Opinion for the Court filed by Circuit Judge WALD.
WALD, Circuit Judge:
Philip O’Donnell, a career officer in the District of Columbia’s Metropolitan Police Department, has brought suit against the District of Columbia, the Mayor, and the Chief of Police. O’Donnell claims that his transfer from the job of Deputy Chief of Police for Investigative Services to the (less desirable) job of head of the Department’s Property Division, followed by his demotion to the rank of captain and the position of district patrol supervisor, violated his free speech and due process rights under the Constitution. The district court dismissed all of O’Donnell’s claims, except for one of the due process claims, as to which it granted summary judgment; O’Donnell has appealed. We affirm, except as to one element of O’Donnell’s First Amendment claim, which we believe requires further factual development.
I. Background
O’Donnell’s account of the events surrounding his transfer and demotion is set out below. Many of the facts he recounts are fiercely contested by the defendants; however, because the plaintiffs claims must be taken as true on a motion to dismiss, and the facts on the basis of which the district court granted summary judgment are not contested, we assume them to be true for purposes of the appeal as well.
O’Donnell began his career with the Police Department almost three decades ago, in 1969. Over the years, he rose through the ranks, until, in January 1995, he was appointed D.C. Deputy Chief of Police for the Investigative Services Bureau of the D.C. Police Department. This was a very senior position; O’Donnell had authority over offices specializing in a range of high-profile criminal matters, including the Homicide Branch (a fact whose importance will soon become clear). In his new position, O’Donnell reported to the then Assistant Chief of Police for the Patrol Services Bureau, Larry Souls-by; Soulsby, in turn, reported directly to the Chief of Police. O’Donnell avers in his complaint that he did an excellent job during his tenure, making a number of important reforms in. the work of the offices under his command.
The first signs of trouble came in May of 1995, when Soulsby called O’Donnell into his office to discuss a newspaper article that
reported that the Police Department had failed to resolve a total of 123 unexplained deaths of African-American women over''-a ten-year period. The article had caused considerable public controversy, and
Soulsby was “incensed” because he believed that the story had been leaked by sources within the Police Department — specifically, by Captain William J. Hennessy, the head of the Homicide Branch, who was O’Donnell’s immediate subordinate. O’Donnell denied that Hennessy was the source of the news report, and added that the Homicide Branch was aware of the unsolved homicides and was already conducting an investigation. According to O’Donnell, Soulsby “responded angrily that the article was untrue and ordered that nothing should be done to investigate the allegations.” O’Donnell strongly disagreed, noting that a serial killer might be involved; .at-a routine staff meeting later in..the day, the Chief of Police, Fred Thomas, ordered' an investigation.
Two months later, Thomas retired and Soulsby was appointed intexim chief of police. The week after his appointment, O’Donnell sent him a long memorandum, titled “Setting Priorities for the Metropolitan Police Department,” which ranked crime-related issues and proposed solutions. Soulsby did not respond; O’Donnell claims that this was because Soulsby wished to keep things quiet pending his confirmation as Chief of Police by the D.C. City Council.
Soon afterwards, Soulsby told O’Donnell that Mayor Marion Barry was blocking O’Donnell’s promotion to Assistant Chief of Police (a promotion Soulsby had promised to secure). Soulsby said that the Mayor believed that O’Donnell was a racist; there was, O’Donnell avers, no reason that anyone should believe that this was true. O’Donnell then spent several months off the job ber cause of a heart condition. When he returned, in September 1995, he learned that he was to be transferred to a less responsible position, that of Commander of the Property Division. (O’Donnell did, however, retain his rank as Deputy Chief of Police.) Once again, Soulsby said that the transfer was because Barry thought O’Donnell was a racist.
Soulsby also reassigned Captain Hennessy to a less responsible job, that of Night Supervisor. According to O’Donnell, Soulsby then told the media in an off-the-record statement that Hennessy had been transferred because he was under investigation and likely to be indicted for violating the civil rights of a black suspect. O’Donnell says that when Hennessy learned of this claim he confronted Soulsby, who admitted he had lied. Hennessy apparently taped this conversation, and later told Soulsby that he had done so. As a result, the two allegedly entered into a pact (which O’Donnell refers to as the “Soulsby-Hennessy agreement”) under which Hennessy got a better job (but not in the Homicide Branch), in exchange for silence, especially at Soulsby’s pending confirmation hearing. Soulsby was confirmed as Chief of Police in December of 1995.
