dissenting:
I respectfully dissent. Because Mickens has failed to establish that his attorney’s conflict of interest adversely affected his representation, I would affirm the decision of the district court and deny Mickens’ petition for habeas corpus relief.
*219I.
I accept, for the purposes of argument, the majority’s conclusion that Mickens’ attorney labored under an actual conflict of interest. I must disagree, however, with the majority’s conclusion that the Supreme Court’s decision in Wood v. Georgia, 450 U.S. 261, 101 S.Ct. 1097, 67 L.Ed.2d 220 (1981), relieves Mickens from his burden of establishing that his attorney’s conflict of interest adversely affected his representation.
As the majority has noted, a criminal defendant’s Sixth Amendment right to effective assistance of counsel includes a right to counsel unhindered by conflicts of interest. Wood, 450 U.S. at 271, 101 S.Ct. 1097; United States v. Tatum, 943 F.2d 370, 375 (4th Cir.1991). The general standard for ineffective assistance of counsel has most clearly been articulated by the Supreme Court in Strickland v. Washington, 466 U.S. 668, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984). To prevail, a petitioner alleging ineffective assistance of' counsel must establish, (1) that his attorney’s representation fell below an objectively reasonable performance and (2) that the inadequate performance prejudiced the petitioner’s case. Strickland, 466 U.S. at 687, 104 S.Ct. 2052. To establish prejudice, “[t]he defendant must show that there is a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding 'would have been different.” Strickland, 466 U.S. at 694, 104 S.Ct. 2052. This constitutional standard for establishing ineffective assistance of counsel applies equally in collateral ha-beas corpus proceedings as it does on direct appeal. Strickland, 466 U.S. at 697, 104 S.Ct. 2052.
Even as it clarified the applicable standard for ineffective assistance of counsel, however, the Strickland Court harmonized existing case law and noted exceptions to the new standard it articulated, recognizing special circumstances addressed by its prior decisions. The Strickland Court recognized that a claim of ineffective assistance of counsel predicated on a defense attorney’s conflict of interest presents a special case. Strickland, 466 U.S. at 692, 104 S.Ct. 2052. To establish ineffective assistance of counsel predicated on a conflict of interest a defendant must establish only that (1) his attorney labored under an actual conflict of interest that (2) adversely affected the representation. See Cuyler v. Sullivan, 446 U.S. 335, 348, 100 S.Ct. 1708, 64 L.Ed.2d 333 (1980). After a defendant satisfies this two-part test, prejudice is presumed and a defendant need not show that the conflict altered the outcome of the proceeding. Sullivan, 446 U.S. at 349-50, 100 S.Ct. 1708.
In Wood, the Court addressed a conflict of interest issue apparent on the face of the record even though the conflict had not been addressed below. Wood, 450 U.S. at 262-63, 101 S.Ct. 1097. Because the case arose as an appeal from a revocation of probation, the Court analyzed the conflict under due process, rather than Sixth Amendment grounds. Wood, 450 U.S. at 271-72, 101 S.Ct. 1097. Nonetheless, the Court cited Sullivan for the proposition that the Sixth Amendment ensures a right to counsel free from conflict of interest. Wood, 450 U.S. at 271, 101 S.Ct. 1097. Without the benefit of briefing, the Court could not determine whether counsel labored under an actual conflict of interest. Wood, 450 U.S. at 272, 101 S.Ct. 1097. The record was adequate to establish, however, “that the possibility of a conflict of interest was sufficiently apparent at the time of the revocation hearing to impose upon the [trial] court a duty to inquire further.” Wood, 450 U.S. at 272, 101 S.Ct. 1097. The Wood Court thus clarified that trial courts have an obligation to inquire into potential conflicts of interest when they know or reasonably should know of a conflict.
The Supreme Court remanded the case and instructed the trial court “to determine whether the conflict of interest that this record strongly suggests actually existed” but did not expressly instruct the *220trial court to identify an adverse effect. Wood, 450 U.S. at 273, 101 S.Ct. 1097. The majority concludes that, because the Wood Court did not expressly require the trial court to identify an adverse effect on remand, a defendant need not establish an adverse effect when a trial court had a duty to inquire into a conflict of interest but fails to do so. Thus, the majority reads Wood to have overruled the second part of Sullivan’s two-part standard by implication. “Overruling by implication is not favored.” Catawba Indian Tribe of South Carolina v. South Carolina, 978 F.2d 1334, 1347 (4th Cir.1992).
