the village, but there is no evidence that he was in the village during the massacre or that he personally killed anyone. He acknowledged that he was briefed on the operation beforehand and that all units during the operation had radios capable of radio contact, but he claimed that he had no radio contact with the units in the village that day and did not find out about the massacre until weeks later.

Finding himself a potential target in the wake of his role in the operation, Castañe-da came with his family to the United States on tourist visas in 1991, overstayed, and in 1993 applied for asylum and withholding of removal. By statute, an alien is not eligible for asylum and withholding of removal if the Attorney General determines that “the alien ordered, incited, assisted, or otherwise participated in the persecution of any person on account of race, religion, nationality, membership in a particular social group, or political opinion.” 8 U.S.C. § 1158(b)(2)(A)(i) (emphasis added); see also id. § 1231(b)(3)(B)(i).

It is undisputed that the government met its burden of making a prima facie case that Castañeda had “assisted or otherwise participated in the persecution” and that the burden of showing he had not done so has shifted to Castañeda. 8 C.F.R. §§ 1208.13(c)(2)(h), 1208.16(d)(2). It is also undisputed that the persecution was on account of one of the enumerated grounds in the statute.

The Board of Immigration Appeals (“BIA”), the Immigration Judge (“IJ”), and the initial asylum interviewing officer each found Castañeda had not met his burden because he was not credible. He was found not credible based both on observations of his demeanor and on the content of his testimony. The law requires courts to give great deference to those lack of credibility determinations, especially those based on direct observations of the alien’s demeanor. The majority fails to accord the appropriate deference to the BIA and, rather than determining whether the record could reasonably be read to support the BIA, determines whether the record could reasonably be read to support Castañeda.⁴¹ This is the wrong perspective, and having taken the wrong perspective, the majority reaches the wrong conclusion in holding that substantial evidence does not support the BIA’s finding that Castañeda was not credible.

Having incorrectly applied the standard of review, the majority further errs in refusing to remand to the BIA to reconsider its determinations in light of the supposed failures in its reasoning and errs in concluding for itself that Castañeda was credible. Supreme Court precedent and the precedent of this circuit require that the agency be given an opportunity to reconsider, particularly when the BIA did not find it necessary to comment on all of the evidence relied upon by the IJ. The majority errs in holding as a matter of law that Castañeda was credible and that he had no knowledge of the massacre. The record does not compel either conclusion.

This sequence of errors has led the majority into yet a further error, one of great significance and going well beyond the particulars of this case. The majority takes a one-line statement by the BIA — that Cas-tañeda still had not met his burden even assuming he was credible that he personally had not killed or harmed anyone — and draws two mistaken conclusions. First, it turns this into a statement that the BIA made a holding assuming Castañeda was credible in all of his testimony, including that he had no knowledge of the massacre. The BIA did not say this, and if there is any ambiguity, the case should be remanded for the BIA to clarify. Second, the majority sees in the BIA’s statement a ruling of law on the relevance of Castañe-da’s knowledge about the massacre to the statutory exclusion for those who assist or otherwise participate in persecution. The BIA’s statement means no such thing. The issue of the legal significance of Cas-tañedas knowledge was not raised or argued before the agency in the removal proceedings, but was rather raised sua sponte by this court, which asked for and received post-argument briefing on the issue.

Despite the position of the Attorney General that this issue of statutory interpretation is not before this court, was not raised or exhausted before the agency, and is committed in the first instance to the BIA, the majority nonetheless addresses the issue itself, contrary to both the exhaustion and Chevron doctrines. The majority’s conclusion as a matter of law that Castañeda has met his burden of proving he was not a persecutor depends on the majority’s implicitly deciding that knowledge is a requirement under the statute (contrary to the position stated by the Attorney General in his brief) and then on the majority’s applying this newly formulated legal rule to its version of the facts of this case. The majority refuses to permit the agency to examine on remand the questions both of statutory interpretation and then of application of the legal standard to the facts. The majority concludes that Castañeda has established that he did not assist or otherwise participate in the persecution and remands to the agency to consider only the question of whether he should be granted asylum.

The majority is wrong on all of these points and has overstepped the limited role assigned to the courts in reviewing decisions of the BIA in immigration cases. The proper outcome of this case is a simple affirmance of the BIA’s decision on the grounds that the decision is based on substantial evidence that Castañeda has not met his burden and the record does not compel any other finding. There is no need to reach the other issues the majority reaches. More than that, courts are forbidden from denying the BIA the opportunity to interpret what the statute means. Accordingly, I dissent. I also attach to this dissent the decision of the BIA in this case, so it may be read directly.

II.

A. Standard of Review

We are required to give deferential review to the BIA’s findings. “[A] decision that an alien is not eligible for admission to the United States is conclusive unless manifestly contrary to law.” 8 U.S.C. § 1252(b)(4)(C). As to the BIA’s subsidiary factual findings, including credibility determinations, we apply the highly deferential substantial evidence standard. Dhima v. Gonzáles, 416 F.3d 92, 95 (1st Cir.2005). Under this standard, “[t]he [BIA’s] determination must stand ‘unless any reasonable adjudicator would be compelled to conclude to the contrary,’ ” id. (quoting 8 U.S.C. § 1252(b)(4)(B)), meaning we can overrule the BIA’s findings only if the evidence “points unerringly in the opposite direction.” Laurent v. Ashcroft, 359 F.3d 59, 64 (1st Cir.2004).

“We review the BIA’s legal conclusions de novo, with appropriate deference to the agency’s interpretation of the underlying statute in accordance with administrative law principles.” Gailius v. INS, 147 F.3d 34, 43 (1st Cir.1998); see also Da Silva v. Ashcroft, 394 F.3d 1, 5 (1st Cir.2005) (citing, inter alia, Chevron U.S.A., Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 842-43, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984)).

B. Facts of Record

As I believe the majority’s discussion is materially incomplete, I independently recount the facts as contained in the administrative record.⁴² See 8 U.S.C. § 1252(b)(4)(A).

Castañeda joined the Peruvian military in 1978. In 1983, he graduated from a military academy and was commissioned as a lieutenant. In January 1985, Lieutenant Castañeda was transferred to anti-terrorist Battalion 34, stationed in Ayacu-cho, a province within the “emergency zone” and the historic birthplace of the Shining Path. This battalion, composed of five companies spread throughout the region, engaged in field operations. Cas-tañeda was first assigned to be head of a twenty-man patrol in Sacchaeamba, which had two patrols.

After three months, Castañeda transferred to another base, in Vilcashuan, again as head of one of two patrols. Cas-tañeda was one of only a few officers at this base; besides himself, there was a captain and a base head. Castañeda was stationed there in August of 1985.

In August 1985, Castañeda was involved in a military operation, the details of which are at the center of this case. The operation involved four patrols. Two of the patrols, led by Sub-Lieutenant Telmo Hurtado and Lieutenant Riveri Rondón, were to enter the village of Lloellapampa, in the Accomarca region, to search for Shining Path members. Two other patrols, including Castañeda’s, were assigned to block escape routes from the village.

The two main patrols entered the village and massacred as many as sixty-nine civilians. The massacre was reported by the media, and a formal investigation followed. A Peruvian Senate Human Rights Commission found that Hurtado’s patrol had herded the village inhabitants into two houses, open fired on them, lobbed grenades inside, and set fire to the houses. The Commission concluded that the operation amounted to genocide and found that “[jjudging by the age of the people and the way they responded[,] it was clear that the people attacked and killed were defenseless.”

