United States Court of Appeals FOR THE DISTRICT OF COLUMBIA CIRCUIT Argued February 7, 2000 Decided June 2, 2000 No. 99-3103 United States of America, Appellee v. Susan Viola Klat, Appellant Appeal from the United States District Court for the District of Columbia (No. 96cr00385-01) Mary M. Petras, appointed by the court, argued the cause for the appellant. Thomas S. Rees, Assistant United States Attorney, argued the cause for the appellee. Wilma A. Lewis, United States Attorney, and John R. Fisher and Thomas J. Tourish, Jr., Assistant United States Attorneys, were on brief for the appellee. Before: Silberman, Henderson and Randolph, Circuit Judges. Opinion for the court filed by Circuit Judge Henderson. Karen LeCraft Henderson, Circuit Judge: Susan Viola Klat was convicted of threatening to assault the Chief Justice of the United States and the Clerk of the United States Supreme Court in violation of 18 U.S.C. ss 115, 1114. On appeal after the conviction, we remanded for the district court to determine whether there was a reasonable possibility that appointment of counsel to represent Klat would have changed the outcome of the pre-trial competency hearing in which Klat, upon her insistence, appeared pro se. See United States v. Klat, 156 F.3d 1258, 1267 (D.C. Cir. 1998). On remand the district court found "no reasonable possibility that counsel could have affected the outcome of defendant's competency hearing." United States v. Klat, 59 F. Supp. 2d 47, 55 (D.D.C. 1999). Klat appeals, seeking reversal of the district court's order on remand and a new trial. She argues counsel could have affected the outcome of the competency hearing by (1) sharing his observations regarding her under- standing of the proceedings and her ability to assist counsel, (2) advising her to participate more fully in the psychological evaluation, (3) challenging the findings of the forensic psy- chologist who examined her, (4) retaining an independent expert to examine Klat and (5) advising her not to waive her right to counsel. Her arguments, whether standing alone or in combination, do not persuade us to overturn the district judge's determination that counsel could not have affected the outcome of the competency hearing. Accordingly, we affirm the district court. I. Based on conduct law enforcement officers viewed as threats directed toward the Chief Justice and toward person- nel of the United States Supreme Court, Klat was arrested and presented to a magistrate for a preliminary hearing, where she was represented by counsel from the Federal Public Defender's office. The magistrate judge found proba- ble cause and ordered Klat to undergo a competency exami- nation. On September 3, 1996 Bruce Cambosos, M.D., a psychiatrist, concluded Klat was competent to stand trial and she was released on her own recognizance. She filed a motion three weeks later seeking removal of counsel so that she could represent herself. On October 25, 1996 Klat was indicted on two counts for threatening to assault, in violation 18 U.S.C. ss 115 and 1114, an officer and employee of the United States and a United States judge. Klat first appeared before the district court at a November 1, 1996 arraignment hearing. The record of the hearing reflects Klat's determination to represent herself. See Klat, 59 F. Supp. 2d at 47-48. During the hearing, counsel moved to withdraw from the case, citing a lawsuit Klat had filed seeking to hold the United States liable for his representation of her. The court granted the motion to withdraw but found that Klat's "bizarre behavior" provided reasonable cause to order a competency evaluation. Joint Appendix (JA), tab F, at 2. Klat was then transferred to Carswell Medical Center in Fort Worth, Texas, where James A. Shadduck, a forensic psychologist, performed a competency evaluation. Klat was interviewed and otherwise examined by Shadduck and others who conducted "psychiatric consultation," "social work consul- tation" and "behavioral observation." JA, tab E, at 3. They also interviewed Klat's former counsel and her former co- workers. Shadduck interviewed Klat five times for a total of approximately ten hours. Klat, however, refused to take both an IQ test and the Multi-Phasic Personality Inventory (MMPI). Shadduck submitted his report to the court on December 16, 1996. Citing Klat's in-depth knowledge of the legal process, of the charges against her and of potential outcomes, as well as her at least average intellectual capacity, Shadduck concluded that she was competent to stand trial. He diagnosed her as possessing a narcissistic personality disorder and potential bipolar disorder. One month after Shadduck submitted his report, the dis- trict court conducted a competency hearing at which Klat appeared pro se. No witnesses were called. The govern- ment did not challenge Shadduck's evaluation and Klat, while taking issue with certain aspects of the diagnosis, maintained she was competent. She did agree to the government's proposal that she accept appointment of stand-by counsel. The court specifically cited Klat's demeanor during the hear- ing as indicative of her competence.1 The court accepted Shadduck's opinion and found