dissenting from the order denying the petition for rehearing en banc:
The second opinion in this case, United States v. Ziegler, 474 F.3d 1184 (9th Cir.2007), is troubling for its serious Fourth Amendment implications, as outlined by Judge W. Fletcher’s persuasive dissent. But the opinion is also troubling because it transgresses the boundaries of our institu*900tional competence and disregards settled rules of appellate review. First as to competence: We may not find facts on appeal; we may only review findings made by the courts below us. See, e.g., Anderson v. City of Bessemer City, 470 U.S. 564, 574, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985) (“The trial judge’s major role is the determination of fact, and with experience in fulfilling that role comes expertise.”); United States v. Lang, 149 F.3d 1044, 1046 (9th Cir.1998) (“[T]he district court is in a superior position to judge the accuracy of witnesses’ recollections and make credibility determinations in cases in which live testimony is presented.” (internal quotation marks omitted)). This difference in institutional competence is also reflected in our Rules. E.g., Fed.R.CivJP. 52(a) (“[D]ue regard shall be given to the opportunity of the trial court to judge of the credibility of the witnesses.”).
The panel affirms the district court’s result, but does so by using facts the district court never found. The opinion “conclude[s]” that Reavis, a corporate officer, consented to the government search of Ziegler’s office. 474 F.3d at 1192. But we, and the Supreme Court, have repeatedly and consistently held that consent to search must be found as a matter of fact. See, e.g., United States v. Mitchell, 322 U.S. 65, 69 n. 2, 64 S.Ct. 896, 88 L.Ed. 1140 (1944); United States v. Spires, 3 F.3d 1234, 1236-37 (9th Cir.1993); United States v. Shaibu, 920 F.2d 1423, 1425 (9th Cir.1990). The district court here made no findings as to consent, resolving the case on other grounds (which the panel now finds insufficient). United States v. Ziegler, No. CR-03-08-BU-RFC, at 3 (D.Mont. Sept. 8, 2004) (“Defendant could not have an objectively reasonable belief that the files he accessed on the Internet were private.... ”).
The panel finds consent based on a single passage of testimony from Schneider, an IT department employee, who claimed that he and his supervisor obtained Reav-is’s approval for going into defendants’s office and copying his hard drive. The panel accepts this as the Gospel Truth, but it did not observe Schneider’s testimony and is not in a position to determine whether he was lying or telling the truth. See Anderson, 470 U.S. at 575, 105 S.Ct. 1504 (“[OJnly the trial judge can be aware of the variations in demeanor and tone of voice that bear so heavily on the listener’s understanding of and belief in what is said.”). The district court made no finding on this point.
Even if Schneider’s testimony can be believed, it is insufficient to show that consent was given to the government. Schneider testified that Reavis told him that it was OK to proceed, but he does not say that this was ever communicated to Agent Kennedy. Again, the district court made no finding as to whether Agent Kennedy had been told of Reavis’s apparent consent at the time that he ordered Schneider to copy Ziegler’s hard drive. Most likely, Agent Kennedy was not aware of Reavis’s statement as the conversation among Schneider, his supervisor and Reavis took place after Kennedy had already ordered them to copy the hard drive. See 474 F.3d at 1192 (“[W]hen I returned from the meeting with Agent Kennedy, I spoke to [my supervisor], and then we, in turn, both went up and spoke to Ronald Reavis. Explained the situation to him. Said that, you know, [the FBI] wanted a backup made of this information .... [H]e said that as an officer of the company, he was okay with that.” (emphasis added)). These are matters that the district court is competent to sort out; we are not.
Finally, even if we assume that Reavis did give consent, and that consent was communicated to Kennedy, there is yet *901another factual hurdle: Did Reavis have authority to consent to a search of Ziegler’s office? The majority resolves this by finding that Reavis was Ziegler’s “superior[ ]” at the company, and thus presumably had authority to enter Ziegler’s office. 474 F.3d at 1192. But both of these are contested issues of fact that were not resolved by the district court.
As to being “superior” to Ziegler, Judge Fletcher correctly points out that the only testimony we have is that Ziegler and Reavis shared the role of “second-in-command.” Fletcher dissent, at 892-93. The district court made no finding that Reavis was defendant’s superior, and I am aware of no evidence that could have supported it. Certainly, the record is not so unequivocal on this point that a district court’s finding would be superfluous.
Even if Reavis were found to be Ziegler’s superior, that does not necessarily mean that he had authority to enter Ziegler’s office or to give consent to a government search. “[Ajuthority to consent to a search requires a considered judgment of both factual circumstances and legal issues.” United States v. Kim, 105 F.3d 1579, 1582 (9th Cir.1997). Here, the district court made no findings as to whether Reavis, assuming that he was Ziegler’s superior, had the authority to enter the office himself or consent to entry by anyone else. Had Agent Kennedy known of Reavis’s consent, he might been justified in relying on it when ordering the search of Ziegler’s office, but that depends on a close analysis of the facts and circumstances of the case; it is not purely a matter of law that can be determined for the first time on appeal.
We have held that “[tjhe existence of consent to a search is not lightly to be inferred, and is a question of fact to be determined from the totality of circumstances.” United States v. Patacchia, 602 F.2d 218, 219 (9th Cir.1979) (Sneed, J.). We don’t have the institutional competence to find facts: We don’t observe the witnesses and we don’t have the district court’s first-hand familiarity with the record. Where, as here, the factual question is fraught with uncertainty, and that uncertainty has not been resolved by the district court, we cannot find the facts ourselves. At most, we can remand for a determination of the dispute by the district court.
This brings me to the panel’s second institutional transgression: resolving the case on a basis not briefed or argued on appeal. I have examined the appellate briefs closely and find no discussion of consent. Rather, the parties discussed the question of Ziegler’s expectation of privacy in his office. The panel originally ruled for the government on this point, but then realized its error and corrected it in its second opinion. 474 F.3d at 1188-90. The government’s brief did not argue consent as an independent basis for upholding the search, doubtless for good reason: There are no district court findings that would have supported affirmance on that ground. We have long held that issues not briefed or argued on appeal are waived. See Stuard v. Stewart, 401 F.3d 1064, 1067 (9th Cir.2005) (“[W]e are not going to construct an argument for the state sua sponte, depriving Stuard’s counsel of a fair chance to respond to it.”); Smith v. Marsh, 194 F.3d 1045, 1052 (9th Cir.1999) (“[Ojn appeal, arguments not raised by a party in its opening brief are deemed waived.”). We apply that rule with some vigor against criminal defendants, see, e.g., United States v. Romm, 455 F.3d 990, 997 (9th Cir.2006) (rejecting un-raised argument challenging constitutionality of search in child pornography case); we should be no less vigorous in applying it against the government.
*902By plucking consent out of its judicial top hat, when neither party has argued it and the district court made no findings to support it, the panel gives the unfortunate impression that it is seeking to vindicate a result it has reached on other (now repudiated) grounds. It is not our business to reach particular results, nor may we jiggle the rules of procedure to achieve an outcome we prefer. Our responsibility is to apply the law in an objective and impartial manner, and let the chips fall where they may. Here, the government lost the one issue on which it chose to make its stand— Ziegler’s expectation of privacy in his own office. At that point it was our responsibility to reverse the district court and vacate the defendant’s sentence. Appellate review is not a magic wand and we undermine public confidence in the judicial process when we make it look like it is.