The Commission noted that Castañeda’s patrol was “not involved in any confrontations with fugitive civilians.” But it also found that the massacre could not simply be attributed to a single out-of-control officer; instead it wrote that Hurtado “is only a piece of a larger picture and it is necessary to study whether he acted on virtue of expressed verbal orders.”⁴³ The Commission left it to the Peruvian court system “to determine ... the people responsible ... and apply punishments as dictated by the law.”

The massacre was documented in the U.S. State Department’s 1985 Country Report on Human Rights Practices in Perú. The report stated that “an Army sub-lieutenant and three other officers were responsible for the Accomarca massacre of some 25 to 69 peasants.” Castañeda admitted during testimony before the IJ that he was one of the officers referred to in the report. The report also said the men involved had been charged in the military and civilian court systems in Perú.

The Peruvian Supreme Court decided the military courts, rather than the civilian courts, had jurisdiction, and Castañeda, along with Hurtado and others, was charged in a military court martial in 1986. Castañeda was charged with homicide and abuse of authority.⁴⁴ Castañeda appeared in a military court on at least four occasions, represented by a civilian lawyer, and produced documents in his defense. He also testified in his own defense, answering questions from a judge and prosecutor. Castañeda was apparently acquitted of all charges at the court martial. He returned to duty thereafter and was eventually promoted to captain in 1991.

Castañeda’s application for asylum and other relief was based on his assertion that after the massacre, he was persecuted by the Shining Path.⁴⁵ Castañeda decided to leave Perú. He obtained tourist visas for himself and his family, resigned from the military, and came to the United States in 1991.

In January 1993, he and his family applied for asylum and withholding of removal. The claims for relief were initially denied by an agency hearing officer, who found, in a May 19, 1999 report explaining the denial, that Castañeda was ineligible for asylum because his testimony was “only partially credible.” The report stated that Castañeda’s testimony about his duties as a military officer was “vague and evasive,” and that Castañeda’s testimony regarding military practices was “inconsistent with country condition information.” Castañeda had “denied awareness of any harm that occurred to captives,” testified that suspected terrorists were turned over to a military intelligence unit, and “testified not to have any knowledge of any harm that may have occurred to such persons.” The report found it unlikely that a Peruvian military officer of Castañeda’s era would have had “such limited knowledge of the treatment of suspected terror-ista.” Castañeda’s case was subsequent^ ly referred to an IJ for adjudication. The hearing officer’s report was in the record before the IJ and BIA.

III.

The hearings before the IJ focused on whether Castañeda was ineligible for asylum and withholding of removal because he “ordered, incited, assisted, or otherwise participated in the persecution” of others on account of “race, religion, nationality, membership in a particular social group, or political opinion.” 8 U.S.C. §§ 1158 (b)(2) (A) (i), 1231(b)(3)(B)(i).

Under the applicable regulations, if evidence is produced “indicating]” that an alien engaged in such behavior, the alien must prove by a preponderance of the evidence that he or she did not do so. 8 C.F.R. §§ 1208.13(c)(2)(h), 1208.16(d)(2). The IJ found, and Castañeda does not dispute, that the government produced evidence indicating that Castañeda “assisted, or otherwise participated” in the Accomar-ca massacre. It is, as a result, undisputed that the burden lay with Castañeda to show that he did not do so.

A. Hearing Before the IJ

Castañeda put on two types of evidence during an extended hearing before the IJ: his testimony about the events in question and a document purporting to be a judgment from a military tribunal acquitting him of certain charges.

Castañeda testified extensively about his military experiences both before and during August 1985. He testified that at both bases at which he was stationed in 1985, he headed one of two patrols. While one patrol was out on operations, the other stayed to protect the base. Castañeda indicated that his patrol went out on both combat and reconnaissance missions. The commander of the battalion was the person who determined the instructions the patrols were given.

Sometimes the patrols were coordinated; sometimes they were independent. Cas-tañeda testified that when the patrols were independent, the commander still let each patrol know where the others were. It was, he said, “very important to coordinate the movement.” Castañeda said his patrol would “keep contact with the base by radio,” and at least sometimes would make direct radio contact with other patrols in the area.

Castañeda testified that in August 1985, after several weeks of reconnaissance patrolling, he was ordered to go on a combat patrol in mobilized, coordinated activity with three other patrols. His order came from the base captain, following orders of the division command. Castañeda indicated that he was briefed in advance on the missions of the other three patrols involved in the operation. He was informed that there were forty to sixty Shining Path terrorists at a training center in the village. His patrol was ordered to deploy itself to a place where it could block the path providing an escape route from the village. His deployment position was, according to Castañeda, located three to five miles from the village. A second patrol was supposed to block the other exit to the village, but it never arrived at its position. The two remaining patrols were the active patrols whose mission was to go into the village.

Castañeda’s patrol had twenty men, a radio, and a rocket launcher. He said it took approximately six hours to reach the deployment position. Once there, he testified, he and his troops hid themselves along the sides of the path. When he was in position, he radioed that fact back to the head of the base. He said he would have been in charge of deciding whether to open fire if anyone came along.

Castañeda testified that he knew there would be an attack on the village, but was not told when the two lead patrols were ready to attack or when the attack began. He says he was told, via radio contact from the base, only when the attack was over. During the operation, he testified, no one came down the path and his patrol fired no shots. Other evidence in the record suggests the attack went on for two to three hours.

Castañeda denied hearing any radio communications from the attacking patrols during the mission. At first, when asked if he had been “informed what frequencies the two attacking patrols would be using,” he replied, “That’s correct.”⁴⁶ Shortly thereafter, however, he said he did not have the frequencies of the other patrols because he had not been given them. When asked whether the two attacking patrols had his patrol’s radio frequency so that they could notify him of any escaping guerrillas, he said they did not. He maintained that each patrol could contact the base by radio, and the two attacking patrols could communicate with each other, but the attacking patrols, by contrast, could not communicate with the patrols blocking the escape routes. Castañeda was asked to reconcile this testimony about the lack of radio contact among all the patrols, particularly in a mission where multiple patrols were trying to coordinate an attack, with his earlier testimony that communication among patrols could be important. He responded that the attack had been centrally planned that way and that he was simply ordered to exercise his independent judgment in controlling exit points.

After the attack ended, Castañeda was ordered back to base by his base commander. He acknowledged that there was a debriefing. He testified that he did not find out about the massacre until weeks later, when he heard a report on civilian radio. He said he later found out that Hurtado had admitted to executing civilians.

As to the subsequent court martial, Cas-tañeda stated that he was acquitted of all charges, and produced a document purportedly showing that the Supreme Council of Military Justice had affirmed the dismissal of the charges against him. The document, dated April 4, 1989, stated in relevant part that Castañeda was charged with homicide and abuse of authority and that the charges were dismissed. It gave no reason for the dismissal.

Castañeda submitted no other documentation relating to the court martial. He did not introduce into his immigration proceedings a transcript of the military court proceedings or the documents he had used in his defense. He also offered no evidence as to what the elements of the charged crimes were, and therefore did little to prove the significance of the purported acquittal in determining whether he assisted or otherwise participated in persecution. However, the U.S. State Department’s 1999 Country Report on Human Rights Practices for Perú does corroborate Castañeda’s ultimate acquittal, stating that “all the other defendants” besides Hurtado were “acquitted.”