Klat competent to stand trial and, in particular, that she had demonstrated both an under- standing of the charges and an ability to conduct her defense. See JA, tab F, at 4. A jury trial commenced in February 1997 and, after open- ing statements and the government's first two witnesses, Klat requested that stand-by counsel conduct the rest of the trial. She was convicted on February 26, 1997 and was held in jail pending sentencing. Thomas Goldman, M.D., a psychiatrist, examined Klat on March 26, 1997 at her counsel's request. Although he determined that she was not competent to continue representing herself, he did not state that she was not competent to be sentenced. On May 27, 1997 Klat was sentenced to a term of 57 months' imprisonment on each count, to run concurrently. Klat appealed and we rejected all of her challenges but one. See United States v. Klat, 156 F.3d 1258 (D.C. Cir. 1998). We held that the district court erred in allowing Klat to appear without representation at the hearing to determine her competence to stand trial while at the same time it expressed misgivings about her competence: In the instant case, appellant had clearly indicated her desire to waive her right to counsel and to proceed pro se. However, at the November 1, 1996 arraignment hearing the district court made an explicit finding that there was "reasonable cause" to believe that appellant was mentally incompetent to stand trial. Under these circumstances, we find that the district court erred in __________ 1 The district court found Klat's demeanor "controlled and appro- priate for the situation." JA, tab F, at 4. allowing appellant's appointed counsel to withdraw with- out appointing new counsel to represent appellant until the issue of her competency to stand trial had been resolved. This finding is based on our conclusion that, where a defendant's competence to stand trial is reason- ably in question, a court may not allow that defendant to waive her right to counsel and proceed pro se until the issue of competency has been resolved. 156 F.3d at 1262-63 (footnotes omitted). We held that the district court's error constituted a violation of Klat's sixth amendment right to counsel. See id. at 1263. Accordingly, we remanded "for an evidentiary hearing to determine wheth- er counsel could have made a difference in the outcome of appellant's competency hearing." Id. at 1267. We instructed the district court to vacate the conviction and sentence only if it determined counsel could have affected the outcome but otherwise affirmed both her conviction and sentence. See id. The district court held an evidentiary hearing on April 14, 1999 during which the government presented only one wit- ness, James Shadduck, and Klat through counsel presented four: Shelly Stanton, M.D., a psychiatrist; Thomas Goldman, the psychiatrist who had examined her before sentencing; Richard Schmitt, a psychologist; and Klat's former counsel who withdrew at arraignment. On cross-examination Shad- duck testified that it was possible, but not probable, that the MMPI test (which Klat refused) could have affected his opinion to some degree and that it was possible, but not probable, that Klat was suffering from a delusional disorder when he evaluated her. All of Klat's expert witnesses examined her at some point after the competency hearing; Goldman's exam occurred approximately two months later and the rest were even later. Goldman concluded, on the basis of a fifty-minute interview and his review of Klat's trial counsel's case file, that Klat was not competent to proceed pro se. He testified that if he had been asked in March 1997 about her competence to stand trial one month earlier, he would have deemed her incompe- tent. Stanton examined Klat later, in August 1997, after she had begun a hunger strike, and found she had a delusional disorder. Stanton testified that, if Klat was in the same condition at the time of the competency hearing in January 1997 as she was when he saw her several months later in August 1997, he would opine she was not competent at the earlier time. Schmitt, who examined Klat in September 1997, testified that she suffered not only from a narcissistic disor- der but also from a delusional disorder. Based in large part on his conclusion that Klat misapprehended the facts of her case and exhibited "a paranoid distrust of her attorneys," Transcript (Tr.) 4/14/97 at 157, Schmitt opined that Klat was not competent when he examined her. Shadduck, Goldman and Stanton agreed, however, that Klat's conviction could have caused "decompensation," that is, a worsening of her condition. See Tr. 4/14/97 at 18-19 (Shadduck), 69 (Stanton) and 110 (Goldman). Klat's former counsel testified that Klat's conduct after the August 30, 1996 preliminary hearing and Cambosos's initial competency determination on September 3 led him to believe she was incompetent. He cited Klat's tape-recording a meet- ing with him and her irrational views on how to proceed in the case: for example, refusing the negotiated pre-trial diver- sion offer and seeking indictment so that she could go to trial and expose the government conspiracy against her. He gave as reasons for his