In the meantime, according to O’Donnell, he was doing an exemplary job of running the Property Division. In his new position, he was in charge of the Department’s facilities for storing property and automobiles, and supervised personnel at a police drug laboratory. O’Donnell wrote a number of reports about problems at the Division, which Soulsby again ignored; O’Donnell nevertheless implemented a number of the suggestions that they contained.
On May 9, 1996, Soulsby made a speech at the District’s Ballou High School in which he claimed that the backlog of 123 unsolved cases involving African-American women had only been discovered after O’Donnell and Hennessy had been transferred out of then-jobs. O’Donnell protested to Soulsby orally and in writing, and ultimately said that he would retire. Shortly thereafter, the news media learned about the Soulsby-Hennessy agreement, and Hennessy’s tape of his conversation with Soulsby; subsequent news reports were highly critical of the alleged agreement, and of Soulsby’s performance as Chief of Police. For example, on July 14, 1996, the CBS television program
60 Minutes ran a
report that said that the Police Department had performed poorly under Soulsby and that (in O’Donnell’s words) Soulsby had “low credibility.” The report also criticized the Mayor’s poor response to
the scandal surrounding the Soulsby-Hen-nessy agreement, and said that senior executives in the Homicide Branch believed that the reassignment of Captain Hennessy was unjustified and had damaged the Police Department. There were many demands for Hennessy and Soulsby to release the tape. Soulsby called on Hennessy to release the tape, and Hennessy, somewhat inexplicably, declined to do so until Soulsby released him from their agreement. Senior D.C. officials (including a D.C. Council member and the Corporation Counsel) eventually criticized Soulsby for entering into the claimed agreement with Hennessy, and Soulsby publicly apologized.
On July 18, 1996, in the midst of these events, the
Washington Post
published a letter to the editor by O’Donnell, which said:
I am writing to confirm the credibility of Capt. William L. Hennessy of the Metropolitan Police Department. I currently am employed by the Metropolitan Police Department as a deputy chief of police assigned to the technical services bureau as director of the property division.
I was Capt. Hennessy’s bureau chief during the better part of the two years he spent as the commander of the homicide branch. During this time, I observed Capt. Hennessy to be a very loyal and extremely knowledgeable and hard-working individual — one who clearly is dedicated to the Metropolitan Police Department and the citizens of the District of Columbia. Capt. Hennessy is a “cop’s cop”; he is a true leader.
I did not support Capt. Hennessy’s transfer to a “nondescriptive” assignment. It is my opinion that Capt. Hennessy’s transfer from the homicide branch not only was a tragic loss to the Metropolitan Police Department but a disservice to the citizens of the District.
Capt. Hennessy is an extremely credible person — unlike the chief of police.
The following day, Soulsby and Barry demoted O’Donnell to the rank of captain, and ordered him to a comparatively low position as district patrol supervisor.
O’Donnell retired two weeks later, on July 31, 1996. As a result of the fact that he was demoted before he retired, his severance pay and payments for unused annual leave were reduced by an amount that he estimates as $11,250. O’Donnell also says that he had intended to seek out a job as chief of police of a large city, but that his demotion “irreparably stigmatized” him in “the close-knit world of chiefs of police of large cities and municipalities and eliminated him from any real consideration for positions such as these.” O’Donnell ended up with a job as Chief of Police of Brunswick, Maryland.
On September 23, 1996, O’Donnell brought suit against the District of Columbia, against Police Chief Soulsby and Mayor Barry in their official capacities, and against Soulsby in his individual capacity. O’Donnell’s complaint claimed that the foregoing events (1) violated his First Amendment rights; (2) deprived him of a liberty interest in his good name and professional reputation in violation of the due process clause; and (3) deprived him of “his ability to pursue his chosen occupational field,” also in violation of the due process clause. The defendants moved to dismiss or in the alternative for summary judgment. The district court granted the motion to dismiss as to the first two counts, and the motion for summary judgment as to the third. O’Donnell now appeals this order.
The District also argued before the district court that Soulsby was entitled to qualified immunity. The district court did not reach this issue; the District presses it again on this appeal.
II. Analysis
A.
