Moore, William v. Hartman, Michael

388 F.3d 871, 363 U.S. App. D.C. 350, 2004 U.S. App. LEXIS 23384, 2004 WL 2514383

• Court: Court of Appeals for the D.C. Circuit
• Decided: November 9, 2004
• Docket: 03-5241
• Status: Published

Opinion

Opinion for the Court filed by Circuit Judge TATEL.

TATEL, Circuit Judge:

Qualified immunity generally shields public officials from civil damages “insofar as their conduct does not violate clearly established statutory or constitutional rights of which a reasonable person would have known.” Harlow v. Fitzgerald, 457 U.S. 800, 818, 102 S.Ct. 2727, 2738, 73 L.Ed.2d 396 (1982). In this case, appellee William G. Moore, Jr., claims that government officials — in particular six postal inspectors — pursued criminal charges against him in retaliation for his political activities. The postal inspectors argue that even though the criminal charges against Moore were dismissed, they enjoy qualified immunity because probable cause supported the prosecution. At the time of Moore’s indictment, however, the clearly established law of this circuit barred government officials from bringing charges they would not have pursued absent retaliatory motive, regardless of whether they had probable cause to do so. Because a reasonable jury could find on the basis of the record before us that Moore’s prosecution violated this standard, we reject the inspectors’ immunity defense and affirm the district court’s denial of summary judgment on this issue.

I.

In the mid-1980s, William G. Moore, Jr., served as CEO of Recognition Equipment, Inc. (“REI”), a company specializing in optical scanning technology. Among other products, REI produced a multi-line optical character reader (“MLOCR”) — a device capable of mechanically interpreting multiple lines of text. Encouraged by some $50 million in research and development funding REI had received from the U.S. Postal Service (“USPS”), Moore urged Postmaster General (“PMG”) William F. Bolger to consider purchasing REI’s MLOCRs to aid the USPS in automating its mail sorting functions. Moore was disappointed, however. Since the late 1970s, the USPS had been pursuing an initiative, known as “Zip + 4,” to add four digits to existing five-digit zip codes; with the new nine-digit codes, efficient automatic sorting required scanning only a single line of text, rather than the multiple lines read by REI’s device. Accordingly, PMG Bolger — a staunch supporter of Zip + 4— announced in late 1983 that the USPS would stick with single-line optical character readers (“SLOCRs”) instead of using REI’s product.

Zip + 4, however, was politically controversial. “Bureaucratic arrogance,” one senator called it. Another urged the USPS to “Zap the ZIP!!” In December 1981, the House Committee on Government Operations accused the USPS of “repeatedly overstating] and misrepresenting] the benefits that might accrue” due to the nine-digit codes. And despite PMG Bolger’s testimony that prohibiting Zip + 4 would “cut the Postal Service from the only major opportunity it now has to meet all its obligations at controlled costs,” Congress imposed a two-year moratorium on Zip + 4 in July 1981 and barred the USPS from making the nine-digit codes mandatory.

Chagrined by PMG Bolger’s procurement of SLOCRs, Moore plunged REI into the political fray. To members of Congress and USPS governors, he argued that REI’s MLOCRs were superior technology because they were not dependent on Zip + 4. He also pointed out that unlike SLOCRs, REI’s MLOCRs were American-made. USPS managers reacted angrily: PMG Bolger told Moore to “back off,” and another top official told Moore REI would never receive USPS business. Moore’s position nevertheless gained influence. Several members of Congress pressed REI’s case with the USPS Board of Governors, and Representative Martin Frost, working closely with Moore, introduced legislation (later withdrawn) to force USPS to buy American-made MLOCRs. More important, the General Accounting Office (now the Government Accountability Office) and the Office of Technology Assessment (“OTA”) produced reports concluding that the USPS’s operational losses due to the use of SLOCRs rather than MLOCRs exceeded one million dollars a day. The OTA report attributed the procurement of SLOCRs to unrealistic expectations for Zip + 4, noting that while MLOCR technology might have been inferior in the past, it was now “fully competitive,” making it unreasonable for USPS to continue using single-line technology despite low usage of the nine-digit codes.

Responding to these pressures, the USPS Board of Governors voted in July 1985 to make a “mid-course correction” and switch to multi-line technology. Although this was just what Moore’s media and lobbying campaign had sought, the result turned out unhappily for Moore and his company.

In the months following the mid-course correction, the USPS Postal Inspection Service uncovered two criminal schemes relating, at least incidentally, to REI. The first, a kickback arrangement, involved a USPS Governor, Peter Voss, and a consulting firm, Gnau & Associates, Inc. (“GAI”), that REI had hired in connection with its lobbying campaign. As it turned out, GAI was paying Voss for referrals, and three GAI officers' — John Gnau, Jr., Michael Marcus, and William Spartin — had agreed to share the proceeds of the REI contract with Voss. The second scheme, the details of which are unimportant to this case, involved Spartin’s and REI’s role in the search for a new PMG. In connection "with these two schemes, Voss, Gnau, and Marcus pleaded guilty to criminal charges, while Spartin accepted immunity in exchange for cooperation.

Having uncovered these crimes, the postal inspectors sought to determine whether anyone at REI had participated in them. Following an investigation we describe in more detail below, a grand jury returned a seven-count indictment against Moore, REI, and REI’s Vice President for Marketing, Robert Reedy, in October 1988. The case went to trial a year later, but six weeks into the proceedings at the close of the government’s case, the district court issued a judgment of acquittal. See United States v. Recognition Equip. Inc., 725 F.Supp. 587, 587-88, 602 (D.D.C.1989). Emphasizing a “complete lack of direct evidence to suggest the Defendants knew of the illegal payoff scheme,” id. at 596, the district court concluded, “The government’s evidence is insufficient, even when viewed in the light most favorable to it, for a trier of fact to find guilt beyond a reasonable doubt. Much of what the government characterizes as incriminating evidence is not persuasive of guilt when viewed in its full context. In fact, some of the government’s evidence is exculpatory and points toward innocent conduct of the Defendants.” Id. at 587-88.