B. The IJ’s Decision

The IJ heard testimony from Castañeda in four separate sessions over the course of more than a month.'⁴⁷ On October 4, 2004, the IJ concluded that Castañeda was ineligible for asylum and withholding of removal because he had not proven that he did not “assist[ ] or otherwise participate[ ]” in politically motivated killings. 8 U.S.C. § 1158(b) (2) (A) (i); id. § 1231(b)(3)(B)(i). The IJ had three grounds for this conclusion. First, he rejected Castañeda’s argument that the one-page Peruvian document dismissing the charges against him was sufficient to carry his burden. Second, he held that Castañeda was not credible and had not carried his burden for that reason. Third, the IJ held that even setting aside his credibility determination, Castañeda’s “motivation and intent [were] irrelevant,” given that the “objective effect of [his] participation in the Accomarca operation would have been to ... aid even in some small measure with the execution and murder and rape of those unarmed men, women, and children.”

The IJ offered numerous grounds for the finding that Castañeda’s testimony was not credible. First, the IJ wrote that he had “carefully observed [Castañeda’s] demeanor ... and found him to be vague, evasive, and non-responsive during questioning by the Government prosecutor relating to his participation in the massacre at Accomarca and more importantly whether he was aware of that massacre.” The IJ also found that Castañeda was evasive and non-responsive when asked why he did not have documentation from his court martial. The IJ further found that Castañeda’s testimony that he did not have the radio frequencies for the other patrols the day of the massacre was “wholly incredible,” in part because he had earlier testified that radio contact among patrols could be important and in part because his demeanor indicated a lack of veracity. As to this demeanor finding, the IJ wrote that he “observed [Castañeda’s] demeanor during this part of his testimony and found him to be extremely incredible.... [Castañeda was] blinking his eyes in an unusually rapid rate as compared with the rest of his testimony.” The IJ also found incredible Castañeda’s testimony that he did not find out about the massacre until weeks later, and then only from the media, especially considering his rank and his presence at briefings and debriefings. Finally, the IJ found incredible both Castañeda’s testimony about his awareness of human rights violations in the emergency zone and his testimony about his knowledge of Hurtado’s fate.

C. The BIA Decision

On appeal to the BIA, Castañeda first argued that the IJ should have deferred to the Peruvian tribunal’s dismissal of the charges against him; he argued that the IJ’s finding- — that Castañeda failed to prove he had not assisted in persecution— was “in direct contravention” of the military court’s judgment.

The BIA responded on several levels. First, it held that the IJ was correct about potential flaws in Castañeda’s documentary evidence: there was some doubt about whether the acquittal document was properly authenticated, and further, other documents suggested that the military tribunal may have been a mechanism “to grant impunity to the alleged persecutors.”

Second, the BIA held as a matter of law that dismissal of the charges in the military tribunal did not establish that Cas-tañeda was not a persecutor. The government had argued that “fail[ure] to meet the burden of proof in a criminal forum never ... proves that the defendant was factually innocent [and] ... [t]he legal presumption of innocence at a criminal proceeding cannot be substituted for independent facts which could have established [Castañeda’s] burden of proof.” Moreover, the government had noted that “[t]here are many possible scenarios at a criminal trial that could conclude the proceedings in an acquittal without a determination of innocence.” Agreeing with these arguments, the BIA concluded that Cas-tañeda’s “evidence of his dismissal is not necessarily the legal equivalent of a criminal finding that he was innocent.” It was Castañeda’s burden to prove he did not assist or otherwise participate in the massacre; the military court document, which did not specify the reason for dismissal of the charges or offer any details on the charges themselves, did not establish that he had not assisted or otherwise participated.

In concluding that Castañeda had failed to meet his burden of proof, the BIA held that “even if the respondent testified credibly that he did not harm or execute anyone on August 8, 1985, there is substantial evidence to support the Immigration Judge’s conclusion that the respondent did aid in the persecution of others by helping to confine them, inasmuch as his patrol was charged with blocking an escape route for the villagers.... ”

The BIA also affirmed on the separate basis that the IJ had not clearly erred in finding Castañeda less than credible. The BIA specifically referenced the IJ’s finding that Castañeda’s demeanor during a line of significant questioning indicated a lack of veracity, and affirmed the finding that Castañeda was “evasive” regarding his court martial. It also said that Castañe-da’s argument on appeal that he did not have access to a transcript or other documents from the court martial “does not seem plausible,” and that Castañeda did not “provide any explanation for not providing [such documents].”

The BIA also affirmed the IJ’s conclusion that crucial portions of Castañeda’s testimony were inconsistent and incredible and reached the same conclusion on its own review. The BIA found that the IJ had not erred in doubting the veracity of Castañeda’s testimony that he had not been in radio contact with the other patrols the day of the killings (and thus had not been aware of the unfolding massacre), particularly because Castañeda’s own testimony established that patrols sometimes made contact and that “radio contact was important for purposes of coordination as well as efficiency and protection of the patrols.” Further, the BIA affirmed the IJ’s findings that Castañeda was not forthcoming about the extent of human rights violations by the Peruvian military and about the fate of Hurtado, the patrol commander charged with primary responsibility for the massacre. It concluded: “In light of obvious media attention and international documentation of the massacre, as well as [Castañeda’s] military position and subsequent accusations about his role, it was reasonable to expect Mm to provide answers regarding what occurred in Perú on and after August 8, particularly since he failed to provide adequate documentary evidence of his innocence.”

IV.

A. The Adverse Credibility Determination

Substantial evidence supports the BIA’s adverse credibility determination, which in itself is sufficient to support the BIA’s conclusion that Castañeda did not meet his burden. The IJ and BIA based their findings on Castañeda’s demeanor and evasiveness and on the tendency for his testimony to be contradictory and non-responsive. The majority errs in dismissing this evidence and imposing its own view of the record.

In order to carry his burden, “an alien must support his claim ... through credible testimony.” Nikijuluw v. Gonzáles, 427 F.3d 115, 121 (1st Cir.2005). “The testimony of the applicant, if credible, may be sufficient to sustain the burden of proof without corroboration.” 8 C.F.R. § 1208.13(a) (emphasis added). However,

[tjhis does not mean that a reviewing court must take every applicant’s uncon-tradicted testimony at face value, for testimony sometimes is internally inconsistent or belied by the prevailing circumstances. Furthermore, a witness’s demeanor is often a critical factor in determining his veracity. And when a hearing officer who saw and heard a witness makes an adverse credibility determination and supports it with specific findings, an appellate court ordinarily should accord it significant respect.

Aguilar-Solís v. INS, 168 F.3d 565, 570-71 (1st Cir.1999) (internal citation omitted). In other words, the strong deference we owe to BIA findings of fact under the substantial evidence standard applies even more so to adverse credibility determinations based on witness demeanor. See Rodríguez Del Camen v. Gonzáles, 441 F.3d 41, 43 (1st Cir.2006) (“Matters of witness credibility and demeanor are peculiarly for the factfinder,” and credibility determinations supported with specific findings are treated “ ‘with great respect.’ ” (quoting Laurent, 359 F.3d at 64)).

Here, the IJ said he had “carefully observed [Castañeda’s] demeanor” and had concluded that Castañeda was not telling the truth about his role in the massacre. More specifically, the IJ had observed Castañeda’s demeanor during Ms testimony about radio contact and found Mm “extremely incredible.”

The IJ also found that Castañeda was “extremely vague, evasive, and non-responsive” when cross-examined by the government attorney about the massacre and the events following and about his lack of documentary evidence of the court martial. The BIA found no error in any of these findings, and its decision was reasonable, as the following examples show.

During Castañeda’s testimony, he appeared to contradict himself, and failed to give direct answers, in response to a series of questions about whether he had had the other patrols’ radio frequencies on the day of the massacre. Castañeda initially indicated that he had had such frequencies, and then stated that he had not. At some point, the IJ stepped in to get a straight answer from an evasive witness:

IJ: Sir, it still hasn’t been answered though. Could you contact the other two patrols by radio, yes or no?

