Sylvia A. LIPSON, Petitioner-Appellant, v. UNITED STATES of America, Respondent-Appellee, 233 F.3d 942

DIANE P. WOOD, Circuit Judge. Sylvia Lipson, along with a number of other people, was charged and convicted of conspiring to distribute crack cocaine in violation of 21 U.S.C. § 846 and of using her home to store and distribute crack cocaine in violation of 21 U.S.C. § 856. She received stiff sentences for those convictions, of 360 months and 240 months respectively, which run concurrently. After both her convictions and sentences were affirmed on direct appeal, see United States v.

Smith, 26 F.3d 739 (7th Cir.1994), she filed the present motion under 28 U.S.C. § 2255. (The case was remanded on other grounds, see 26 F.3d at 760, and later affirmed on a subsequent appeal. United States v. Smith, No. 96-2724, 1997 WL 58804 (7th Cir. Feb.7, 1997).) Her argument for relief has shifted during the course of these proceedings, but it is very generally an ineffective, assistance of counsel claim based on conflict of interest.

We conclude that, under the circumstances, further proceedings are necessary in the district court before it will be possible for that court to decide whether Lip-son’s Sixth Amendment rights were violated because her lawyer was paid by one of her co-defendants. I The underlying basis for Lipson’s claim of ineffective assistance rests on her allegation that her representation was tainted by a conflict of interest because of the relationship between her lawyer and her co-defendant, Greg Collins. Collins was the ringleader of the conspiracy.

Lip-son (whose version of the facts on these points we accept, given the lack of an evidentiary hearing) was an indigent single parent. She was romantically involved with, and habitually battered by, Collins. Eventually, Lipson, Collins, and eight other individuals were charged in the multicount indictment that produced the convictions Lipson is now trying to challenge. A careful reading of the original pro se § 2255 motion that Lipson filed reveals that she raised two complaints about her representation at trial that related to conflicts of interest.

Specifically, her motion made the following assertions: This request lie. for an order appointing counsel and setting an evidentiary hearing on the § 2255 motion] to the court is based on the ineffective assistance of counsel, namely, Ricky Black. Black should not have been allowed to represent Collins and Lipson, due to the conflict of interest and others violations of Constitutional and Due Process Rights. Lipson simply did not have the “baggage” that Collins and the other conspirators [sic] were found guilty or convicted.

Black was paid his attorneys fees by Collins who was interested in controlling the defense in his favor in this case. Other portions of the same memorandum make it clear that she was asserting that Collins had selected Black to be her lawyer and that he was paying Black, so that he could control Lipson’s representation. Her motion also raised numerous other grounds for relief, including recanted testimony of a witness, the government’s failure to prove that the substance was really crack, a claim that her sentence should be reduced because she had been abused, and so on.

We focus here on the ineffective assistance claim, because this is the only one on which a certificate of appealability has been granted. This court granted that certifícate after the trial court denied Lip-son’s request, and it identified the following issue for appeal: ‘Whether trial counsel was ineffective due to a conflict in representing both petitioner and her co-defendant, Greg Collins.” We also appointed appellate counsel for Lipson.

Unfortunately, the precise theory of ineffective assistance identified in our certificate of appealability does not appear to have been the theory — or at least the only theory — before the district court, as Lip-son’s original motion demonstrates. Furthermore, it is clear from a reading of the district court’s opinion that the judge understood what Lipson was trying to say.

After rejecting a number of specific instances of ineffective counsel that Lipson had alleged, that related to failures to investigate her claim or to communicate with her, the court wrote the following: Petitioner further alleges that counsel had a conflict of interest. Specifically, petitioner argues that counsel neglected her interests to benefit a co-defendant, Gregory Collins, her boy friend and co-defendant who paid for attorney’s fees. Lipson v. United States, No. 97-CV-891-WDS, Mem. & Order at 6 (S.D.Ill. August 26, 1998). After reciting the standard established by Strickland v.

Washington, 466 U.S. 668, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984), and noting that Lipson was relying on Strickland rather than Cuyler v. Sullivan, 446 U.S. 335, 100 S.Ct. 1708, 64 L.Ed.2d 383 (1980) (which would have permitted her to demonstrate an actual conflict of interest which, if shown, would have lightened the burden of demonstrating ineffective assistance), the court rejected her argument with the following explanation: ...

To demonstrate prejudice [under Strickland], petitioner must demonstrate that counsel’s errors were so serious as to deprive her of a fair trial, that is, of a trial whose result is reliable.... At trial, the government provided extensive evidence demonstrating petitioner’s deep involvement to amply support her conviction.... Her counsel certainly participated in the trial on her behalf, not on behalf of any co-defendant. This case involved a joint defense of a lack of conspiracy. That commonality does not rise to the level of conflict.

