STEPHEN F. WILLIAMS, Circuit Judge,
concurring in the denial of rehearing in banc:
The imperfections of our “direct evidence” rule are plain to see. Kimberlin v. Quinlan, 6 F.3d 789, 798 (D.C.Cir.1993) (Williams, J., concurring). But it represents our effort, most expressly stated in Seigert v. Gilley, 895 F.2d 797, 800-02 (D.C.Cir.1990), aff'd on other grounds, 500 U.S. 226, 111 S.Ct. 1789, 114 L.Ed.2d 277 (1991), to reconcile the extension of Bivens liability to constitutional torts based on motive, see, e.g., Davis v. Passman, 442 U.S. 228, 99 S.Ct. 2264, 60 L.Ed.2d 846 (1979), with the ’ Supreme Court’s determination in Harlow v. Fitzgerald, 457 U.S. 800, 102 S.Ct. 2727, 73 L.Ed.2d 396 (1982), that qualified immunity should afford officials substantial protection not merely from ruinous financial liability, but also from the burdens of litigation, including the burdens of discovery. Id. at 814-19, 102 S.Ct. at 2736-39. It was, of course, precisely because of the Court’s recognition that “questions of subjective intent so rarely can be decided by summary judgment”, id. at 816, 102 S.Ct. at 2737, that it reformulated qualified immunity in Harlow to exclude subjective intent. There may well be better reconciliations than the “direct evidence” rule, ranging from officials’ full exposure to conventional discovery, to withdrawal of Bivens liability for constitutional torts involving subjective intent. Drawing the line depends, under the Supreme Court’s view of the Bivens remedy, on the trade-off between the benefits of assuring citizens compensation for *1526unconstitutional acts and the costs of exposing officials to many suits that, though ultimately meritless, can only be proven merits less after great expense in time, stress and money. Indeed, as Justice Harlan observed in Bivens itself, the “range of policy considerations” to be taken into account is “at least as broad as the range a legislature would consider with respect to an express statutory authorization of a traditional remedy.” Bivens v. Six Unknown Named Federal Narcotics Agents, 403 U.S. 388, 407, 91 S.Ct. 1999, 2010, 29 L.Ed.2d 619 (1971) (Harlan, J., concurring). As existing precedents and our status in the judicial hierarchy preclude us from adopting many of these solutions, it seems to me on balance preferable to continue with the imperfections of Seigert until the Supreme Court resolves the issue.
SILBERMAN, Circuit Judge,
concurring in the denial of rehearing en banc:
I very much appreciate Judge Edwards’ endorsement of my views as to the core ingredient of judicial restraint — avoidance of judicial policymaking, policymaking often exercised in pursuit of such elusive goals as “justice,” a tranquil population, or the rearrangement of our society through the discovery of new constitutional rights for favored groups or individuals. He overlooks the point, however, that Justice Harlan made when he concurred in Bivens v. Six Unknown Named Agents of Federal Bureau of Narcotics, 403 U.S. 388, 398, 91 S.Ct. 1999, 2005, 29 L.Ed.2d 619 (1971) (Harlan, J., concurring), and which the whole Court has subsequently endorsed. See Bush v. Lucas, 462 U.S. 367, 376-78, 103 S.Ct. 2404, 2410-12, 76 L.Ed.2d 648 (1983). Recognizing with commendable candor that the Court had itself created a cause of action that neither the Constitution nor Congress had fashioned, Justice Harlan emphasized the open policy-making role that the Court had assumed. See Bivens, 403 U.S. at 407-10, 91 S.Ct. at 2010-12. And in a series of cases since Bivens, the Supreme Court and the lower federal courts have struggled with the costs and benefits of a Bivens action, perhaps more openly resembling legislatures than courts normally do. See Bush, 462 U.S. at 378, 103 S.Ct. at 2411; Hubbard v. EPA Administrator, 809 F.2d 1, 9 (D.C.Cir.1986), adhered to on reh’g en banc sub nom. Spagnola v. Mathis, 859 F.2d 223 (D.C.Cir.1988) (per curiam).