Exonerated of the criminal charges, Moore set about obtaining civil damages for the harm to his life and career. Joined by his wife, Moore began by filing a complaint in the Northern District of Texas asserting constitutional claims under Bivens v. Six Unknown Named Agents of Federal Bureau of Narcotics, 403 U.S. 388, 91 S.Ct. 1999, 29 L.Ed.2d 619 (1971), against the prosecutor and six postal inspectors (one of whom is now deceased). Shortly thereafter, the Moores filed a second complaint, also in the Northern District of Texas, seeking recovery from the United States under the Federal Tort Claims Act (“FTCA”), 28 U.S.C. §§ 2671-2680. The Texas federal court dismissed Mrs. Moore’s claims for lack of standing; found that absolute immunity barred the claims against the prosecutor; and, citing qualified immunity, threw out a Fifth Amendment abuse-of-process claim against the inspectors. Moore v. Valder, No. 91-2491 (N.D. Tex. Sept. 21,1992). The court transferred the remaining claims to the U.S. District Court for the District of Columbia, which dismissed the entire suit. Moore v. Hartman, No. 92-2288, 1993 WL 405785 (D.D.C. Sept. 24,1993).

Reviewing the decisions of the D.C. and Texas district courts, we reinstated certain claims against the prosecutor and the United States along with a retaliatory prosecution Bivens claim against the postal inspectors. Moore v. Valder, 65 F.3d 189 (D.C.Cir.1995) (“Moore I”). On remand, the district court denied the inspectors’ motion for summary judgment, allowing limited discovery on the retaliatory prosecution claim. As to the prosecutor and the United States, however, the court again dismissed Moore’s claims. Moore v. Valder, No. 92-2288 (D.D.C. Feb. 5, 1988). Moore appealed a second time, and we affirmed the district court’s ruling except as to one FTCA claim not relevant here. Moore v. United States, 213 F.3d 705 (D.C.Cir.2000) (“Moore IF).

The inspectors, setting up the issue we now face, again sought summary judgment on the retaliatory prosecution claim, this time on the theory that they enjoy qualified immunity because probable cause supported Moore’s prosecution. In the alternative, the inspectors argued that the record contained insufficient evidence of retaliatory motive. The district court denied the inspectors’ motion in the following one-paragraph order:

Upon consideration of the motion of defendants, United States and Michael Hartman, et al., for summary judgment and the response thereto, the Motion for Summary Judgment is DENIED. There are material facts in dispute. The most significant are the facts surrounding the presentation of evidence to the grand jury and the disclosure of grand jury testimony as to a key prosecution witness.

The inspectors now appeal, arguing, as they did in the district court, that they enjoy qualified immunity because they had probable cause to pursue the criminal charges against Moore.

II.

Before addressing the merits of the inspectors’ qualified immunity claim, we must consider whether we have jurisdiction over this interlocutory appeal. Though 28 U.S.C. § 1291 permits us to hear appeals only from “final decisions” of the district court, denial of a claim of qualified immunity falls within the “small class” of collateral orders subject to immediate appeal under that statute despite the absence of a final judgment. See Mitchell v. Forsyth, 472 U.S. 511, 524-25, 530, 105 S.Ct. 2806, 2814-15, 86 L.Ed.2d 411 (1985). The reason for this is simple: appeal after trial cannot remedy an erroneous denial of qualified immunity, since by then the defendant will already have suffered the burdens of litigation the immunity is intended to prevent. See id. at 525-30, 105 S.Ct. at 2814-18; Int’l Action Ctr. v. United States, 365 F.3d 20, 23 (D.C.Cir.2004). As Moore observes, however, the collateral order doctrine applies only “to the extent [the denial of qualified immunity] turns on an issue of law.” Mitchell, 472 U.S. at 530, 105 S.Ct. at 2817. Pointing out that many facts in the record are disputed, Moore argues that the inspectors cannot establish a “purely legal” issue subject to interlocutory appeal, id. at 530, 105 S.Ct. at 2817, unless they concede the plaintiffs view of the facts — something Moore says the inspectors refuse to do. Accordingly, Moore argues, we lack jurisdiction to entertain the inspectors’ appeal.

We have little trouble rejecting Moore’s argument. Although in one interlocutory case where we found jurisdiction, we did describe the facts as “effectively conceded,” see Farmer v. Moritsugu, 163 F.3d 610, 614 (D.C.Cir.1998), we never suggested that such a concession was required for jurisdictional purposes. In fact, such a requirement would conflict with Behrens v. Pelletier, 516 U.S. 299, 116 S.Ct. 834, 133 L.Ed.2d 773 (1996), which held that denial of a claim of qualified immunity remains an appealable collateral order even if the underlying facts are disputed — indeed, even if, as in this case, the district court denied the motion for summary judgment due to the presence of material issues of fact. See id. at 312-13, 116 S.Ct. at 841-42. While noting in reliance on Johnson v. Jones, 515 U.S. 304, 115 S.Ct. 2151, 132 L.Ed.2d 238 (1995), that “determinations of evidentiary sufficiency at summary judgment are not immediately appealable merely because they happen to arise in a qualified-immunity case,” Behrens explained that the solution to a disputed record on qualified immunity is the same as in any other summary judgment case: the court determines “what facts the dis trict court, in the light most favorable to the nonmoving party, likely assumed,” performing “a cumbersome review of the record” if necessary. Behrens, 516 U.S. at 313, 116 S.Ct. at 842 (quoting Johnson, 515 U.S. at 319, 115 S.Ct. at 2159). Once the facts are established under that standard, an immunity claim like the inspectors’ raises “the purely legal question of whether or not an official’s actions violate clearly established law,” no less than in an appeal based on agreed facts. See Meredith v. Fed. Mine Safety & Health Review Comm’n, 177 F.3d 1042, 1048-49 (D.C.Cir.1999). Such legal questions — which sharply divide the parties in this case — fall squarely within the collateral order doctrine as expounded in Mitchell v. Forsyth.