C: No. I did not on my radio.

IJ: That’s not what I asked. Could you contact them by radio, yes or no?

C: No. I could not.

Later in the cross-examination, Castañe-da was asked what had happened in the village the day of the massacre, and he replied: “One of the patrols ... committed excesses.” He then claimed that he did not know until his appearance before the IJ how many villagers had been killed, and that such information was never discussed during his court martial. Information about the number of deaths, however, appears to have been widely circulated following the massacre, and such information was discussed in the report of the Peruvian Human Rights Commission, which led to the court martial.⁴⁸

Still later, Castañeda was confronted with a 1985 State Department report stating that “an army sub-lieutenant and three other officers” were responsible for the massacre. The government attorney asked him whether the report was “talking about you.” Again, the witness was evasive:

C: They are talking about the patrol head of that commission.

IJ: Sir, are they talking about you, yes or no?

C: Yes. They are talking about me.

There is also substantial record support for the finding that Castañeda was evasive about his knowledge about Hurtado.⁴⁹ In response to repeated questions, Castañeda denied having any positive knowledge about whether Hurtado had been discharged from the military. Yet when confronted with a human rights report from the U.S. State Department, he admitted to “finding] out through public media when [he] was here in the United States” that Hurtado had been promoted to captain.

These examples suffice to demonstrate a substantial basis in the record for the conclusion that Castañeda was evasive and non-responsive, and for the BIA’s refusal to believe his story.

Castañeda and the majority take issue with several of the subsidiary bases for the BIA’s and IJ’s adverse credibility determinations. Castañeda argues, for instance, that the BIA erroneously disbelieved his testimony about lack of radio contact with other patrols, and that it erroneously found him evasive regarding the extent of human rights violations committed by the Peruvian military. The majority agrees and adopts similar criticisms of the BIA.

As to the radio issue, the BIA found unlikely Castañeda’s testimony that on the day of the massacre, not only did he not have radio contact with the other patrols, but he did not even know the frequency they were using. The BIA relied in part on Castañeda’s earlier testimony that he sometimes had had direct radio contact with other patrols; the BIA said this testimony was “inconsistent” with Castañeda’s assertions about the day of the massacre. In a sense, it was not “inconsistent” for Castañeda to deny radio contact on the day of the massacre, since, as the majority notes, he had said that he only sometimes had radio contact with other patrols. But this criticism misses the BIA’s point that this operation was a coordinated combat mission, out of the ordinary, and it was much more likely in such a circumstance that the two attacking patrols would have at least had the capability to contact Cas-tañeda’s patrol by radio, both to provide Castañeda with information on escaping guerrillas and to call in the assistance of Castañeda’s patrol if needed. If radio contact among patrols was at least sometimes used, this would have seemed the time to make contact possible.

As to the issue of human rights violations, Castañeda did, after some questioning, admit that he was aware of extrajudicial killings by the Peruvian army, and that members of the armed forces had been responsible for disappearances, torture, and arbitrary detentions. But one could view the record as showing evasiveness. The BIA could reasonably have considered that Castañeda’s new-found relative candor as to human rights abuses contrasted sharply with Castañeda’s initial asylum interview, when he was asked the same question and found both not candid and not credible. The BIA could have considered that Castañeda might have wanted to downplay the significance of his knowledge of a pattern of human rights abuses by the Peruvian military before the events in question.

The BIA’s additional bases for its adverse credibility determination are also substantially supported. The BIA agreed with the IJ that Castañeda’s failure to provide more documentary evidence from his court martial reflected negatively on his credibility because such evidence should have been available to him. See Estrada-Henao v. Gonzáles, 453 F.3d 38, 40 (1st Cir.2006) (holding that lack of documentation is “not automatically fatal” but that “where documentation would naturally be expected, its lack can count against the applicant”). Castañeda was asked during proceedings before the IJ whether there was any written accounting of the facts that would support his account of the day of the massacre. He said he had submitted documents to the court martial, but did not have them for the immigration proceeding. Given the formal nature of the court martial and the fact that Cas-tañeda was able to procure the dismissal-of-charges document and other documents from Perú, the BIA’s conclusion that Cas-tañeda’s credibility was weakened by his failure to produce additional documentation is supported.

The BIA also upheld the IJ’s conclusion that given Castañeda’s position as head of one patrol in the joint operation, his “entire testimony regarding how he found out about [the massacre] only from the media is wholly incredible.” The IJ reasonably concluded that it was unlikely that information that two patrols had just killed over five dozen men, women, and children would not find its way to the third patrol that participated in the same operation, and particularly to the head of that patrol. The practice of debriefings only made it even more likely that Castañeda would have learned about the atrocities engaged in by the other patrols before the media did.⁵⁰

In short, the BIA’s adverse credibility determination survives substantial evidence review, and that requires affir-mance. The majority sees in the record insufficient evidence of Castañeda’s participation in the massacre, finding the evidence of evasiveness and inconsistencies to be too ambiguous. The majority forgets, however, that the burden was on Castañe-da to prove he had not assisted or otherwise participated, and that therefore the question is not whether there was sufficient evidence that he participated, but whether there was sufficient evidence that he did not. Since there is substantial evidence to support the BIA’s finding that Castañeda’s credibility was questionable, and since Castañeda relied almost exclusively on his testimony for evidence, the BIA’s decision must be affirmed.

B. Signifícame of the Peruvian Court’s Dismissal of Charges

Castañeda and the majority fault the BIA for failing to accord the proper weight to the Peruvian court’s dismissal of the court martial charges against Castañeda. They are mistaken, however, in arguing that the dismissal is itself sufficient to carry his burden of proof, and indeed, the BIA was not mistaken in according the bare dismissal very little weight.

Castañeda’s main argument is that the BIA erred in concluding that the Peruvian court martial was a whitewash and in “dis-regardfing]” that court’s dismissal of the charges against him. The BIA did not disregard the Peruvian judgment; it merely evaluated that evidence and found it not to be particularly helpful for Castañeda. The BIA also did not rest its conclusion as to the dismissal’s limited significance on a finding that the court martial was a fraud. The BIA did say that the IJ had so concluded, and noted that “there [was] evidence” to support the conclusion.⁵¹ The BIA then gave two other reasons. First, there was no evidence as to the grounds on which charges were dismissed, and a dismissal of criminal charges may not be the result of an acquittal on the merits. Second, dismissal of criminal charges on any grounds — even acquittal on the merits — is not the same as a finding of innocence on those particular charges. Given the divergent burdens of proof, the mere fact that a defendant has not been proven guilty beyond a reasonable doubt does not mean he is more likely innocent than not.

The BIA’s reasoning is entirely sound. Furthermore, Castañeda was charged in the court martial with homicide and abuse of authority. The elements of those crimes are not the same as the standards under immigration law to find an alien statutorily ineligible for asylum. A finding that an alien was acquitted on homicide and abuse-of-authority charges does not establish that he did not assist or otherwise participate in persecution. The BIA was justified in affording only limited significance to the dismissal of charges against Castañeda.⁵²

The majority improperly faults the BIA for the “unreasonable” requirement of “an actual finding of ‘innocent.’ ” But this is not what the BIA required. The BIA required Castañeda to provide a different showing — that he did not assist or otherwise participate in the persecution. This requirement is not only reasonable; it is mandated. See 8 C.F.R. §§ 1208.13(c)(2)(ii), 1208.16(d)(2).