Those who criticize this circuit’s direct versus circumstantial evidence distinction, which stems from Martin v. D.C. Metropolitan Police Dept., 812 F.2d 1425, 1435-36 (D.C.Cir.1987), have a logical point, but it is a point that focuses on the trees rather than the forest. As Judge Williams suggests, the Supreme Court is better positioned than we to resolve the broader problem, and, in the meantime, it can hardly be thought “an extraordinary use of judicial authority” to rest on our precedent.
Of course, in Bivens cases, individual government officers are sued personally, and are sometimes obliged to pay attorneys fees out of modest government salaries. That is why the Court has recognized that early dismissal of all but the few obviously meritorious cases is desirable. Apparently, the wisdom of that proposition is easier for some judges to perceive when the potential defendants are part of the judiciary. See Whitacre v. Davey, 890 F.2d 1168 (D.C.Cir.1989), cert. denied, 497 U.S. 1038, 110 S.Ct. 3301, 111 L.Ed.2d 810 (1990), in which no judge dissented from the denial of rehearing.
HARRY T. EDWARDS, Circuit Judge,
with whom Chief Judge MIKVA and Circuit Judge WALD
concur, dissenting from the denial of the suggestion for rehearing en banc:
The panel majority in this case relied on a rule that distinguishes between “direct” and “circumstantial” evidence in reversing the District Court and dismissing the plaintiffs suit for failure to satisfy so-called “heightened pleading” requirements. In response to plaintiffs petition for en banc review, and after consultation with the Solicitor General’s office, the Government notified the court that it would not defend the “direct evidence” standard that had been adopted by the panel majority. Indeed, even the concurring *1527member of the majority acknowledged that the panel’s judgment rested on a rule that is “completely arbitrary and unrelated to the strength of the plaintiffs case.” Kimberlin v. Quinlan, 6 F.3d 789, 798 (D.C.Cir.1993) (Williams, J., concurring). Yet, by a margin of 5 (in favor of en banc review) to 4 (against), with one judge recused, a plurality of the court has blocked en banc consideration of this case. This result is unfathomable.
In particular, I find it incomprehensible that this court has refused to rehear a case which is so clearly of great importance, and which rests upon a rule that is concededly “completely arbitrary and unrelated to the strength of the plaintiffs case.” The result of our action is all the more unintelligible in light of the fact that a recent decision of the Supreme Court calls into question whether federal courts may ever apply “heightened pleading” requirements not provided for in the Federal Rules of Civil Procedure. Even assuming the validity of heightened pleading requirements generally, Kimberlin creates a regime under which a civil rights plaintiff must prove more to survive a motion to dismiss than he or she must prove in order to win at trial. This rule is nothing short of Kafkaesque, and ours is the only circuit that has failed to recognize this fact.
First, the Supreme Court’s decision in Leatherman v. Tarrant County Narcotics Intelligence & Coordination Unit, — U.S. -, 113 S.Ct. 1160, 122 L.Ed.2d 517 (1993), issued just six months before Kimberlin, strongly suggests that federal courts may not impose so-called “heightened pleading” requirements in civil rights cases. However, under the rule announced by the majority in Kimberlin, a plaintiff making a civil rights claim that requires proof that a Government official acted with an unconstitutional motive must plead “direct evidence” of the official’s intent. Plaintiffs alleging unconstitutional motive are not permitted any discovery, and are subject to dismissal unless they meet this standard. See Kimberlin, 6 F.3d at 804-05 n. 5 (Edwards, J., dissenting). The majority’s opinion in Kimberlin cannot be squared with Leatherman.
The Court in Leatherman held that a “ ‘heightened pleading standard’ — more stringent than the usual pleading requirements of Rule 8(a) of the Federal Rules of Civil Procedure — could not be applied in civil rights cases alleging municipal liability under [42 U.S.C. § 1983 (1988)].” — U.S. at -, 113 S.Ct. at 1161. Although Leather-man did not consider the permissibility of heightened pleading requirements in cases involving individual Government officials, because that question was not before the Court, the logic of the decision indicates that all such requirements are improper, except as expressly required by the Federal Rules. This conclusion can be gleaned from the Court’s unequivocal statement that “it is impossible to square the ‘heightened pleading standard’ ... with the liberal system of ‘notice pleading’ set up by the Federal Rules.” Id. — U.S. at -, 113 S.Ct. at 1163 (em*1528phasis added). Further, Leatherman observed that Rule 9(b) imposes a particularity-requirement only in cases alleging fraud or mistake, concluding that under the familiar doctrine of expressio unius est exclusio al-terius, federal courts were not free to establish “a more demanding rule for pleading a complaint under § 1983 than for pleading other kinds of claims.” Id.