Though neither party raises the issue, we also note that our statement in Moore I that “Moore’s retaliatory prosecution claim ... does allege the violation of clearly established law,” 65 F.3d at 196, neither deprives us of jurisdiction nor controls our resolution of the issues before us. The denial of qualified immunity at summary judgment is a “final decision” subject to immediate appeal even if the defendant previously appealed a denial of the same claim on a motion to dismiss. See Behrens, 516 U.S. at 309-11, 116 S.Ct. at 840-41. Thus, although the inspectors conceded in the appeal from their motion to dismiss that Moore’s claim stated a violation of clearly established law, they are free to assert qualified immunity now: the “legally relevant factors bearing upon the [qualified immunity] question will be different on summary judgment than on an earlier motion to dismiss,” because the court now conducts the immunity inquiry based on “the evidence before it,” rather than the pleadings. Id. at 309, 116 S.Ct. at 840. Furthermore, as we explained in Moore II, our opinion in Moore I “said nothing about the elements of [a retaliatory prosecution claim], or whether Moore could succeed on his complaint.” Moore II, 213 F.3d at 709. Accordingly, whether Moore’s cause of action requires lack of probable cause remains a live issue.

III.

As the Supreme Court has recognized, although damages suits like Moore’s “may offer the only realistic avenue for vindication of constitutional guarantees,” such suits also carry substantial social costs, including the expense of litigation, the diversion of official energy, and the risk of deterring legitimate official action. See Harlow, 457 U.S. at 814, 102 S.Ct. at 2736. Striking “a balance between the evils inevitable in any available alternative,” id. at 813, 102 S.Ct. at 2736, qualified immunity protects “government officials performing discretionary functions ... from liability for civil damages insofar as their conduct does not violate clearly established statutory or constitutional rights of which a reasonable person would have known,” id. at 818, 102 S.Ct. at 2738. Underlying this doctrine is the basic principle of fair notice: officials may be held liable if “[t]he contours of the right [are] sufficiently clear that a reasonable official would understand that what he is doing violates that right,” Anderson v. Creighton, 483 U.S. 635, 640, 107 S.Ct. 3034, 3039, 97 L.Ed.2d 523 (1987); otherwise, the unfairness of holding officials responsible on grounds they could not have anticipated trumps the individual’s interest in vindicating transgressed rights. See id. at 641, 107 S.Ct. at 3039; Crawford-El v. Britton, 523 U.S. 574, 590-91, 118 S.Ct. 1584, 1592-93, 140 L.Ed.2d 759 (1998). To ensure that shielding public officials from unclear law does not freeze the law in place, however, courts facing qualified immunity claims ordinarily engage in a two-step inquiry, considering first what the law is, and only then whether that law was clearly established. See Wilson v. Layne, 526 U.S. 603, 609, 119 S.Ct. 1692, 1696, 143 L.Ed.2d 818 (1999). Were the procedure otherwise, constitutional avoidance might lead courts to rest on findings of uncertainty without first clarifying the law for future cases — a result contrary to the interest of both government officials and individuals claiming that such officials violated their constitutional rights. See id.; County of Sacramento v. Lewis, 523 U.S. 833, 841 n. 5, 118 S.Ct. 1708, 1714 n. 5, 140 L.Ed.2d 1043 (1998).

Because the qualified immunity inquiry focuses on whether the officials could have known “what [they were] doing” was unlawful, Anderson, 483 U.S. at 640, 107 S.Ct. at 3039, defining the right “at the appropriate level of specificity” is critical. Wilson, 526 U.S. at 615, 119 S.Ct. at 1700; see also Butera v. District of Columbia, 235 F.3d 637, 646 (D.C.Cir.2001). While the right need not have arisen in identical or even “fundamentally” or “materially similar” circumstances, see Hope v. Pelzer, 536 U.S. 730, 739, 741, 122 S.Ct. 2508, 2515, 153 L.Ed.2d 666 (2002), the right can be considered clearly established only if the unlawfulness was “apparent” in light of pre-existing law, see Anderson, 483 U.S. at 640, 107 S.Ct. at 3039. The “salient question,” then, is “whether the state of the law [at the relevant time] gave [the officials] fair warning that their alleged treatment of [the plaintiff] was unconstitutional.” Hope, 536 U.S. at 741, 122 S.Ct. at 2516.

In this case, Moore seeks to vindicate his right to be free from prosecution undertaken in retaliation for First Amendment activity. The inspectors, though conceding that right generally exists, see Crawford-El, 523 U.S. at 592, 118 S.Ct. at 1593 (describing the “general rule” that “the First Amendment bars retaliation for protected speech” as one that “has long been clearly established”), urge us to define the claim more specifically. Insisting the record shows that they acted with probable cause, the inspectors argue that what they were doing could violate a clearly established right only if the First Amendment prohibits retaliatory prosecution even when probable cause exists. Based on cases from other jurisdictions requiring lack of probable cause as an element of a retaliatory prosecution claim, the inspectors argue that no such right exists, much less a clearly established one. Moore disputes both points in the inspectors’ syllogism: this circuit, he insists, clearly permitted liability despite probable cause at the time of his indictment, and in any event the inspectors acted without sufficient grounds for suspicion.

