We hold that when a criminal defendant is represented by two different attorneys from the same public defender’s office at trial and on direct appeal, res judicata bars a claim of ineffective assistance of trial counsel raised for the first time in a petition for postconviction relief when such claim could have been made on direct appeal without resort to evidence beyond the record, unless the defendant proves that an actual conflict of interest enjoined appellate counsel from raising a claim of ineffective assistance of trial counsel on direct appeal.
In State v. Perry (1967), 10 Ohio St.2d 175, 39 O.O.2d 189, 226 N.E.2d 104, this court found res judicata to be a proper basis upon which to dismiss without hearing a petition for postconviction relief. This court held that res judicata bars a convicted defendant from raising in a postconviction relief petition any defense that was raised or could have been raised by the defendant at trial or on direct appeal. Id. at paragraph nine of the syllabus.
After Perry, this court and lower courts recognized exceptions to the absolute application of res judicata in postconviction relief proceedings where a claim of ineffective assistance of counsel is raised. This court’s decision in Cole, supra, addressed and distilled those post-Perry decisions and forms the applicable law in the area. Cole approvingly attributes to State v. Carter (1973), 36 Ohio Misc. 170, 65 O.O.2d 276, 304 N.E.2d 415, the notion that since “counsel cannot realistically be expected to argue his own incompetence, res judicata does not act to bar a defendant represented by the same counsel at trial and upon direct appeal from *530raising a claim of ineffective assistance of counsel in a petition for postconviction relief.” Cole, 2 Ohio St.3d at 114, 2 OBR at 663, 443 N.E.2d at 171, fn. 1.
That statement is implicitly included in the Cole syllabus, which states:
“Where defendant, represented by new counsel upon direct appeal, fails to raise therein the issue of competent trial counsel and said issue could fairly have been determined without resort to evidence [beyond] the record, res judicata is a proper basis for dismissing defendant’s petition for postconviction relief.” (Emphasis added.) Cole at syllabus.
Cole recognizes that res judicata does not apply when trial and appellate counsel are the same, due to the lawyer’s inherent conflict of interest. The applicability of that conflict of interest to co-workers is questionable — Cole recognizes a conflict of interest that is highly personal, and thus it may be logically argued that the exception to res judicata applies only when trial and appellate counsel are the same person.
However, Ohio’s Code of Professional Responsibility states that a lawyer’s conflict of interest is imputed to his law firm. DR 5-105(D). Therefore, an important question is whether a private law firm and a county public defender’s office are analogous in regard to the question at hand.
At the threshold level, the lack of a financial stake in the case’s outcome sets the public defender apart from the private firm. A lawyer in private practice who is still being paid by a defendant would be less willing to admit that his firm’s representation in an earlier stage of the proceedings was substandard. Also, unlike the public defender, the private attorney is in competition with other law firms for clients’ business, so diminished reputation more directly affects the finances of private sector attorneys.
While a public defender’s office may not have the financial conflicts of a private law firm, conflicts driven by loyalty, reputation and esprit de corps may be just as likely to arise in a public defender’s office as in a private law firm.
Still, the doubts or awkwardness such feelings engender does not give rise to the same level of conflict of interest that would occur were a lawyer representing competing parties or co-defendants with differing interests, and certainly does not create a level of conflict serious enough for this court to find a per se conflict of interest. A lawyer’s supreme duty of loyalty is to his client (EC 5-1), and that is a duty that we should not assume will be ignored due to the possibility of embarrassing a co-worker.
To find that counsel from the same public defender’s office is per se not “new counsel” for purposes of asserting a claim of ineffective assistance of counsel at the postconviction stage would be to find that there is a per se conflict at the appellate level as well. We are not prepared to say, as amicus suggests, that in *531every case where ineffectiveness of trial counsel may be raised there is a conflict of interest disqualifying the public defender’s office from representation, and necessitating the appointment of private counsel.
Upon considerations of petitions for postconviction relief, the trial court should conduct a case-by-case analysis to determine whether an actual conflict of interest existed which kept appellate counsel from effectively arguing ineffectiveness of trial counsel. If circumstances peculiar to the case indicate the presence at the appellate level of an actual conflict of interest, then res judicata does not bar a claim for postconviction relief based upon ineffectiveness of trial counsel.
Of course, it should not be necessary to instruct public defenders that if an actual conflict does exist at the appellate level that they should either procure a waiver of the conflict from the defendant or cease representation in the case. Any lawyer who does not do so is violating his ethical obligation to avoid conflicting representation, and also creates a viable claim for ineffective assistance of appellate counsel. Murnahan, supra.
We accordingly reverse the judgment of the court of appeals and remand the cause to the trial court to determine whether an actual conflict of interest enjoined appellate counsel from asserting the ineffectiveness of Lentz’s trial counsel.
Judgment reversed and cause remanded.
Moyer, C.J., Douglas and Resnick, JJ., concur. F.E. Sweeney, J., concurs in the syllabus and judgment only. A.W. Sweeney and Wright, JJ., dissent.