C. The Majority’s Failure to Remand

Even if the majority were correct in finding fault with the BIA’s credibility decision, the majority commits serious error in (a) not remanding to the BIA for further consideration, (b) determining itself that Castañeda is credible, and (c) finding that his credible testimony means that, as a matter of law, he has met his burden and the BIA may no longer consider the question of whether Castañeda assisted or otherwise participated in persecution. In addition, and for other reasons, the issue of statutory interpretation implicit in the majority’s holding should also have been remanded, as discussed later.

When a reviewing court finds flaws in the reasoning of the BIA as to the evidence of record, it is ordinarily required to remand to the agency for further investigation or explanation. The “ordinary” rule is that if a federal court reviewing an agency action finds that

the record before the agency does not support the agency action, or ... [that] the agency has not considered all relevant factors, or ... the reviewing court simply cannot evaluate the challenged agency action on the basis of the record before it, the proper course, except in rare circumstances, is to remand to the agency for additional investigation or explanation.

Fla. Power & Light Co. v. Lorion, 470 U.S. 729, 744, 105 S.Ct. 1598, 84 L.Ed.2d 643 (1985).

A court of appeals reviewing an agency decision “ ‘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.’ ” INS v. Ventura, 537 U.S. 12, 16-17, 123 S.Ct. 353, 154 L.Ed.2d 272 (2002) (per curiam) (quoting Lorion, 470 U.S. at 744, 105 S.Ct. 1598); see also R.I. Higher Educ. Assistance Auth. v. Secretary, U.S. Dept. of Educ., 929 F.2d 844, 857 (1st Cir.1991) (“[A] reviewing court, having determined that an administrative agency did not consider all the relevant factors, should ordinarily remand the matter to the agency rather than compensating for the agency’s oversight by launching a free-wheeling judicial inquiry into the merits.”). “When an agency ... has provided insufficient explanation for its action, the reviewing court ordinarily should remand the case to the agency.” Seavey v. Barnhart, 276 F.3d 1, 12 (1st Cir.2001).

We have particularly applied this rule in immigration cases. In Cordero-Trejo v. INS, 40 F.3d 482 (1st Cir.1994), we held, as does the majority here, that the IJ’s adverse credibility determination was not supported by substantial evidence, because the claimed inconsistencies were not supported by the record. Id. at 491. We then followed the “usual practice and rule” of “remand[ing] for further consideration by the Board,” even where “the IJ’s conclusions are not drawn from any perspectives offered by the unique vantage point of the factfinder, such as witness demean- or, conflicting or confused testimony, etc., from which credibility is typically assessed.” Id. at 491-92. The majority violates circuit law in failing to remand here.

Similarly, in Gailius v. INS, 147 F.3d 34 (1st Cir.1998), this court found that the BIA’s conclusion was not supported by substantial evidence, in that the BIA had failed to address portions of the petitioner’s evidence that would corroborate his testimony. Id. at 46-47. The majority here points, for example, to “a portion of the [State Department] report the government failed to read at the hearing.” We held in Gailius that remand was the appropriate remedy and that in general, remand is appropriate “when a reviewing court cannot sustain the agency’s decision because it has failed to offer legally sufficient reasons for its decision.” Id. at 47.

The Supreme Court has stated that the “ordinary remand requirement” has “obvious importance in the immigration context.” Ventura, 537 U.S. at 16-17, 123 S.Ct. 353; see also Gonzáles v. Thomas, — U.S. -, -, 126 S.Ct. 1613, 1615, 164 L.Ed.2d 358 (2006) (per curiam). This is because immigration decisions often involve matters on which the BIA can bring its particular expertise to bear. See Ventura, 537 U.S. at 17, 123 S.Ct. 353 (finding that agency expertise on country conditions supported remand as a remedy). Indeed, this court has held that the more an issue relates to matters within an administrative agency’s particular expertise, the less willing a court should be to decide an issue itself. See R.I. Higher Educ. Assistance Auth., 929 F.2d at 857 (“[T]he more imbricated a matter, the more cautious a reviewing court should be about attempting to resolve the issue itself, rather than remanding to the agency.”).

The ordinary rule applies with full force here, and even if in rare circumstances a reviewing court could eschew remand, this is not such a case. Even if the BIA’s reasoning as to the finding that Castañeda lacked credibility were flawed, the majority is not correct in finding that the evidence compels exactly the opposite conclusion: that Castañeda is credible. That is especially so because Castañeda bears the burden to show his credibility. The BIA did not find it necessary to comment on all of the reasons the IJ gave for finding Castañeda not credible — for example, on the point that Castañeda was vague and evasive concerning how he learned about the massacre. The BIA must be given the opportunity to reconsider such evidence. An appellate court should not, as the majority does, make a credibility finding, especially in a case such as this one involving “witness demeanor” and “conflicting or confused testimony.” Cf. Cordero-Trejo, 40 F.3d at 491. The majority cites no case in which an appellate court made its own credibility finding.

Moreover, there are other reasons to remand. Although the majority faults the BIA for allegedly failing to accord appropriate deference to the judgment of the Peruvian military court, the majority apparently does not assert that this judgment alone satisfies Castañeda’s burden. The BIA must be allowed to reconsider its decision in light of the majority’s ruling on the appropriate view of the foreign judgment.

V.

The majority further errs in holding that Castañeda met his burden to show that he did not assist or otherwise participate in persecution. The majority improperly reads into the BIA’s decision a holding it did not make, then improperly decides a novel issue of statutory interpretation, and finally improperly applies its newly formulated rule to the facts of this case.

The majority reads the BIA’s opinion to have stated a holding assuming that Cas-tañeda did not know about the massacre. No such holding can be found. The BIA’s opinion addressed the dismissal of the court martial charges against Castañeda and the correctness of the credibility determination made by the IJ. The only sentence discussing an alternate holding states that in the BIA’s opinion,

even if the respondent testified credibly that he did not harm or execute anyone on August 8, 1985, there is substantial evidence to support the Immigration Judge’s conclusion that the respondent did aid in the persecution of others by helping to confine them, inasmuch as his patrol was charged with blocking an escape route for the villagers and people located in the emergency zone.

(emphasis added). The majority finds that this lone sentence, together with its citation to a page in the IJ’s opinion and to In re Rodríguez-Majano, 19 I. & N. Dec. 811 (BIA 1988), suffices to open the door to the majority’s conclusions: (a) that as a matter of law, a person without prior or contemporaneous knowledge of the persecution cannot have assisted or otherwise participated in the persecution within the meaning of the statutes at issue here by blocking an escape route for those persecuted, and (b) that because the record purportedly compels the conclusion that Castañeda had no such knowledge in this case, Cas-tañeda has met his burden to show that he did not assist or otherwise participate in persecution.

The majority is simply wrong in all of this. First, the majority is wrong to find in the BIA’s opinion any holding that takes into account the state of Castañeda’s knowledge. The BIA did not say that it was assuming all of Castañeda’s testimony was credible, only that even if Castañeda had not harmed or killed anyone on the day of the massacre, this did not establish that he did not assist or otherwise participate in the massacre. The cited portion of the IJ’s opinion merely supports this proposition by finding Castañeda’s “motivation and intent” largely irrelevant: the majority agrees that this is the established rule. Rodríguez-Majano also supports this proposition by incidentally mentioning that one can participate in persecution without directly harming anyone by having an “objective effect.” Id. at 815. Neither the IJ nor Rodríguez-Majano considered whether knowledge has any bearing on a finding that an alien assisted or otherwise participated in persecution. The BIA did not consider this point either.