failure to voice his concerns to the magis- trate judge his then-recent appointment and insufficient infor- mation, and for his subsequent failure to express his misgiv- ings to the district court the legal action Klat had by then instituted and his reliance on the government to challenge her competence. In the end, the district court rejected Klat's arguments, the same five arguments she makes here, and concluded there was no reasonable possibility that counsel could have changed the result of the competency hearing. The district court emphasized that Klat "repeatedly demonstrated her total unwillingness to follow advice from others, especially counsel" and that "three separate mental health experts evaluated Klat at three separate times,"2 none of whom found her incompe- tent. 59 F. Supp. 2d at 51. The district court discounted counsel's doubts as to Klat's competence as cumulative in light of the court's observations.3 See id. at 53. The court also noted that counsel had represented Klat only at the preliminary hearing stage. Rejecting Klat's argument that counsel could have advised her to participate in further psychological testing, the court stated that, given her resistance to counsel's advice and efforts, "there is absolutely no evidence to give rise to the inference of a reasonable possibility that defendant would have acted in accordance with such advice." 59 F. Supp. 2d at 53. The court dismissed the argument that counsel's cross-examination of Shadduck, assuming testimony had been taken, would have persuaded it that Klat was not competent. The court again relied on the opinions of the two experts (Cambosos and Shadduck) whose evaluations were before it when it originally concluded she was competent. See id. at 54. Responding to Klat's fourth argument that counsel could have secured another expert opinion, the court noted that she proceeded in forma pauperis and found no reasonable possi- bility that the court would have approved the hiring of an independent expert to conduct what would have been the third evaluation in three and one-half months. See id. Final- ly, the district court concluded that counsel's advice not to proceed pro se would not have affected the outcome in light of Klat's earlier rejection of the advice of both counsel and the court that her pro se stance was imprudent. See id. at 54-55. __________ 2 Although the district court's count included Goldman who evalu- ated Klat after conviction, in its analysis the court relied on the opinions of Cambosos and Shadduck who had evaluated Klat before the court found Klat competent to stand trial. See 59 F. Supp. 2d at 54. 3 The district court concluded that counsel's description of Klat's obstinacy in ignoring his efforts to communicate and her inflated opinion of her defense could not have affected the outcome because it had itself noted the same behavior. See 59 F. Supp. 2d at 53. II. The applicable standard of review is, not surprisingly, in dispute. Klat argues that the determination is a mixed question of law and fact which should be reviewed de novo. She likens the issue before us to our review of the "reason- able probability of a different outcome" determination in an ineffective assistance of counsel claim.4 The government, on the other hand, contends that the determination amounts to a finding of fact warranting review only for clear error. It compares the issue before us to a competency determination which we review for clear error. See United States v. Cald- well, 543 F.2d 1333, 1349 (D.C. Cir. 1974).5 In our review of the district court's conclusion that there was no reasonable possibility counsel could have effected a different outcome, the factual basis of the competency deter- mination necessarily comes into play. Nevertheless our task is to decide the legal question whether a constitutional error may have so affected the proceeding below that the reason- able possibility standard is met. We must therefore apply a legal standard to a particular set of facts. Accordingly, we are reviewing a mixed question of law and fact. See Barbour v. Browner, 181 F.3d 1342, 1345 (D.C. Cir. 1999). In Barbour we discussed the factors involved in determin- ing whether a mixed question of law and fact deserves deferential or independent review: __________ 4 Our standard of review for ineffective assistance of counsel claims is itself subject to debate. See United States v. Askew, 88 F.3d 1065, 1070 (D.C. Cir. 1996) (unnecessary to pass on unsettled issue of standard of review). 5 "Competence to stand trial requires 'sufficient present ability to consult with his lawyer with a reasonable degree of rational under- standing and ... a rational as well as factual understanding of the proceedings against him.' " Caldwell, 543 F.2d at 1348 (quoting Dusky v. United States, 362 U.S. 402 (1960) and citing Pouncey v. United States, 349 F.2d 699, 701 (D.C. Cir. 1965)). Competence to stand trial vel non is decided by a preponderance of the evidence. See 18 U.S.C. s 4241(d). As to so-called "mixed questions of law and fact," ... there is no obvious way to decide whether determinations made at the trial level should be reviewed deferentially or independently. Therefore, the reviewing court must make a reasoned judgment whether the risk of an erro- neous trial level decision, or the need to clarify the governing law, or any other value secured by review de novo, is warranted in view of the added costs of such review. 