As instructed by Wilson, we consider this debate in two stages, asking first what the law is, and second whether that law was clearly established at the time of Moore’s indictment. Because, as we shall explain, we agree with Moore that the inspectors may be liable even if they had probable cause, we have no need to determine whether, as the inspectors insist, they actually had probable cause to pursue Moore’s indictment.

Were Moore’s Rights Violated?

The question presented under the first element of the qualified immunity test — does the retaliatory prosecution cause of action require a lack of probable cause? — has already been answered by this circuit. In Haynesworth v. Miller, 820 F.2d 1245 (D.C.Cir.1987), we described the “essential elements of a retaliatory-prosecution claim” as follows:

The Court should consider whether the plaintiffs have shown, first, that the conduct allegedly retaliated against or sought to be deterred was constitutionally protected, and, second, that the State’s bringing of the criminal prosecution was motivated at least in part by a purpose to retaliate for or to deter that conduct. If the Court concludes that the plaintiffs have successfully discharged their burden of proof on both of these issues, it should then consider a third: whether the State has shown by a preponderance of the evidence that it would have reached the same decision as to whether to prosecute even had the impermissible purpose not been considered.

Id. at 1257 n. 93 (quoting Wilson v. Thompson, 593 F.2d 1375, 1387 (5th Cir.1979) (footnote omitted)). Nowhere does this statement suggest that lack of probable cause is an element of the claim, nor does its silence imply such a requirement. The standard Haynesworth articulated is this: once a plaintiff shows protected conduct to have been a motivating factor in the decision to press charges, the burden shifts to the officials to show that they would have pursued the case anyway. Given that probable cause usually represents only one factor among many in the decision to prosecute — some others being the strength of the evidence, the resources required for the prosecution, the relation to enforcement priorities, and the defendant’s culpability — there is no reason to expect that the mere existence of probable cause will suffice under Haynesworth to protect government officials from liability.

The inspectors insist that this circuit has never “squarely addressed” the issue they raise, leaving us free to require lack of probable cause. (Appellant’s Br. at 25.) Again reading Haynesworth, we disagree. The relevant passage reads as follows:

We share the conviction ... that retaliatory prosecution unconstitutionally impinges on the right of access to the courts guaranteed by the First Amendment. Haynesworth alleged that he was charged with disorderly conduct solely because he refused to release his civil claims against the arresting officers. That averment, we think, partakes from the circumstances enough substance to entitle him to proceed directly under the First Amendment for damages.

Haynesworth, 820 F.2d at 1257 (footnotes omitted). Because this conclusion — that plaintiff had stated a claim for retaliatory prosecution — required some vision of what the claim entailed, Haynesworth's articulation of the elements was central to its holding. True enough, plaintiff described his prosecution as “unmerited,” id. at 1255, and the opinion said the charges arose “solely because” of protected activity, id. at 1257 (emphasis added), implying, perhaps, that plaintiff was prosecuted without probable cause. As we noted above, however, Haynesworth’s description of the cause of action left little doubt that probable cause would not automatically immunize a retaliatory prosecution. Because that description of the tort was part of Haynesworth’s holding, we lack authority to disregard it.

Haynesworth, moreover, is not the only case in which we have suggested liability may arise regardless of probable cause. In Martin v. D.C. Metropolitan Police Department, 812 F.2d 1425 (D.C.Cir.1987), overruled on other grounds by Crawford-El v. Britton, 93 F.3d 813 (D.C.Cir.1996) (en banc), rev’d, 523 U.S. 574, 118 S.Ct. 1584, 140 L.Ed.2d 759 (1998), in which plaintiff alleged that police pressed charges to deter the vindication of civil rights, we noted “the at least arguable existence of probable cause,” yet went on to consider the sufficiency of the motive allegations. Id. at 1434. That disposition implied that a showing of probable cause, by itself, is insufficient to preclude liability. Id. at 1434. As the inspectors argue, Martin could be read to have simply assumed the validity of the claim so as to reach the motive issue, but the case at least reinforces the view that probable cause is not conclusive. Demonstrating the continuing vitality of Haynesworth, moreover, our two prior opinions in this case relied on that decision in discussing retaliatory prosecution. See Moore I, 65 F.3d at 196 & n. 12; Moore II, 213 F.3d at 709.

As the inspectors point out, several other circuits require lack of probable cause in retaliatory prosecution actions. See, e.g., Mozzochi v. Borden, 959 F.2d 1174, 1180 (2d Cir.1992); Merkle v. Upper Dublin Sch. Dist., 211 F.3d 782, 796-97 (3d Cir.2000); Keenan v. Tejeda, 290 F.3d 252, 260 (5th Cir.2002); Smithson v. Aldrich, 235 F.3d 1058, 1063 (8th Cir.2000); Redd v. City of Enterprise, 140 F.3d 1378, 1383-84 (11th Cir.1998). These cases, however, are not the law of this circuit — Haynes- worth is. Besides, two other circuits agree with Haynesworth. See Greene v. Barber, 310 F.3d 889, 897-98 (6th Cir.2002); Poole v. County of Otero, 271 F.3d 955, 961 (10th Cir.2001). Our approach, moreover, comports with the Supreme Court’s framework in Mt. Healthy City School District Board of Education v. Doyle, 429 U.S. 274, 97 S.Ct. 568, 50 L.Ed.2d 471 (1977). In that case, which involved an untenured public school teacher’s claim that the school board fired him because of his First Amendment activity, the Court explained that if the teacher could show his constitutionally protected conduct to have been a “motivating factor” in the firing, the burden would shift to the board to establish that “it would have reached the same decision ... even in the absence of the protected conduct.” Id. at 287, 97 S.Ct. at 576. While the criminal context, of course, involves considerations of prosecutorial discretion absent in a school employment decision, Mt. Healthy provides considerable support for Haynesworth.