If there is any ambiguity as to what the BIA held, or questions about the basis for the BIA’s holding, the majority is required to remand to the BIA for further explanation. See Halo v. Gonzáles, 419 F.3d 15, 16 (1st Cir.2005) (“The BIA offered no further explanation for its conclusions. Lacking such explanation, and given that the final agency decision does not rest on a lack-of-credibility determination, we are left with significant questions about the justifications for the denial. We therefore vacate the BIA’s order and remand.”). The majority is not free to decide the case on grounds not addressed by the BIA. See Gailius, 147 F.3d at 44 (“ ‘[A] reviewing court ... must judge the propriety of [administrative] action solely by the grounds invoked by the agency.’ ” (alteration in original) (quoting SEC v. Chenery Corp., 332 U.S. 194, 196, 67 S.Ct. 1575, 91 L.Ed. 1995 (1947))).

The majority attempts to obscure the fact that it is deciding the issue, unaddressed by the BIA, of whether knowledge of a massacre can be a requirement to a finding that an alien assisted or otherwise participated in that massacre. The majority does so by repeatedly asserting its prior improper conclusion that Castañeda did not know about the massacre, and then assuming this lack of knowledge as a background fact in its analysis of the “objective effect” of Castañeda’s actions.

It may be that the majority is saying that even if Castañeda knew full well that the villagers would be massacred, he still would not have assisted or otherwise participated in the massacre. If that is what the majority means, its holding is both irrational and flatly contrary to the “objective effect” line of cases that the majority cites.⁵³ The majority cites no case holding that those with lesser tasks — such as blocking escape routes — should not be barred as persecutors.⁵⁴ To be clear, the calculus might be different if Castañeda did not know about the atrocities, but that issue is one for the BIA to address in the first instance.

The logic of the majority’s opinion thus depends upon its reaching and resolving an issue of statutory interpretation, namely whether knowledge is a factor that must be considered in determining whether an alien “assisted or otherwise participated in” persecution within the meaning of 8 U.S.C. § 1158(b)(2)(A)(i) and 8 U.S.C. § 1231(b)(3)(B)(i). The majority errs in deciding this issue without a ruling on it by the BIA.

The principles in this ease are clear. “We review de novo an agency’s construction of a statute that it administers, although subject to established principles of deference.” Griffiths v. INS, 243 F.3d 45, 49 (1st Cir.2001). “Under those principles of deference, if the intent of Congress is clear, it must govern, but where the statute is silent or ambiguous on an issue, the question for the court is whether the agency’s interpretation is based on a permissible construction of the statute.” Id. (citing, inter alia, Chevron, 467 U.S. at 842-43, 104 S.Ct. 2778). “Since agency officials acting in the immigration context ‘exercise especially sensitive political functions that implicate questions of foreign relations,’ deference to administrative expertise is particularly appropriate.” Id. (internal quotations omitted) (quoting Abudu, 485 U.S. at 110, 108 S.Ct. 904).

The BIA must be given the first opportunity to rule on questions of statutory interpretation. It is only through such a ruling that the BIA can fulfill its role in “ ‘the formulation of policy and the making of rules to fill any gap left, implicitly or explicitly, by Congress.’ ” Elien v. Ashcroft, 364 F.3d 392, 396 (1st Cir.2004) (quoting Morton v. Ruiz, 415 U.S. 199, 231, 94 S.Ct. 1055, 39 L.Ed.2d 270 (1974)). Such a ruling is also necessary for the BIA to bring to bear its specialized “expertise in matters of foreign policy.” Id.

The BIA has not had this opportunity here. The record shows that this know-ingness issue of statutory interpretation was not raised or argued before it, and was only the subject of supplemental briefing in this court, after this court sua sponte raised the issue at oral argument. As a result, because Castañeda did not clearly present this issue to the BIA, we in fact lack the authority to even consider it. See Olujoke v. Gonzáles, 411 F.3d 16, 23 (1st Cir.2005). Further, even if Castañeda had exhausted his administrative remedies, he would have waived the argument by failing to make it in his initial brief to this court. See Cipes v. Mikasa, Inc., 439 F.3d 52, 55 (1st Cir.2006).

The majority cannot defend its decision to interpret the statute itself as encompassing a knowledge requirement on the ground that the role it assigns knowledge is required by the plain meaning of the statute. That argument is foreclosed by Fedorenko v. United States, 449 U.S. 490, 101 S.Ct. 737, 66 L.Ed.2d 686 (1981). Fe-dorenko involved a similar provision, which precluded admission to the United States under the Displaced Persons Act of aliens “who had assisted the enemy in persecuting” civilians. Id. at 495, 101 S.Ct. 737. Fedorenko, who said he was a prisoner of war who had been forced to serve as a guard in a Nazi concentration camp, argued the exclusion did not apply to him because his service had been involuntary. The Supreme Court- held that the statute did not include the term “voluntary” and the Court was “not at liberty to imply a condition” into the statute. Id. at 513, 101 S.Ct. 737. The Court commented that the administrative interpretation of the exclusion was that it applied whether or not the assistance was voluntary and whether or not the alien had participated in the persecution himself. Id. at 499, 101 S.Ct. 737. The Court also noted the need for strict compliance with all statutory preconditions to naturalization given that “Congress alone has ... constitutional authority” in this area. Id. at 506, 101 S.Ct. 737. Finally, the Court held that the federal courts lacked equitable power to vary the terms of the statute. Id. at 516-18, 101 S.Ct. 737.

The plain text of the statutes at issue here contains no “knowledge” qualifier to a finding that an alien “assisted or otherwise participated” in persecution. Indeed, the majority does not claim that the statute unambiguously contains a knowledge qualifier. Rather, the majority effectively holds that knowledge is a factor which the BIA must consider. Not only is such a factor nowhere to be found in the statutory text, but determining which factors to consider in applying a statute is even more clearly the province of the BIA. See Aguirre-Aguirre, 526 U.S. at 424, 119 S.Ct. 1439 (faulting the Ninth Circuit for “fail[ing] to accord the required level of deference” to the BIA by requiring that it “examin[e] additional factors”). Such factors are to be developed by the BIA “ ‘through a process of case-by-case adjudication’ ”, id. at 425, 119 S.Ct. 1439 (quoting INS v. Cardoza-Fonseca, 480 U.S. 421, 448-49, 107 S.Ct. 1207, 94 L.Ed.2d 434 (1987)), and cannot be found in the statutory text.

Finally, the majority errs in making its own findings of fact and then in determining that these facts dictate a holding that Castañeda did not assist or otherwise participate in persecution. It is the BIA’s role to “determin[e] the facts and decid[e] whether the facts as found fall within a statutory term.” Thomas, 126 S.Ct. at 1615. The majority claims that

we are presented with the following set of facts: (1) the uncontradicted evidence is that the purpose of the Operation was directed at the Shining Path guerillas in the village, not civilians; (2) Castañeda and his men were hidden in the jungle three to five miles from the village and about thirty meters from the path they were guarding; (3) during the time they were there, they did not see or hear anyone, nor did they fire any shots; (4) they never entered the village and never directly or indirectly participated in the killings of the villagers, which was carried out by a rogue patrol led by Hurta-do in a previously unplanned manner; (5) Castañeda and his men lacked any control or authority over this rogue patrol; and (6) they did not know what occurred in the village during the Operation.

This is not true. Neither the IJ nor the BIA found or assumed that Castañeda had established that there was no order allowing the patrols to harm civilians who were known to have harbored the Shining Path, that Hurtado’s patrol was a rogue patrol acting entirely on its own, or that Castañe-da had no knowledge of the massacre before or as it was happening. The record does not compel any of these conclusions. Even if the record compelled these conclusions, the BIA must be given the opportunity to apply the majority’s newly formulated rule to the facts.