Similarly, the mere fact that a co-defendant footed the bill for her defense counsel does not, in and of itself, create a conflict of interest. This is not a situation like that presented in Stoia [v. United States, 109 F.3d 892 (7th Cir. 1997)] or similar conflict cases. The Court finds that petitioner’s claims of conflict are wholly unsupported by the record. Lipson, Mem. & Order at 7.

It is therefore clear that, until the certificate of appealability issued from this court, everyone knew that this case was about the type of conflict that can arise when one co-defendant retains, pays, and controls the lawyer representing another co-defendant. Unfortunately, matters became confused in part because of our order and in part because counsel for Lipson must not have investigated the record very carefully. Lipson’s opening brief in this court relies solely on the proposition that the issue was joint representation of her and Collins.

The government correctly responded in its brief that the record clearly showed that no such thing occurred. Lip-son was represented by attorney Richard R. Black, of East St. Louis, Illinois, while Collins was represented by attorney Robert H. Rice of Belleville, Illinois at the pretrial, trial, and appellate stages, and attorney Burton H. Shostak of St. Louis, Missouri, after remand and on the new appeal. Lipson’s reply brief concedes the error, in effect, and argues that her right to effective counsel was nonetheless infringed by the financial arrangements between Collins and Black.

There is plenty of blame to spread around for this confused state of affairs, but none of it in our view properly goes to Lipson. In fact, the government fully briefed the question whether Black labored under a conflict of interest because of the fact that Collins paid for his services to Lipson, and Lipson’s appointed counsel also briefed that issue (albeit in the reply brief). In these somewhat unusual circumstances, we think it best to expand the certificate of appealability to include the actual conflict theory Lipson originally alleged, the district court decided, and the parties addressed.

We hereby do so, and now turn to the merits of this claim. II As we recently explained in Cabello v. United States, 188 F.3d 871 (7th Cir. 1999), defendants who wish to raise claims of ineffective assistance of counsel based on conflicts of interest may proceed under either of two theories. If the defendant can establish that the trial judge knew or should have known that a potential conflict of interest existed, then we presume that the defendant was prejudiced by that conflict if the judge made no inquiry into it. Id. at 875.

If, on the other hand, the trial judge was not put on notice of a potential conflict, we will find prejudice “only if the defendant demonstrates that her counsel actively represented conflicting interests and that the conflict adversely affected the counsel’s performance.” Id., citing Cuyler, 446 U.S. 335, 350, 100 S.Ct. 1708, 64 L.Ed.2d 333 (1980). Both parties claim that Cabello entitles them to prevail, and so we take a moment to review the facts and holding of that case.

In Cabello, the petitioner (Maria Cabello) argued that her trial counsel provided ineffective assistance because he had been retained and paid by her boyfriend, the alleged ringleader of the drug conspiracy in which she was accused of participating. Like Lipson, Cabello also argued that her indigence should have put the district court on notice of the likelihood that the boyfriend was paying for the lawyer, when the lawyer showed up ready to participate in the proceedings without any request for granting Cabello IFP status.

Importantly, the court considering Cabello’s § 2255 motion held an evidentiary hearing on her allegations ,at which it heard testimony from both Cabello and the retained lawyer. On this record, we concluded in Cabello that there was nothing that should have alerted the trial court to the potential for a conflict, and that nothing in the record indicated that Cabello had been actually prejudiced by her lawyer’s performance. We see no distinction between Ca-bello’s case and Lipson’s with respect to the first of these points.

As far as the trial court knew, Lipson and Collins were represented by different lawyers, and the judge specifically found in his § 2255 memorandum and order that Lipson’s lawyer gave her adequate representation. Lipson suggests that the court should have seen red flags in the facts that she and Collins were being tried together, that they had a romantic relationship, and that Collins had abused her in the past — all of which were known at the time of trial.

But we see no reason why the existence of the romantic relationship should have triggered the thought that Collins was paying for Lipson’s lawyer, much less why a history of abuse should have supported such an inference. We must turn, therefore, to the second question identified in Cabello— whether Lipson can show an actual conflict of interest and an adverse effect on her representation, and, in her case, whether the district court abused its discretion in failing to hold a hearing to explore these points.

Lipson alleged in her motion that Black allowed Collins to dictate her legal defense because Collins was paying his fees. If true — and because there was no hearing we must assume that it is- — ’this would be enough to demonstrate an actual conflict of interest. Stoia v. United, States, 22 F.3d 766, 771 (7th Cir.1994) (“An actual conflict of interest results if the defense attorney was required to make a choice advancing his own interests to the detriment of his client’s interests.” (Quotations and citation omitted.)). See also Ciak v. United States, 59 F.3d 296, 307 (2d Cir. 1995).