Although Haynesworth’s binding effect is enough to end the first part of our qualified immunity inquiry, the inspectors have raised serious objections to our approach, so we think it useful to flesh out the reasons why the existence of probable cause should not necessarily preclude liability. To begin with, probable cause, requiring no more than “information sufficient to warrant a prudent man in believing the suspect has committed or is committing an offense,” United States v. Kayode, 254 F.3d 204, 209 (D.C.Cir.2001) (internal quotations and alterations omitted), is designed for the ordinary arrest or prosecution where courts may presume that government officials exercised their discretion in good faith, so long as their actions were not obviously unfounded. Yet when plaintiffs demonstrate hostility to free speech to have been a motivating factor in the decision to prosecute — as in a prima facie case under Haynesworth— courts may no longer presume that appropriate considerations guided the government’s decision-making. In such circumstances, were courts to demand no more than a showing of probable cause, as the inspectors urge, law enforcement officers could freely bring marginal cases against advocates of disfavored views, even if the officers’ only reason for doing so were hostility to those views. The inspectors’ approach, in other words, interprets the First Amendment to prevent only baseless prosecutions, i.e., prosecutions lacking probable cause. As the inspectors see it, constitutional free speech protections say nothing about prosecutions brought only because the defendant is, say, a peace activist, a Klan member, a Democrat, or a Republican.

In our view, the First Amendment prohibits such targeted prosecutions, just as it prohibits legislation aimed at punishing free speech. To be sure, prosecutorial discretion is a “core executive constitutional function,” United States v. Armstrong, 517 U.S. 456, 465, 116 S.Ct. 1480, 1486, 134 L.Ed.2d 687 (1996), but as the Supreme Court has made clear, “the decision to prosecute may not be deliberately based upon an unjustifiable standard such as race, religion, or other arbitrary classifica tion, including the exercise of protected statutory and constitutional rights.” Wayte v. United States, 470 U.S. 598, 608, 105 S.Ct. 1524, 1531, 84 L.Ed.2d 547 (1985) (internal quotations and citations omitted). Respectful of executive discretion, Haynesworth’s framework allows the government to proceed with prosecutions that, though motivated in part by hostility to First Amendment activity, can be justified on legitimate grounds. When hostility to speech represents a but-for cause of the prosecution, however, the charges are “deliberately based upon an unjustifiable standard.” Id.

We also disagree with the inspectors that analogous First Amendment Bivens claims call for imposing an “objective” threshold requirement relating to the defendant’s culpability. Pointing out that “courts in other contexts have interposed rules requiring some objective showing before scrutinizing a criminal prosecution for bad faith or other ill motive,” the inspectors argue that lack of probable cause should be required in the retaliatory prosecution context because it affords “a ready-made (but not insurmountable) objective criterion as a first step in assessing prosecutorial discretion.” (Appellant’s Br. at 29-30.) Yet the two defenses the inspectors cite in support of their theory— selective prosecution and vindictive prosecution — are hardly irreconcilable with Haynesworth. It is true that a selective prosecution claim requires proof not only that prosecutors acted with bad intent, but also that “similarly situated individuals [outside the protected category] were not prosecuted.” Armstrong, 517 U.S. at 465, 116 S.Ct. at 1487. But once that showing has been made, the accused has a defense to the charges. See United States v. Mangieri, 694 F.2d 1270, 1273 (D.C.Cir.1982); United States v. Steele, 461 F.2d 1148, 1151-52 (9th Cir.1972). Thus, contrary to the inspectors’ theory, selective prosecution doctrine supports our view that constraints on prosecutorial motive may at times override the interest in punishing objectively culpable conduct.

As for vindictive prosecution, that defense entails a framework much like the one Haynesworth adopted for retaliatory prosecution: if evidence indicates a “reasonable likelihood” that the government acted “to punish a defendant for exercising his legal rights,” a presumption of vindictiveness arises, which the government may rebut with “objective information in the record justifying the increased sentence or charges.” United States v. Gary, 291 F.3d 30, 34 (D.C.Cir.2002) (quoting Maddox v. Elzie, 238 F.3d 437, 446 (D.C.Cir.2001)). Though the government’s burden under this standard may be lighter than under Haynesworth, in both casés prima facie evidence of bad motive triggers an obligation on the government’s part to show that permissible considerations supported its action. The standard the inspectors propose, in contrast, requires plaintiffs— the alleged victims — to establish lack of justification in the first instance.

Our reluctance to impose objective limitations finds support in the logic of Crawford-El. In that case, the Supreme Court held that while qualified immunity protects officers who comply with an objectively reasonable view of the law, it affords no protection against claims under clearly established law that entail the subjective element of improper intent, See Crawford-El, 523 U.S. at 593-94, 118 S.Ct. at 1594-95. Consistent with this reasoning, we see no reason why compliance with the objective probable cause standard should bar scrutiny of subjective motivations here. Other constraints identified in Crawford-El — procedural mechanisms for limiting discovery and facilitating summary judgment, as well as the opportunity to show the prosecution would have happened anyway, id. at 592-93, 597-601, 118 S.Ct. at 1593-94, 1596-98 — may screen out baseless motive claims without precluding recovery in cases where officers pursue retaliatory charges they would not have undertaken but for their unconstitutional animus.