VI.

For the reasons stated above, the majority opinion has not only erred, but it is at odds with the statutory mandate, the decisions of the Supreme Court, and our own prior case law. Asylum is not a right; it is a privilege. See Fedorenko, 449 U.S. at 518, 101 S.Ct. 737 (“ ‘An alien who seeks political rights as a member of this Nation can rightfully obtain them only upon the terms and conditions specified by Congress.’ ” (quoting United States v. Ginsberg, 243 U.S. 472, 474-75, 37 S.Ct. 422, 61 L.Ed. 853 (1917))). Congress has decided that privilege does not extend to those who assist or otherwise participate in persecution within the meaning of the statute, as reasonably interpreted by the BIA. I respectfully dissent.

ADDENDUM

U.S. Department of Justice

Executive Office for Immigration Review

Falls Church, Virginia 22041

Decision of Board of Immigration Appeals

Files: A70 700 784 — Boston

A70 700 783

A70 700 782

A70 701 180

Date: Sep 09 2005

In re: DAVID EDUARDO CASTANEDA CASTILLO

CARMEN JULIA DE LA CRUZ CASTANEDA

PIERA DINA CASTANEDA

PIA MARIBEL CASTANEDA

IN REMOVAL PROCEEDINGS APPEAL

ON BEHALF OF RESPONDENTS: Jane C. Chiang, Esquire

ON BEHALF OF DHS: John P. Marley Assistant Chief Counsel

CHARGE:

Notice: Sec. 237(a)(1)(B), I & N Act [8 U.S.C. § 1227(a)(1)(B)] - In the United States in violation of law (all respondents)

APPLICATION: Asylum; withholding of removal; voluntary departure

In a decision dated October 4, 2004, an Immigration Judge denied the respondents’ application for asylum and withholding of removal and granted the three female respondents voluntary departure (70 700 783, 70 700 782, and 70 701 180) but denied such relief to the lead respondent (70 700 784).¹ The respondents, a family consisting of father (lead respondent), mother, and two female children have appealed the Immigration Judge’s decision; the appeal will be dismissed.

The Immigration Judge did not actually address the respondents’ contention that they would be persecuted by the Shining Path if returned to Peru (I.J. at 44). Rather, the Immigration Judge denied the respondents’ applications for asylum and withholding of removal under sections 208 and 241(b)(3) of the Immigration and Nationality Act, 8 U.S.C. §§ 1158, 1231(b)(3), respectively, because he found the lead respondent² statutorily ineligible for having ordered, incited, assisted, or otherwise participated in the persecution of others on account of race, religion, nationality, membership in a particular social group, or political opinion. See sections 208(b)(2)(A)(i) and 241(b)(3)(B)® of the Act, 8 U.S.C. §§ 1158(b)(2)(A)®, 1231(b)(3)(B)®. Specifically, the Immigration Judge found that the lead respondent, who was a member of the armed forces in Peru and patrol leader of one of four units assigned to patrol an “emergency zone” in the state of Accomarca (I.J. at 7), participated in a military operation on August 8, 1985, which resulted in a massacre of innocent civilians, an event which a human rights commission has labeled as genocide (Exh. 35; I.J. at 8). In concluding its investigation, the human rights commission apparently recommended that the respondent be tried in civil court, as opposed to a military court martial (I.J. at 9; Exh. 35), and as noted by the Immigration Judge, the Department of State Country Reports for 1985 (Exh. 40) detail that the respondent and other military leaders involved in the incident were charged in both civil and military courts (I.J. at 9-11). The respondent admitted that he was one of the leaders referred to in the Country Reports which also indicated that the Peruvian Supreme Court was to decide final jurisdiction (Tr. at 390-93; Exh. 40; I.J. at 11, 25). Although no evidence confirmed the outcome of the jurisdictional issue, the Immigration Judge concluded that the Supreme Court had decided that the respondent would be tried in military court because the respondent was able to produce a copy of the Peruvian Supreme Council of Military Justice’s affirmance of a lower court’s dismissal of murder charges against him (I.J. at 11; Exh. 24).

On appeal, the respondent argues that the Immigration Judge erred in not giving deference to the foreign military court’s decision dismissing the charges against him, a decision which goes against the “principle of international comity” (Respondents’ Br. at 9-10). Furthermore, he asserts that the documentary evidence relied on by the Immigration Judge to deem the military court’s dismissal as having less weight were general in nature and did not contradict the respondent’s evidence (Respondents’ Br. at 10-11).

As noted by the Immigration Judge, there is some doubt about the validity of the respondent’s evidence inasmuch as it was not properly authenticated and also appeared to have been issued by the clerk’s office of the Supreme Council of Military Justice and not by the Supreme Council itself (I.J. at 11-12). Furthermore, no other documents regarding the respondent’s trial were submitted in order to determine the exact grounds for dismissing the respondent’s military court martial and the documentary evidence of record indicates that there is evidence that the military tribunal was a way in which to grant impunity to the alleged persecutors who took part in the August 8, 1985, massacre (Exh. 41; I.J. at 34-36; Tr. at 439-46; see also Exh. 43). As pointed out by the Department of Homeland Security (DHS, formerly the Immigration and Naturalization Service), the respondent’s evidence of his dismissal is not necessarily the legal equivalent of a criminal finding that he was innocent, particularly without additional evidence to indicate the court’s reasoning (DHS’s Br. at 4). In sum, we find no error in the Immigration Judge’s conclusion that the respondent failed to meet his burden of proving that he did not assist or otherwise participate in the persecution of villagers in Accomarca based on their political opinion (I.J. at 36); even if the respondent testified credibly that he did not harm or execute anyone on August 8, 1985, there is substantial evidence to support the Immigration Judge’s conclusion that the respondent did aid in the persecution of others by helping to confine them, inasmuch as his patrol was charged with blocking an escape route for the villagers and people located in the emergency zone (I.J. at 36; See Tr. at 201, 351-55). Matter of Rodriguez-Majano, 19 I & N Dec. 811 (BIA 1988).

In addition, we find no clear error in the Immigration Judge’s determination that the respondent was not completely credible and affirm the denial of relief on this basis as well (I.J. at 37-43). The Immigration Judge found the respondent evasive regarding his military court martial and, as already noted, the respondent failed to provide documentary evidence such as a transcript of that hearing which would have shed greater light on his activities during the August 8, 1985, patrol (I.J. at 37-38, 40; Tr. at 411-32, 435-53, 456-73). The respondent’s assertion on appeal that he may not have had access to these documents does not seem plausible nor does he provide any explanation for not providing them (Respondents’ Br. at 13).

In addition, the Immigration Judge relied on the fact that the respondent gave inconsistent testimony regarding whether and when he was in radio contact with the other patrols involved in the August 8 patrol, indicating early in his testimony that it was important always to have radio contact and later stating that he was not in radio contact with the patrols on the day of the massacre (I.J. at 38-41; Tr. at 314-16, 329-34, 366-75); the Immigration Judge noted that the respondent’s demeanor during this portion of questioning was also indicative of a lack of veracity (I.J. at 40). The respondent argues that the Immigration Judge misconstrued his statements because he testified that radio contact between patrols was made sometimes (Respondents’ Br. at 14; Tr. at 314) and, therefore, erroneous when the Immigration Judge stated that it was “completely illogical” for the respondent to have no contact during the attack on August 8 (Respondents’ Br. at 14; I.J. at 41). We disagree with the respondent that the Immigration Judge’s determination “ventured into the realm of Peruvian military tactics, an arena outside of his knowledge and expertise” (Respondents’ Br. at 15). Based on the respondent’s own testimony, it is clear that radio contact was important for purposes of coordination as well as efficiency and protection of the patrols (See, e.g., Tr. at 314).