Although recognizing that the pleading standard invalidated in Leatherman was prompted by legitimate policy concerns, the Supreme Court underscored its holding by stating that heightened pleading requirements could only be imposed by “amending the Federal Rules, and not by judicial interpretation.” Id. Thus, it seems clear that this court has no authority to depart from the Federal Rules on the ground that some judges believe that the purposes of qualified immunity might be better served by imposing more stringent pleading requirements. Leatherman’s emphatic language seems to admit of only one reading: heightened pleading standards other than those set forth in Rule 9(b) are simply impossible to square with the Federal Rules of Civil Procedure, and are therefore invalid.
Second, even assuming that some form of heightened pleading requirement may be imposed in this case, the rule announced in Kimberlin strayed from binding circuit precedent by holding that our heightened pleading standard requires a plaintiff to plead direct, as opposed to circumstantial, evidence of intent. In Hobson v. Wilson, 737 F.2d 1, 29 (D.C.Cir.1984), cert. denied, 470 U.S. 1084, 105 S.Ct. 1843, 85 L.Ed.2d 142 (1985), the case which is the foundation for later decisions purporting to announce a direct evidence requirement, this court held that when the unconstitutional motive of Government officials is an essential element of a claim, a plaintiffs pleading must rest on “non-conclu-sory allegations of evidence of such intent” in order to proceed to discovery on the claim. Hobson did not distinguish between direct and circumstantial evidence — indeed, the intent of three of the five defendants in that case was established solely by circumstantial evidence. See Kimberlin, 6 F.3d at 805-06 (Edwards, J., dissenting). The direct evidence requirement grows out of this court’s later decisions interpreting Hobson, particularly Siegert v. Gilley, 895 F.2d 797, 804 (D.C.Cir.1990), aff'd on other grounds, 500 U.S. 226, 111 S.Ct. 1789, 114 L.Ed.2d 277 (1991). See Kimberlin, 6 F.3d at 805-808 (Edwards, J., dissenting) (discussing genesis of direct evidence rule in misreadings of Hobson). At best, then, the law of the circuit is in disarray, and en banc review is needed to repair this situation.
In addition, it is noteworthy that, when the Supreme Court upheld Siegert on grounds not relied upon by the panel in that case, four Justices made plain that they believed that the direct evidence rule was error. In a separate concurrence, Justice Kennedy wrote that he would “reject ... the Court of Appeals’ statement that a plaintiff must present direct, as opposed to circumstantial, evidence.” Siegert, 500 U.S. at 235-36, 111 S.Ct. at 1795 (Kennedy, J., concurring). Instead, Justice Kennedy concluded that a plaintiff should be required to “put forward specific, nonconclusory factual allegations which establish malice, or face dismissal.” Id. In another separate opinion authored by Justice Marshall (joined by Justices Black-mun and Stevens), the dissenters in Siegert stated that this court “erred in holding” that a plaintiff must proffer “direct evidence of the unconstitutional motive.” Id. 500 U.S. at 246, 111 S.Ct. at 1800 (Marshall, J., dissent ing).
Finally, neither logic nor law offers any support for the direct evidence requirement. The probative value of circumstantial evidence is “intrinsically no different from testimonial evidence,” Holland v. United States, 348 U.S. 121, 140, 75 S.Ct. 127, 137, 99 L.Ed. 150 (1954), and such evidence can in some cases be “more certain, satisfying and persuasive than direct evidence.” Michalic v. Cleveland Tankers, Inc., 364 U.S. 325, 330, *152981 S.Ct. 6, 10, 5 L.Ed.2d 20 (1960). In both civil and criminal contexts, motive is usually established by circumstantial evidence because it is virtually impossible to obtain any other evidence of a defendant’s state of mind. See generally Kimberlin, 6 F.3d at 809 (Edwards, J., dissenting) (collecting and discussing cases). Moreover, neither this nor any other circuit requires civil rights plaintiffs to prove intent at trial via direct evidence— circumstantial evidence is perfectly adequate proof. Indeed, circumstantial evidence of intent is sufficient proof even in criminal prosecutions. Thus, Kimberlin countenances dismissal of claims that would easily be meritorious at trial, simply because a plaintiff cannot, before even beginning discovery, produce direct evidence as to a defendant’s motive.