In sum, the law of this circuit, as expressed in Haynesworth, affords damages liability for prosecutions that would not have occurred without retaliatory motive, even if the officers involved acted on the basis of probable cause. This theory of liability, we hasten to stress, is limited. Given that probable cause ordinarily suffices to initiate a prosecution, that showing will be enough in most cases to establish that prosecution would have occurred absent bad intent. A Bivens recovery remains possible, however, in those rare cases where strong motive evidence combines with weak probable cause to support a finding that the prosecution would not have occurred but for the officials’ retaliatory animus. In such circumstances, government officers cannot prevail under Haynesworth because they cannot establish that legitimate considerations supported their action.

Moore’s case appears to be an example of this rare circumstance, at least when we view the evidence in the light most favorable to Moore, as we must at summary judgment, see, e.g., Beckett v. Air Line Pilots Ass’n, 59 F.3d 1276, 1278 (D.C.Cir.1995). Looking at the record through that lens, we detect not only strong evidence of retaliatory motive, but also quite weak indicators of probable cause.

Beginning with motive, we think the record permits, at the least, a reasonable inference that the inspectors had Moore’s lobbying campaign in mind as they pursued his indictment. The inspectors referred explicitly to Moore’s political activities in two reports summarizing the evidence in the REI investigation. The first, titled “Arguments for Indicting the Corporation,” lists the following as the first of nine “bas[es]” for indicting REI:

Independent of Voss/GAI actions, the corporation and its PAC funded a media and political campaign to discredit USPS management and cause financial harm to USPS, for example:

a. staged questions and testimony before Congress

b. Frost amendment to freeze USPS appropriations bill.

Similarly, a “Details of Offense” memorandum submitted to the U.S. Attorney’s Office refers to REI’s lobbying activities as evidence that Moore and Reedy (the REI vice president) had “intent to defraud the USPS”:

Moore’s, Reedy’s and REI’s intent to defraud the USPS is evident in the following events and transactions that related to Voss’ official influence but were independently initiated by Moore and Reedy.

• On or about July 25,1985, at Moore’s and Reedy’s suggestion and with their substantial input relative to its drafting, Congressman Frost proposed an amendment to a USPS appropriate [sic] bill that in effect would freeze USPS revenue until MLOCRs were purchased from REI.

• During the period August 1985 to April 1986, REI continued to undermine the competitive testing program [an aspect of OCR procurement] via the media and Congress.

Read in Moore’s favor, these documents suggest that the inspectors regarded Moore’s speech and lobbying — activities clearly subject to First Amendment protection — as grounds for prosecution, even though these activities were “independent of Voss/GAI actions” (presumably a reference to the alleged conspiracy). Consistent with this view of the inspectors’ motives, subpoenas in the REI investigation targeted speech and lobbying activity, seeking, among other things, “articles placed with trade publications and reporters,” “interviews with journalists and reporters,” “meetings with United States Congressmen,” and “consulting services or meetings with or regarding the REI Political Action Committee.”

Reinforcing the inference that the prosecution would not have happened without retaliatory motive, the evidence supporting the government’s case — again, viewed in the light most favorable to Moore — appears quite weak. To begin with, the strongest evidence connecting REI to the conspiracy related to Moore only indirectly. Though Voss, the corrupt USPS governor, called Moore at one point to ask “why hadn’t it [a contract with GAI] been done,” Voss gave his initial referral not to Moore, but to Reedy. Reedy may have known the GAI contract was fishy; at the least, information from Gnau about conversations he had with Reedy gave the inspectors reason to suspect as much. See REI, 725 F.Supp. at 593-94. In addition, Reedy, perhaps revealing a guilty conscience, initially lied to the inspectors about the source of the GAI referral. Id. at 595-96. Yet no record evidence indicates that Reedy shared with Moore whatever misgivings he may have had about the contract.

Attempting to connect Moore to the conspiracy, the inspectors point to several scribbles about Voss and GAI in a notebook Moore labeled “Postal.” One entry, apparently dating from December 18, 1984, reads as follows:

Get John Knau [sic] involved — have broad

scale assoc w/ John — get together

* Call Peter Voss

“The business to be had here is substantial”

Another entry dated April 29, 1985 again mentions Gnau and Voss, while also referring to Zip + 4; to John McKean, the Chairman of the USPS Board of Governors; to Electrocom Automation, Inc., a competing producer of scanning technology; and to James Jellison, a top USPS official:

USPS — prudent to do contingency planning

• ZIP + 4 not going well

• Consultant — wired (Peter Voss)

• Inside vs outside control

• 100 systems — $150m-$250m

• McKean — West Point/airborne/Gonzaga HS

• Upgrade at Electrocom

• Jellison

Elsewhere, the notes appear to refer to information from a “closed session” of the USPS Board of Governors, and an entry dating from January 27, 1987 — more than six months after Voss’s guilty plea — suggests that Moore gave advice to employees in preparation for Postal Inspection Service interviews:

Critical Incident • Final “fishing trip”

• Lawyer in DC — late for hearing — Martin Luther King — no copy of transcript

(plea arrangement) — date of plea — conversation between judge +

U.S. Attorney

• lot of homework

• drive a wedge between people (intimidate)

• answer “I don’t know, I really can’t remember”

• excitable

• all kinds of scenarios

• ask same questions over and over

• don’t show him how smart you are

• don’t relax

• long interrogation (tough questions

at end)

• possible subpoena

Note [illegible] B/S list based (1/27) on high number of charges

The inspectors interpret these notes to show that Moore (1) formed a “broad scale” criminal association with Gnau and Voss, (2) sought “inside control” of the Board of Governors through a “wired” consultant (i.e., Gnau), and (3) obstructed the Postal Inspection Service investigation.