Finally, the Immigration Judge found the respondent evasive regarding the extent of human rights violations by the military which has been documented by the Department of State (I.J. at 41-42) and that the respondent was not forthcoming about the judgement entered against one of his fellow commanders (I.J. at 42^43). The respondent argues that knowledge of another’s fate is irrelevant and immaterial to his case (Respondents’ Br. at 15); he states that the Immigration Judge misconstrued the respondent’s testimony about his fellow commander’s fate (Respondents’ Br. at 17 (citing Tr. at 440)). While it may be true that the respondent, a patrol leader, lacked knowledge regarding the events which took place following the August 8 massacre, including whether and how many people were killed as well as the outcome of his fellow officers’ court mar-tials, we also cannot find clear error in the Immigration Judge’s conclusion that the respondent was evasive. In light of obvious media attention and international documentation of the massacre, as well as the respondent’s military position and subsequent accusations about his role, it was reasonable to expect him to provide answers regarding what occurred in Peru on and after August 8, particularly since he failed to provide adequate documentary evidence of his innocence.

ORDER: The respondents’ appeal is dismissed.

FURTHER ORDER: Pursuant to the Immigration Judge’s order and conditioned upon compliance with conditions set forth by the Immigration Judge and the statute, the female respondents (A70 700 783, 70 700 782, and 70 701 180) are permitted to voluntarily depart from the United States, without expense to the Government, within 30 days from the date of this order or any extension beyond that time as may be granted by the Department of Homeland Security (DHS). See section 24013(b) of the Immigration and Nationality Act; 8 C.F.R. §§ 1240.26(c), (f). In the event the respondents fail to so depart, the respondents shall be removed as provided in the Immigration Judge’s order.

NOTICE: If the respondents fail to depart the United States within the time period specified, or any extensions granted by the DHS, the respondents shall be subject to a civil penalty of not less than $1,000 and not more than $5,000, and shall be ineligible for a period of 10 years for any further relief under section 240B and sections 240A, 245, 248, and 249 of the Immigration and Nationality Act. See section 240B(d) of the Act.

/s/ Roger A. Pauley FOR THE BOARD

41 . Thus, for example, even if it might be true that translation difficulties could explain some of the problems with Castañeda's testimony — a point not raised by Castañeda himself — the issues raised by the BIA and the IJ are such that it was entirely reasonable to conclude otherwise. Moreover, those who actually heard Castañeda's testimony are in a much better position than this court to weigh the effects of any translation difficulties. Similarly, the majority's assertions that Cas-tañeda misunderstood some questions or was interrupted by the judge might be possible readings of the record, but the record certainly does not compel such conclusions, and we are in a poor position to judge such matters from the transcript alone.

42 . The majority objects to my account of the facts as unduly recounting "the actions of persons for whom Castañeda was not responsible and who were without his command or authority.” I am merely recounting, however, the persecution that Castañeda is supposed to have assisted or participated in. The evidence of such persecution was properly before the IJ and BIA. The majority also appears to complain that the extent of the persecution is irrelevant and should be ignored. There is no rule that the BIA cannot weigh the extent of persecution. Again, that is a decision for the BIA.

43 .The majority notes that any such orders must have come from those with a higher rank than Hurtado, rather than from Castañe-da. The point, however, is that the BIA could consider that if Hurtado was acting on orders, rather than in the heat of the moment, it is much more likely that Castañeda would have known about the massacre before it happened.

44 . It is unclear whether he also was charged with first degree murder.

45 . The IJ and BIA never reached the issue of whether Castañeda could meet the standards for asylum or withholding of removal. See 8 U.S.C. §§ 1158(b)(1), 1231(b)(3)(A). Because the BIA did not reach the issue, I do not address it either.

46 . The majority claims that the remainder of Castañeda's response makes “clear that Cas-tañeda misunderstood the question.” This is far from clear. An equally plausible interpretation of the language quoted by the majority is that Castañeda testified that while he had the frequencies, he did not need to use them. Between two reasonable interpretations, we must defer to the one adopted by the IJ and BIA.

47 . Castañeda also testified several times in 2001 and 2002 before another IJ. After the case was reassigned, the second IJ agreed to re-hear Castañeda's testimony at the request of Castañeda’s counsel.

48 . The majority reads Castañeda’s testimony here to have been that he did not know the exact number because "the exact number could not be determined.” As the majority notes, however, this remains true "[t]o this day," and hence such a reading is not consistent with Castañeda’s assertion that he did not learn the exact number until the day of his hearing before the IJ. In any event, regardless of whether the majority's reading is a plausible one, it is equally plausible, and reasonable on the part of the IJ and BIA, to read Castañeda’s testimony as unduly attempting to minimize his knowledge of the massacre.

49 . The majority's claim — that the issue of Castañeda's knowledge about Hurtado is a "red herring”' — is puzzling. As the majority itself notes, "Castañeda's case is based on his disassociation from Hurtado's actions.” The BIA and IJ were entirely justified in presuming that evasiveness as to Castañeda's later knowledge about Hurtado's fate indicated a lack of credibility regarding his earlier knowledge about Hurtado's actions.

50 . Of course, such information need not have been "openly discuss[ed]” in order to have been shared.

51 . For example, Lieutenant Hurtado, who led one of the combat patrols into the village, was acquitted of first-degree murder charges and convicted only of abuse of authority. Even if the BIA had rested its conclusion on the ground that the Peruvian proceeding was a fraud, we would owe deference under Chevron to such a conclusion about the significance of a foreign tribunal's proceedings. See INS v. Aguirre-Aguirre, 526 U.S. 415, 424-25, 119 S.Ct. 1439, 143 L.Ed.2d 590 (1999) (applying Chevron deference and noting 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.' " (quoting INS v. Abudu, 485 U.S. 94, 110, 108 S.Ct. 904, 99 L.Ed.2d 90 (1988))).

52 . Castañeda also argues that the BIA erred when it deemed him a persecutor because his “behavior was directed toward the defense of the government, and harm which results from such behavior does not rise to the level of persecution.” As the majority correctly notes, this argument fails.

53 . The case most on point is Naujalis v. INS, 240 F.3d 642, 647 (7th Cir.2001), in which the Seventh Circuit found that a soldier who guarded a railway station had assisted in the persecution engaged in by his battalion. The majority objects that Naujalis is a case about Nazi guards and that Peruvian military guards, unlike Nazi guards, can serve legitimate purposes. The statute makes no such distinction, and in any event, such a distinction has no bearing on the culpability of one who participates in an operation involving persecution.

54 . In only one of the many "objective effect” cases cited by the majority did the court reverse a finding that the alien was a persecutor. That case is Hernández v. Reno, 258 F.3d 806 (8th Cir.2001), a case involving a man forcibly recruited to join a guerrilla movement and turning on whether the man's actions were sufficiently coerced. None of the "objective effect” cases cited by the majority involved aliens who might not have acted knowingly.

1 . Although the respondents did not apply for relief under the Convention Against Torture and conceded that they could not establish that it is more likely than not the Peruvian government would torture them or acquiesce to their torture (I.J. at 32), the Immigration Judge addressed such relief in his decision and found that the respondents had failed to meet their burdens of proof (I.J. at 31-33, 44-46).

2 . All reference to respondent in the singular refer to the lead respondent.