At the risk of self-immolation, I should say that the decision in Wood, to me, has not turned out to be a model of clarity. The Wood Court was concerned about the potential conflict of interest faced by a defense attorney hired and compensated by the defendants’ employer whose interests appeared to diverge from their own. Wood, 450 U.S. at 269-71, 101 S.Ct. 1097. The Court cited Sullivan for the proposition that a defendant’s right to effective counsel includes a right to “representation that is free from conflicts of interest.” Wood, 450 U.S. at 271, 101 S.Ct. 1097. The Court then noted that “[o]n the record before us, we cannot be sure whether counsel was influenced in his basic strategic decisions by the interests of the employer who hired him.” Wood, 450 U.S. at 272, 101 S.Ct. 1097 (emphasis added). If the defendants’ attorney had been so influenced, the Court ruled, “the due process rights of petitioners were not respected.” Wood, 450 U.S. at 272, 101 S.Ct. 1097. That language may only mean that the Court required not just an actual conflict, but also an indication that the conflict adversely affected the counsel’s basic strategic decisions. When read in this light, the Wood Court’s terse instruction to the trial court on remand does not bear the Constitutional weight the majority ascribes to it. As this court has recognized, the two requirements of the Sullivan test, “an actual conflict of interest resulting in an adverse effect on counsel’s performance, are often intertwined, making the factual analyses of them overlap.” United States v. Tatum 943 F.2d 370, 375 (4th Cir.1991). Thus, the Wood Court’s instruction to the trial court amounts to no more than shorthand for an explicit two part test that the Wood Court did not even have occasion to quote in its majority decision much less to overrule. Indeed, subsequent to Wood, found in 450 U.S. 261, 101 S.Ct. 1097, 67 L.Ed.2d 220, the holding of Sullivan was underscored by its citation in Strickland: “Prejudice is presumed only if the defendant demonstrates that counsel ‘actively represented conflicting interests’ and that ‘an actual conflict of interest adversely affected his lawyer’s performance.’ ” Strickland, 466 U.S. at 692, 104 S.Ct. 2052 (quoting Sullivan, 446 U.S. at 348, 100 S.Ct. 1708).
Only in extraordinary circumstances not present here has the Supreme Court reversed a conviction because of an actual conflict of interest without a determination that the conflict of interest adversely affected the representation. The only occasions I have found in which the Supreme Court has not inquired into an adverse effect are those cases in which a defense attorney objected to the conflict of interest and the trial court nonetheless declined to address the conflict. In such a case, the Supreme Court has held, reversal is required without inquiry into either actual conflict or adverse effect. See Holloway v. Arkansas, 435 U.S. 475, 488-89, 98 S.Ct. 1173, 55 L.Ed.2d 426 (1978).
The majority decision is also contrary to this court’s interpretation of Wood. In United States v. Gilliam, 975 F.2d 1050 (4th Cir.1992), this court addressed Wood without concluding, as does the majority here, that the Supreme Court had partially overruled Sullivan. In Gilliam, the defendant did not raise the issue of conflict resulting from multiple representation until he obtained new counsel at the sentencing phase of his trial and, as a result of his delay, the trial court declined to conduct a *221hearing into the possibility of conflict. Gilliam, 975 F.2d at 1054. On appeal, this court concluded that the defendant had alleged an actual conflict and noted that “if the court is aware, or should be aware, of a particular conflict, it should conduct a sua sponte inquiry into its existence.” Gilliam, 975 F.2d at 1053 (quoting United States v. Akinseye, 802 F.2d 740, 744 (4th Cir.1986)). Without citing Wood, the court held that “failure of the court to make inquiry under Rule 44(c) is not, standing alone, sufficient to require reversal of a conviction” and remanded to the district court to conduct a hearing into the conflict. Gilliam, 975 F.2d at 1053-54. The court cited the Sullivan two-part test as - the applicable standard. See Gilliam, 975 F.2d at 1054 n. 5. Rather than reverse; this court remanded, noting that “we do not mean to dictate the result to be reached by the lower court” on remand, and left open the possibility that the district court might deny the claim on the grounds that there was “no suggestion that the alleged conflict impaired the ability of counsel to present the case to the jury.” Gilliam, 975 F.2d at 1054 (emphasis added).
Again, in Beaver v. Thompson, this court noted that “[t]o prevail on a claim of conflict of interest, Beaver must present convincing evidence of an actual conflict and a resulting adverse effect on performance.” 93 F.3d 1186, 1192 (4th Cir.1996). This court applied the two-part Sullivan test without citing Wood in a case where the defense attorney was a part-time prosecutor for the neighboring .county and where the local trial court would presumably have been aware of the potential conflict.