Reading the notes in Moore’s favor, however, we think it at least as plausible that the notes reflect perfectly innocent business considerations, such as Moore’s interest in forming a legitimate relationship with a well-connected lobbyist and protecting his employees from potentially damaging litigation. To be sure, as the inspectors point out, Moore’s interview advice includes no instruction to tell the truth, while the instruction to answer “I don’t know, I really can’t remember” could suggest a coverup. But Moore points to evidence suggesting he did encourage his employees to tell the truth, and in light of that evidence, Moore’s “I don’t know” statement could mean nothing more than that he cautioned employees to avoid guesswork and speculation — guidance that, like Moore’s other notes, reflects standard deposition advice.

Next, the inspectors think it suspicious that Moore’s notebook was missing thirty-six of its eighty pages, and that REI failed to locate certain subpoenaed phone records from late 1984 and early 1985 — the “critical period,” as the inspectors see it, for the formation of the conspiracy. Because Marcus told the inspectors he had heard from Spartin that “Reedy, Moore, Gnau and Voss ... met and developed a story to cover up their involvement,” the inspectors suspected that REI officials removed the pages and records to cover their tracks. Yet phone records were also missing from early 1984 — long before the Voss referral — while REI produced other evidence (including at least one phone message) revealing contacts between Moore and Voss. As for the notebook, Moore explained during his deposition that he often tore out sheets for his secretary to type. Given the posture of this case, we must resolve these ambiguities in Moore’s favor, leading us to conclude that the missing notes and records fail to establish a coverup.

Other evidence points to Moore’s innocence. Though lacking any evident reason to protect Moore, not one of the conspiracy’s admitted members fingered him. Spartin, for example, failed to corroborate Marcus’s assertion that Moore and Reedy agreed to a coverup. In fact, despite extraordinary pressure — at one point as many as ten inspectors surrounded Spartin while an Assistant United States Attorney tore up his immunity agreement — Spartin never indicated that Moore knew of the conspiracy. Instead, Spartin stated that although he “[didn’t] give a hoot and hell about Bill Moore,” he would not “make up a story” to incriminate Moore. Asked whether “anyone in REI knew Peter Voss was involved in this scheme ... or was being paid,” Spartin offered only that other witness statements the inspectors had shown him suggested Moore’s guilt: “Let me answer you this way,” Spartin said. “Being paid, no sir, I don’t. I have no knowledge of that at all. Peter Voss being part of the deal, no knowledge. But, you know I read that goddamn testimony and I’m not a lawyer but Jesus, there’s enough there to seem to me to hang REI from the yardarm.” Voss even told the inspectors there was “no way Moore knew” of anything improper.

Recognizing the deficiencies in the inspectors’ evidence, the U.S. Attorney’s Office hesitated to indict — even though the inspectors urged it to do so. “The facts underlying this [proposed] indictment are complicated, and the evidence is entirely circumstantial,” the Chief and. Deputy Chief of Special Prosecutions wrote in a memo to,the U.S. Attorney. “If this matter goes to trial it will be a very difficult case and consume significant resources.” While concluding — incorrectly, as it turned out — that “there is enough evidence to get by an MJOA [motion for judgment of acquittal],” the memo described the chances of convicting Moore, Reedy, and REI as “questionable.” As to Moore specifically, the two Assistant United States Attorneys observed:

[N]one of the evidence shows direct knowledge by Moore of the payments to Voss through GAL Even when the evidence is considered in light of Moore’s close association with Reedy — from which one can infer that Moore knew of at least some of Reedy’s conversations with Gnau — it proves no more than that Moore probably knew of the payments to Voss.

True enough, Joseph Valder, the AUSA handling the investigation, disagreed with the memo, stating in a response that “hundreds, if not thousands, of pieces of direct evidence ... show that the defendants are guilty beyond a reasonable doubt.” Nevertheless, the opinion of the Chief and Deputy Chief of Special Prosecutions — two experienced prosecutors — that the evidence was “questionable” adds weight to Moore’s assertion that unbiased officials would never have pressed charges against him.

The record also suggests that unusual prodding from the Postal Inspection Service contributed to the eventual decision to indict — an inference that could, again, support Moore’s theory of retaliatory motive. The Chief Postal Inspector, C.R. Clauson, twice wrote to the U.S. Attorney, Jay Stephens, urging him to press charges against Moore, Reedy, and REI. In the second letter, which followed the AUSAs’ memorandum, Clauson wrote, “Frankly, Jay, I am disappointed by your office’s failure to act on this matter and the series of broken promises from your staff (review committee) relative to the date and nature of their recommendation.” Both Clauson and another inspector (one of the defendants in this case) said in their depositions that they were unable to recall the Postal Inspection Service ever sending a similar letter.

Moreover, some record evidence could lead a reasonable trier of fact to conclude that when the U.S. Attorney’s Office finally decided to indict, the inspectors behaved before' the grand jury as if their case heeded bolstering. For example, when Robert Bray, an REI Vice President, wanted to explain in his grand jury statement that to his knowledge Moore and Reedy knew nothing about the payoffs, Valder and the inspectors refused to let him say any such thing, despite protracted negotiations with Bray’s lawyer. Valder apparently circled portions of Bray’s draft statement,and wrote “don’t reveal.” The record also suggests that the inspectors and Valder showed the prepared grand jury statements to Spartin during his polygraph examination, and that they shared investigative materials — allegedly including grand jury evidence — with Bolger’s ousted successor as PMG, Paul Carlin.