Quite simply, as noted by the Seventh Circuit in a decision rejecting the direct evidence requirement, and by Judge Williams in his concurrence to the panel’s opinion in Kimberlin, the direct evidence requirement will prove fatal to all attempts to prove motive, except in the rare case in which a defendant confesses. See id. at 798 (Williams, J., concurring); Elliott v. Thomas, 937 F.2d 338, 345 (7th Cir.1991), cert. denied, — U.S. -, 112 S.Ct. 973, 117 L.Ed.2d 138, and cert. denied, — U.S. -, 112 S.Ct. 1242, 117 L.Ed.2d 475 (1992). As Judge Williams correctly observed, the rule adopted by the Kimberlin panel requires dismissal of a claim brought by a plaintiff who adduces overwhelming circumstantial evidence of guilt, but would permit another plaintiff to proceed based upon a single perjured contention that he overheard a defendant’s confession. Kimberlin, 6 F.3d at 798 (Williams, J., concurring). Such a result is, as the Kimberlin concurrence forthrightly admits, “completely arbitrary and unrelated to the strength of the plaintiffs case.” Id. It is no wonder that all of the other circuits that have considered a direct evidence requirement have explicitly rejected it. See id.; Branch v. Tunnell, 937 F.2d 1382, 1386-87 (9th Cir.1991) (rejecting direct evidence requirement “[bjecause evidence of intent is largely within the control of the defendant and often can be obtained only through discovery”); Elliott, 937 F.2d at 345 (noting that there is no principled difference between direct and circumstantial evidence”); Crutcher v. Kentucky, 883 F.2d 502, 504 (6th Cir.1989) (rejecting direct evidence requirement because it would require plaintiff “to come forth with more evidence than he would have to produce to prevail on the merits”).
By refusing to rehear this case en banc, a plurality of this court has imposed a rule which is irrational and arbitrary, and which finds no support in the Federal Rules of Civil Procedure. There has been no suggestion from the judges voting to deny rehearing that Supreme Court precedent compels this circuit to adopt a direct evidence rule. The plurality instead makes a self-styled “tradeoff’ between competing ends. Concurrence to Denial of Suggestion for Rehearing En Banc (Williams, J.). This “trade-off,” however, effectively closes the courthouse doors to any plaintiff to whom it applies, denying relief to a whole class of claims, apparently only on the basis of the plurality’s conviction that the Supreme Court has gone too far in permitting them.
The court’s action in denying en banc review rests on an extraordinary use of judicial authority, one that I thought was not tolerad ed by those who purport to believe in “judicial restraint.” See, e.g., Laurence H. Silber-man, Chevron — The Intersection of Law & Policy, 58 GeoWashL.Rev. 821, 821 (1990) (noting that the author counts himself among those who have “decried the extraordinary expansion of judicial power in the latter half of this century;” and observing that the one concept that most distinguishes those who advocate “judicial restraint” is “avoidance of judicial policy making”). The plurality’s decision has its roots in policy choices, not in law. This is not surprising, however. For, as one of my colleagues has observed, “When judges think of themselves as bearing responsibility for the results dictated by a neutral application of the law, ... they tend to exceed the appropriate bounds of judicial restraint.” United States v. Pryce, 938 F.2d 1343, 1355 (D.C.Cir.1991) (Silberman, J., dissenting in part).
Bivens v. Six Unknown Named Agents of Federal Bureau of Narcotics, 403 U.S. 388, *153091 S.Ct. 1999, 29 L.Ed.2d 619 (1971), and its progeny are the law of the land and control the actions of this court. I am amazed that a plurality of the members of this circuit have chosen to adopt an admittedly arbitrary and illogical rule that effectively renders impossible all Bivens-type civil rights actions that turn on the intent of a Government official. This court is bound by Supreme Court precedent, by fidelity to our own precedents, and-by simple logic, both to rehear and revise Kimberlin v. Quinlan. It is my hope that the Supreme Court will see fit to reverse the error that this decision introduces into the law of this circuit.
I dissent.