The majority relies on language in United States v. Tatum that “[i]n Wood v. Georgia, the Court flatly stated that a conflict situation which is not addressed by the trial court requires reversal.” Tatum, 943 F.2d at 379 (citations omitted). The Tatum court cited Wood in response to an argument, raised by the government and accepted by the district court, that the court’s review of the conflict of interest issue was premature because facts of record had not yet been developed. Tatum, 943 F.2d at 379. The implication was that the defendant would be free to raise the issue on collateral habeas corpus review rather than on direct appeal. Nonetheless, the Tatum court cited the two-part test of Sullivan as the applicable standard. Tatum, 943 F.2d at 375. The court then applied the test concluding that “known facts lead inevitably to the conclusion that [defendant’s counsel] had unacceptable conflicts of interest” and “that pretrial strategies were adversely affected by those conflicts.” Tatum, 943 F.2d at 380 (emphasis added). The court concludéd that when, as in the ease before it, “the record supports these conclusions, we may confront the issue on direct appeal.” Tatum, 943 F.2d at 380 (citing Wood, 450 U.S. at 272, 101 S.Ct. 1097). Thus, contrary to the majority’s holding, Tatum expressly upheld and applied the Sullivan two-part test and cited Wood for the more narrow jurisdictional point that this court could appropriately address the conflict of interest issue on direct appeal.
II.
Mickens asserts that, when a trial court has failed to inquire into a potential conflict even though it knew or should have known of the conflict, footnote 18 of the Wood decision requires automatic reversal without any inquiry into the existence of either an actual conflict or an adverse effect. See, e.g., United States v. Levy, 25 F.3d 146, 153 (2nd Cir.1994); Unites States v. Burney, 756 F.2d 787, 791 (10th Cir.1985). The relevant footnote states that “Sullivan mandates a reversal when the trial court has failed to make an inquiry even though it ‘knows or reasonably should know that a particular conflict exists.’ ” Wood, 450 U.S. at 273 n. 18, 101 S.Ct. 1097 (internal quote from Sullivan, 446 U.S. at 347, 100 S.Ct. 1708).
*222The majority declines to reach Mickens’ argument about footnote 18 because it concludes he has identified an actual conflict and that the Wood Court’s instruction to the trial court on remand eliminates the requirement of an adverse effect. See Majority opinion at 212 n. 3. On its face, the majority’s rule is distinct from the automatic reversal rule in that it requires a reviewing court to find that an actual conflict of interest existed before it may reverse a conviction. See Majority opinion at 210-11. In practice, however, the rule adopted by the majority, which eliminates the adverse effect requirement when a trial court fails to inquire into a conflict about which it should have known, is functionally equivalent to the automatic reversal rule.
In the absence of an objection from defense counsel, the trial court need not conduct an inquiry unless it “knows or reasonably should know that a particular conflict exists.” Sullivan, 446 U.S. at 347, 100 S.Ct. 1708 (emphasis added). In most cases, the same facts that would lead an appellate court applying Mickens’ automatic reversal rule to conclude that the trial court should have been aware of a “particular conflict” would also lead a court applying the majority’s rule to conclude that an actual conflict existed. Under either rule, the result is reversal of the defendant’s conviction. In this case, for example, the majority finds that the trial court should have been aware of the conflict because Mickens faced charges which might require his defense counsel to counter evidence about the victim and the impact of the murder on the victim’s family. See majority opinion at 212. By the same token, the majority finds an actual conflict because Mickens’ counsel was obligated to investigate and to consider using evidence about the victim to challenge the nature of the crime and the impact of the murder on the victim’s family. See majority opinion at 216-17.
For the same reasons that I reject the majority’s proposed rule, which requires automatic reversal in practice, I must reject the rule asserted by Mickens, which requires automatic reversal on its face. Despite Mickens’ assertions, footnote 18 of Wood can not mean that automatic reversal is required when a trial court fails to inquire into a conflict about which it knew or should have known. That rule would be inconsistent with the manner in which the Wood Court disposed of the case before it. In Wood, the Court ruled that the “possibility of a conflict of interest was sufficiently apparent ... to impose upon the [trial] court a duty to inquire further” and noted that the trial court had failed to conduct such an inquiry. Wood, 450 U.S. at 272, 101 S.Ct. 1097. These holdings satisfy the predicates necessary to trigger the automatic reversal rule asserted by Mickens. Nonetheless, the Wood Court chose to remand the case before it to the trial court for an inquiry into the conflict rather than to reverse the trial court’s decision. Wood, 450 U.S. at 273, 101 S.Ct. 1097.
I should add that the footnote language upon which Mickens relies should be considered as a response by the Wood majority to Justice White’s dissent, not an attempt to create a new rule requiring automatic reversal. Justice White argued that the Court did not have jurisdiction to address the conflict of interest apparent from the record because the conflict had not been raised in the courts below. Wood, 450 U.S. at 280, 101 S.Ct. 1097. In response, the Wood Court’s footnote 18 defended its jurisdiction by indicating that the trial court’s failure to conduct an inquiry constituted error under the rule of Sullivan and noting that Sullivan did not prohibit “the raising of a conflict-of-interest problem that is apparent in the record.” Wood, 450 U.S. at 273 n. 18, 101 S.Ct. 1097 (quoting Sullivan, 446 U.S. at 347, 100 S.Ct. 1708). Subsequent decisions have cited Wood for this procedural point. See Izumi Seimitsu Kogyo Kabushiki Kaisha v. U.S. Philips Corp., 510 U.S. 27, 33, 114 S.Ct. 425, 126 L.Ed.2d 396 (1993); Tatum, 943 F.2d at 380.