Considering all this evidence together and interpreting it in Moore’s favor, we cannot conclude that the postal inspectors would have prosecuted Moore had they not been irked by his aggressive lobbying against Zip + 4. The evidence of-retaliatory motive comes close to the proverbial smoking gun: in addition to subpoenas targeting expressive activity, Moore has produced not one, but two Postal Inspection Service documents specifically referring to his lobbying as a rationale for prosecution. At the same time, evidence of guilt seems quite weak: not only did none of the admitted conspirators implicate Moore, but even the U.S. Attorney’s Office concluded that, at best, Moore “probably” knew about the charged conspiracies, and even that conclusion rested on the assumption that Reedy likely shared with Moore his misgivings about Gnau and Voss — an assumption the record fails to substantiate. Moreover, the U.S. Attorney’s Office warned that the case would be “complicated” and “consume significant resources” — considerations that, under normal circumstances, might weigh against prosecuting a marginal case. Applying the Haynesworth test, we believe this combination of factors — complexity and expense plus strong indications of retaliation and weak evidence of probable cause — suggests not only that hostility to free expression was at least a motivating factor in Moore’s prosecution, but also that the inspectors may be unable to rebut that inference. Accordingly, Moore has alleged the violation of a constitutional right, precluding summary judgment under the first element of the qualified immunity test.

Was the Lato Clearly Established?

As to the qualified immunity test’s second element, Haynesioorth again stands as the key authority. Decided in 1987, a year before Moore’s indictment, Haynesworth clearly stated the elements of retaliatory prosecution, leaving no doubt that government officials could be liable for pressing charges they would not have pursued without bad motive. Our conclusion, then, that the inspectors’ conduct was actionable under Haynesioorth constrains us to hold that Moore has alleged the violation of a clearly established right.

The inspectors’ argument to the contrary misapprehends the standard for clear law. True, Haynesworth stated the elements of retaliatory prosecution “without analysis in a footnote in an opinion generally addressing other issues.” (Reply Br. at 12.) But as we noted earlier, Haynesworth’s description of the elements was part of its holding, and hence binding precedent, even if it appeared in a footnote. In any event, qualified immunity requires only that the law be clear, not that it be stated prominently or elaborately. Here, Haynesworth established the elements of retaliatory prosecution, making plain that what the inspectors were doing — prosecuting a case they otherwise would have left alone — violated the First Amendment. See Anderson, 483 U.S. at 640, 107 S.Ct. at 3039; Butera, 235 F.3d at 646. Neither Haynesworth’s purported lack of analysis nor its use of a footnote freed the Postal Service from the obligation to take note of the opinion and instruct its inspectors accordingly.

Nor did the decisions of other courts give the government reason to doubt that Haynesworth meant what it said. The law of other circuits may be relevant to qualified immunity, but only in the event that no cases of “controlling authority” exist in the jurisdiction where the challenged action occurred. See Wilson, 526 U.S. at 617, 119 S.Ct. at 1700. Here, a decision of this court — Haynes- worth — provided guidance on exactly the issue the inspectors confronted. Moreover, even if cases from other jurisdictions could somehow infuse Haynesworth with ambiguity, they did not do so before 1988, for nearly all decisions on which the inspectors rely came later. At the time of Moore’s indictment, only the Third Circuit required lack of probable cause, see Losch v. Borough of Parkesburg, 736 F.2d 903, 906-09 (3d Cir.1984), although the Eleventh Circuit had hinted at such a requirement in Motes v. Myers, 810 F.2d 1055, 1060 (11th Cir.1987); see also Redd, 140 F.3d at 1383. In contrast, and also at the time of Moore’s indictment, the Fifth Circuit, though later embracing the Third Circuit’s view, see Keenan, 290 F.3d at 260, had stated that an enforcement practice could be unconstitutional “if those who file such charges upon probable cause can be presumed to be motivated by a retributive purpose,” Gates v. City of Dallas, 729 F.2d 343, 346 (5th Cir.1984); cf. Izen v. Catalina, 382 F.3d 566, 571-72 (5th Cir.2004) (holding that although “the government need not have even reasonable suspicion to undertake an investigation,” an investigation undertaken “with the substantial motivation of retaliating” against protected speech may violate the First Amendment). Against this ambiguous background — at best, two circuits immunizing prosecutions based on probable cause and one apparently not — Postal Service officials could not reasonably have read Haynesworth to require lack of probable cause.

To sum up, because Haynesworth’s framework for Moore’s claim is incompatible with the probable cause-based standard the inspectors advocate, we conclude that the Postal Service had, as the Supreme Court put it in Hope, 536 U.S. at 741, 122 S.Ct. at 2516, “fair warning” that government officers could be liable under the circumstances alleged here. Agreeing with the district court, we therefore reject the inspectors’ claim of qualified immunity.

IV.

Some fifteen years after the district court dismissed the indictment and found evidence probative of Moore’s innocence and thirteen years after Moore filed his first complaint, Moore’s attorney quipped at oral argument: “I suppose I’d be the poster boy that a lawyer has to be crazy to take a Bivens case because you die before it ends.” We trust this opinion will reassure both sides — Moore and the postal inspectors — that this ease may now be resolved within the lifetime of their attorneys. With the inspectors’ immunity theory dispatched, nothing stands in the way of a judgment on the merits; indeed, because the district court found material issues of fact in the record, the next step, presumably, will be preparation for trial. We affirm the decision of the district court and remand the case for further proceedings consistent with this opinion.

So ordered.