181 F.3d at 1345 (citations omitted). We cited the discussion in Miller v. Fenton, 474 U.S. 104, 114 (1985), to the effect that determining the applicable standard of review often turns upon which judicial actor "is better positioned than another to decide the issue in question." Barbour, 181 F.3d at 1345. We cautioned, however, that the goal of clarifying governing law may be elusive if the case is "intensely fact specific." Id. at 1348. Here, the district judge who presided over the competency hearing is better positioned than we to decide if counsel could have affected his factual determination of Klat's competence. See Caldwell, 543 F.2d at 1349 (competency determination is question of fact). We conclude that our review of the mixed question of law and fact before us should be more deferential than independent and, therefore, gov- erned by the clearly erroneous standard. The reasonable possibility inquiry governing the district court's determination is similar to that made in determining whether a trial error of constitutional dimension is harmless: that is, whether the error is harmless beyond a reasonable doubt. See Coleman v. Burnett, 477 F.2d 1187, 1211 n.158 (D.C. Cir. 1973) (noting that "[i]n Chapman it was held that before a federal constitutional error can be held 'harmless,' the court must be of the belief that it was harmless beyond a reasonable doubt [and] must find that there is no reasonable possibility that the error complained of might have contribut- ed to the conviction") (citing Chapman v. California, 386 U.S. 18, 24 (1967), overruled in part by Brecht v. Abrahamson, 507 U.S. 619 (1993), and Fahy v. Connecticut, 375 U.S. 85, 86-87 (1963)) (emphasis added);6 United States v. Saro, 24 F.3d 283, 287 (D.C. Cir. 1994) ("For most constitutional errors, an appellate court is to reverse if it entertains a 'reasonable doubt' about whether the error affected the outcome below."); see also Pyles v. Johnson, 136 F.3d 986, 994 (5th Cir. 1998) (equating reasonable possibility determination regarding ex- trinsic evidence in jury room to harmless error review). Of the five reasons Klat advances for asserting that counsel could have affected the outcome, the district court's rejection of the second and fifth reasons (counsel could have advised her (2) to participate more fully in the psychological evalua- tion and (5) not to waive her right to counsel) is plainly reasonable. We, therefore, consider further only the others. Klat first argues that counsel could have offered his own observations regarding her understanding of the proceedings and her ability to assist him. Because a large part of the competency determination turns upon a defendant's ability to assist counsel, see Caldwell, 543 F.2d at 1348 (competence requires ability to consult with counsel), representation at a competency hearing is important and we have so recognized: This court recognizes that in making a competency deter- mination it may be very useful for the trial judge to question both the defendant and his counsel; the applica- ble criteria measure one's ability to consult with his lawyer and to understand the course of legal proceed- ings. Thus counsel's first-hand evaluation of a defen- dant's ability to consult on his case and to understand the __________ 6 In Chapman, 386 U.S. at 24, the Court stated: There is little, if any, difference between our statement in Fahy v. State of Connecticut about 'whether there is a reasonable possibility that the evidence complained of might have contrib- uted to the conviction' and requiring the beneficiary of a constitutional error to prove beyond a reasonable doubt that the error complained of did not contribute to the verdict obtained. We, therefore, do no more than adhere to the meaning of our Fahy case when we hold, as we now do, that before a federal constitutional error can be held harmless, the court must be able to declare a belief that it was harmless beyond a reasonable doubt. charges and proceedings against him may be as valuable as an expert psychiatric opinion on his competency. This is particularly so when--as in the instant case--trial counsel has independently expressed 'misgivings' about the defendant's competency. United States v. David, 511 F.2d 355, 360 (D.C. Cir. 1975) (footnotes omitted); see also Drope v. Missouri, 420 U.S. 162, 177 n.13 (1975) ("Although we do not, of course, suggest that courts must accept without question a lawyer's representa- tions concerning the competence of his client, an expressed doubt in that regard by one with the closest contact with the defendant, is unquestionably a factor which should be consid- ered.") (citations and internal quotation marks omitted). In rejecting Klat's argument, the district court relied in part on counsel's failure to raise the competency issue at the preliminary hearing. Its reliance is misplaced for several reasons. First, counsel would have had more