*223III.
Because I disagree with the majority’s conclusion that Wood has eliminated the requirement that an actual conflict adversely affect the representation and reject Mickens invitation to adopt the automatic reversal rule, I should address briefly whether Mickens has established an adverse effect in this case.
I note, especially, that the majority decision does not take issue with the finding of the district court that Saunders’ conflict of interest caused no adverse effect in the representation of Mickens in this case. A reference to the reported opinion of the district court is instructive. Commencing at 74 F.Supp.2d, page 606, and continuing to page 613, the district court considered each of Mickens’ claims of adverse effect: Saunders’ failure to raise a consent defense; failure to investigate or raise any negative information about Hall; failure to engage in meaningful plea negotiations; failure to appraise the sentencing court of Hall’s pending charges or strained relationship with his mother; failure to pass along information to his co-counsel; and making a deficient investigation. The district court considered each of these items in great detail, and supported its findings with page and exhibit references from the record. Its findings are largely factual, and the majority does not contend they are clearly erroneous.
The majority correctly states that the existence of a conflict presents a mixed question of law and fact subject to de novo review, Sullivan, 446 U.S. at 342, 100 S.Ct. 1708, and that the district court’s factual findings are subject to the clearly erroneous standard. Fields v. Attorney General, 956 F.2d 1290, 1297 n. 18 (4th Cir.1992). However, the adverse effect inquiry, in my view, is so heavily fact dependent that I believe considerable deference must be given to the findings of the district court.
A defendant has established an adverse effect if he proves that his attorney took action on behalf of one client that was necessarily adverse to the defense of another or failed to take action on behalf of one because it would adversely affect another. See Tatum, 943 F.2d at 376. Thus both taking affirmative actions and failing to take actions “that are clearly suggested by the circumstances” can indicate an. adverse effect. Tatum, 943 F.2d at 376. Based on our holding in Tatum, I.agree with the district court’s thoughtful articulation of a three-part standard for showing an adverse effect which, as in any other civil action, Mickens must establish by a preponderance of the evidence. Mickens, 74 F.Supp.2d at 603-04 (citing Freund v. Butterworth, 165 F.3d 839, 860 (11th Cir.1999) (en banc)). First, the defendant must identify a plausible alternative defense strategy or tactic that his defense counsel might have pursued. Second the defendant must show that the alternative strategy or tactic was objectively reasonable under the facts of the case. Because prejudice is presumed, the petitioner need not show that the tactic or strategy would have altered the outcome of the trial, rather he must only establish that the alternative “possessed sufficient substance to be a viable alternative.” Freund, 165 F.3d at 860. Finally, the defendant must establish that the defense counsel’s failure to pursue that strategy or tactic was linked to the actual conflict.
Mickens challenges this standard, arguing that neither this circuit nor the Supreme Court has required a petitioner to establish a link between an adverse effect and an actual conflict. Such a link is implicit, however, in the Supreme Court’s requirement that a defendant show that “his counsel actively represented conflicting interests” and that “an actual conflict of interest adversely affected his lawyer’s performance.” Sullivan, 446 U.S. at 348, 350, 100 S.Ct. 1708. As this court has noted, the two requirements, “an actual conflict of interest resulting in an adverse effect on counsel’s performance, are often intertwined, making the factual analyses of them overlap.” Tatum, 943 F.2d at 375 (emphasis added).
I conclude that the district court applied the appropriate standard to evaluate the *224adverse effects asserted by Miekens. Because I agree with the district court that many of Miekens’ assertions of adverse effect were not viable defense strategies and that those that were viable defense strategies were not linked to his attorney’s conflict of interest, I agree that there was no adverse effect from Saunders’ conflict of interest.
IV.
To sum up the case, the rule of Cuyler v. Sullivan is that in order to prevail in a case such as this “... a defendant must establish that an actual conflict of interest adversely affected his lawyer’s performance.” 446 U.S. at 350, 100 S.Ct. 1708. That rule was endorsed by the Court explicitly in Strickland v. Washington, 466 U.S. at 692, 104 S.Ct. 2052: “Prejudice is presumed only if the defendant demonstrates that ... ‘an actual conflict of interest adversely affected his lawyer’s performance.’ ” Sullivan was decided in 1980, and Strickland in 1984. The majority holds that the intervening 1981 case of Wood v. Georgia has changed that rule so that an adverse effect is no longer required. I am of opinion that Wood did not change the Sullivan rule and, for that reason, I respectfully dissent.