But demonstrating the actual conflict is not enough by itself. Lipson must also show that Black’s conflict adversely affected his performance. She must suggest something that he did that he should not have done, or something that he failed to do that should have been done, that had a negative impact on her case. To the extent she implies that Black should have raised the point that she was effectively under Collins’s control and thus not responsible for her actions in furtherance of the conspiracy, our examination of the record shows that he did bring out this evidence.

At trial, Black argued that Lipson was an unwilling participant in the conspiracy — the precise theory of innocence that she is now asserting in her § 2255 motion. She was entitled to put this argument before the jury, and in fact, the jury evidently took it quite seriously.

At one point during its deliberations, the jury sent a note to the trial judge explaining that they felt that Lipson had participated in the conspiracy “only because of love, fear, and concern for her children,” and “in much the same way one would be [willing] with a knife to their throat.” The jury asked in its note whether this concern was sufficient reason to find a reasonable doubt, but the judge responded without objection from Black that it was to follow the instructions that had already been given. See Smith, 26 F.3d at 748.

The record reveals that those instructions did not include a possible duress defense. Instead, counsel tendered an instruction defining the term “willful” for the jury, which would have told them that willfully means voluntarily and intentionally acting with bad purpose either to disobey or disregard the law. The judge rejected that instruction, but that was not counsel’s fault. Furthermore, we note that duress is an extremely difficult defense on which to succeed.

Counsel’s decision to try to use Collins’s overbearing influence on Lipson as a way of negating Lipson’s criminal intent was a respectable strategy, and one that might have succeeded. See United States v. Madoch, 149 F.3d 596, 598-600 (7th Cir.1998). See also United States v. Fiore, 178 F.3d 917, 922 (7th Cir.1999); United States v. Otis, 127 F.3d 829, 835 (9th Cir.1997). Her lawyers on appeal now suggest that another adverse effect of Black’s relationship with Collins was Black’s failure to pursue a plea agreement on Lipson’s behalf.

If true, and if the evidence showed that she had told him at the time she might be interested in cooperating, that might qualify as an adverse effect resulting from the conflict of interest. Compare Paters v. United States, 159 F.3d 1043, 1047-48 (7th Cir.1998); Johnson v. Duck-worth, 793 F.2d 898, 902 (7th Cir.1986). Had there been such a plea agreement, Lipson might have agreed to testify against Collins in exchange for a reduced sentence; Collins certainly had every incentive to prevent such an outcome, and to instruct Black not to pursue that strategy.

We think it significant that this is exactly what happened with about half of the defendants, who testified against those who remained and who received lighter sentences. The presence of these other turncoats meant that joint denial was a weak strategy at best for Collins and Lipson. The possibility that Lipson was trying to game the system — first cooperating with her co-defendant Collins in the denial strategy, and then claiming that she might have pleaded guilty and helped the government — -seems unlikely enough on the present facts that it should not defeat her right to a hearing as a matter of law.

The fact that the government accepted guilty pleas from some defendants indicates that it might have been open to negotiating with Lipson as well.

Lipson herself alluded to the lost chance at a guilty plea in her § 2255 motion, where she wrote “because of Black’s loyalty to Collins was inspired to prevent Lip-son from receiving any representation at all from Black, because Black and Collins knew or should have known if Lipson, plead guilty, she would have to testify against, Collins, as Black’s paymaster, rather than Lipson who was also his client.” This statement appears to be her only effort to bring this point to the district court’s attention.

Nevertheless, she was at that time proceeding pro se, and her statement appeared in the § 2255 motion itself; it was not buried in the middle of a large stack of discovery materials or otherwise difficult to find. Ill We cannot tell without an eviden-tiary hearing whether Black really did fail to pursue a plea agreement on Lipson’s behalf, whether Lipson had any interest in pleading guilty at the time, or whether the government would have offered an agreement to her in any event.

The judgment of the district court is therefore ReveRsed and the case is Remanded for further proceedings consistent with this opinion. . Our dissenting colleague is concerned about what a trial judge might do to prevent the kind of complaint that Lipson is now raising. But that point assumes that we are criticizing the trial judge for failing to do something at the original trial, which is not the case. As we explain in the text, this is not a case in which the trial judge either knew or should have known about the potential conflict of interest.

Because it is not, the defendant must satisfy the strict requirements of Cuyler, 446 U.S. at 335, 350, 100 S.Ct. 1708. This kind of claim, just like many other ineffective assistance of counsel claims that a trial judge would similarly be unable to anticipate, sometimes requires an exploration of the facts at the § 2255 stage. We hold here only that Lipson has alleged enough to entitle her to such an exploration.