exposure to Klat had he continued to represent her at the competency hearing. Furthermore, given the sole purpose of the compe- tency hearing (and the distinct purpose of the preliminary hearing), it is more likely that counsel would have voiced his concerns there, particularly if the district court had followed the David holding's lead and sought his input. We may assume the ground for granting counsel's motion to withdraw was valid when the motion was granted at arraignment and that therefore new counsel would have been appointed to represent Klat at the competency hearing. It is also likely that newly appointed counsel would have raised the compe- tency issue. First, he had an ethical duty to do so, assuming he had doubts as to her competence, see, e.g., United States v. Boigegrain, 155 F.3d 1181, 1188 (10th Cir. 1998), and, second, Klat's steadfast refusal to follow advice of counsel, as empha- sized by the court below, suggests that counsel could have thought her unable to consult with him rationally and, there- fore, unable to have a rational understanding of the proceed- ings. See Caldwell, 543 F.2d at 1348 (defendant must have rational, not merely factual, understanding of proceedings). Decisive to our rejection of Klat's argument, however, is that nearly everything counsel could have offered was then apparent to the district court.7 See 59 F. Supp. 2d at 53. It recognized that Klat was untrusting, see id. at 51, that she "had a distorted notion of the merits of her case," id. at 53, and that she refused to communicate meaningfully with her lawyer, see id. at 51, 53-54. In fact, the district court's observations to this effect are what triggered it to order the competency hearing. See JA, tab F, at 2 ("Because of defendant's bizarre behavior in open court, the court ordered that she be committed to the custody of the Attorney General for ... psychiatric examination."); see also 59 F. Supp. 2d at 53. It is highly unlikely on this record that Klat would have been receptive to any counsel. The district court thus justifi- ably deemed any observations of counsel "cumulative." See 59 F. Supp. 2d at 53. At the hearing, it was reassured by two expert opinions that Klat was not incompetent. Accordingly, the district court did not clearly err in finding no reasonable possibility that counsel's observations could have affected the outcome. Next, Klat argues counsel could have challenged Shad- duck's findings. The district court dismissed this argument because no live testimony was taken at the competency hearing and because Klat challenged only Shadduck's conclu- sion, arguing that the facts before him should have led him to the opposite conclusion. See 59 F. Supp. 2d at 54. But counsel could have subpoenaed Shadduck and then questioned his opinion. See Caldwell, 543 F.2d at 1348 (D.C. Cir. 1975) ("While the proceeding need not be lengthy or involved, as a minimum we think the inquiry must be of record and both parties must be given the opportunity to examine all wit- nesses who testify.") (internal quotation marks omitted); see also United States v. Williams, 113 F.3d 1155, 1160 (10th Cir. 1997) (once doubt is raised as to his competence "protections of an adversary proceeding must be afforded the defendant"). __________ 7 As the court pointed out on remand, "Dr. Shadduck, like Dr. Cambosis [sic] before him, also observed this type of behavior and found defendant to be competent." See 59 F. Supp. 2d at 53. The only point on which Shadduck's opinion appears assaila- ble, however, is his inability to administer, and assess, the personality test (MMPI) Klat refused to take. In any event, Shadduck testified at the hearing on remand that, while the lack of the MMPI result limited the accuracy of his diagnosis, it did not affect the accuracy of his competency determina- tion. Perhaps counsel's opportunity to expose this weakness in Shadduck's diagnosis would have enabled counsel to persuade the court to appoint another expert to evaluate Klat, which leads to her last argument. Klat claims counsel could have retained an independent expert to evaluate her. As the district court recognized, however, Klat's argument requires several inferences to be stacked one upon the other. See 59 F. Supp. 2d at 54. The weakest may be the foundational supposition that the district judge would have second-guessed Shadduck's opinion (the second competency determination in three and one-half months), whether based on cross- examination or on counsel's doubts which mirrored his own, and then authorized and credited another evaluation.8 For the foregoing reasons, we conclude that the district court did not err, at least not clearly so, in determining that there was no reasonable possibility counsel could have affect- ed the outcome of Klat's competency hearing. Accordingly, we affirm the district court. So ordered. __________ 8 Although Klat claims she has a right to another evaluation, that decision is committed to the trial court's discretion. See 18 U.S.C. s 4247(b) (providing that additional psychiatric or psychological examinations may be conducted "if the court finds it appropriate"); cf. id. s 4241(b) (same as to initial examination).