The First Amendment Claim
O’Donnell claims his transfer from the position of Deputy Chief for the Investigative Services Bureau to the job at the Property Division, and subsequent demotion to the rank of captain and reassignment to a job as district patrol supervisor, both occurred in retaliation for exercises of his right to free speech. He cites five events that prompted the retaliation against him: (1) his memo to Soulsby ranking crime issues in the District; (2) his discussions with Soulsby about the need to investigate the deaths of the 123 women; (3) his criticism of Soulsby’s claim in
the Ballou High School speech that the unsolved deaths were only discovered after O’Donnell and Hennessy were transferred; (4) his memos on improvements in the function of the Property Division; and (5) his letter to the
Washington Post.
The speech of public employees like O’Donnell enjoys considerable First Amendment protection; the Supreme Court has “unequivocally rejected” the proposition that public employees “may constitutionally be compelled to relinquish the First Amendment rights they would otherwise -enjoy as citizens to comment on matters of public interest.”
Pickering v. Board of Educ.,
391 U.S. 563, 568, 88 S.Ct. 1731, 20 L.Ed.2d 811 (1968). However, the reach of this protection is not unlimited. A public official seeking to make out a claim of retaliation in violation of her First Amendment rights must meet a four-factor test.
See Hall v. Ford,
856 F.2d 255, 258 (D.C.Cir.1988). “First, the public employee must have been speaking on a matter of. public concern. If the speech is not of public concern, it is unnecessary to scrutinize the basis for . the adverse action absent the most unusual circumstances.”
Tao v. Freeh,
27 F.3d 635, 638-39 (D.C.Cir.1994) (citations omitted). Second, the court must consider whether the governmental interest in “promoting the efficiency of the public services it performs through its employees” without disruption,
Pickering,
391 U.S. at 568, 88 S.Ct. 1731, outweighs the employee’s interest, “as a citizen, in commenting upon matters of public concern,”
id.,
and the interest of potential audiences in hearing what the employee has to say,
United States v. National Treasury Employees Union,
513 U.S. 454, 468, 115 S.Ct. 1003, 130 L.Ed.2d 964 (1995). Third, the employee must show that her speech was a substantial or motivating factor in prompting the retaliatory or punitive act of which she complains.
See Mt. Healthy City School Board of Education v. Doyle,
429 U.S. 274, 287, 97 S.Ct. 568, 50 L.Ed.2d 471 (1977). And finally, the employer should have an opportunity to show “by a preponderance of the evidence that it would have reached the same decision even in the absence of the protected conduct.”
Id.
“The first two factors under the
Pickering
test are questions of law for the court to resolve, while the latter are questions of fact ordinarily for the jury.”
Tao,
27 F.3d at 639.
The District asserts that the first two elements of this test are not satisfied here: it claims that O’Donnell’s speech was not on an issue of public concern, and asserts that, even if it was, the city’s interest in the efficient operation of the Police Department far outweighed O’Donnell’s interest in speaking out. We address these two issues in that order.' •
1.
Ptiblic Concern
The district court concluded that O’Donnell’s memos to and discussions with Soulsby were all on topics of public concern, but that the letter to the editor was not. We agree with the first finding, but disagree with the second; we find that all of the communications at issue in this case were on topics of public concern.
“Speech that concerns issues about which information is needed or appropriate to enable members of. society to make informed decisions about the operation of their government merits the highest degree of first amendment protection.”
Hall v. Ford,
856 F.2d 255, 259 (D.C.Cir.1988). All of the speech at issue.here falls into this category. Each of O’Donnell’s memos to Soulsby involved important issues of Police Department policy — questions of how to rank the Department’s law-enforcement priorities, and how to reform the operations of the Property Division (which were, O’Donnell claims, later shown by news reports to be in some need of reform). His conversation with Soulsby about what priority to give an investigation of over a hundred unsolved murders was also obviously of great public concern; indeed, the conversation was itself prompted by a newspaper article critical of Police Department ineptitude.
The district court concluded that O’Donnell’s letter to the editor was not on a subject of public concern because it “related to longstanding internal disputes among O’Donnell, Hennessy, and Soulsby, and not to matters of public concern about the structure, management, or policies of the Department,” and did
not contain any “information to educate or inform the public about the operation of the MPD other than to air Plaintiffs personal grievances.” O’Donnell’s conversation with Soulsby about the Ballou High School speech also seems to have been motivated at least in part by personal grievances.
In
Connick v. Myers,
461 U.S. 138, 103 S.Ct. 1684, 75 L.Ed.2d 708 (1983), the Court considered the case of an employee of the New Orleans District Attorney’s office, who, dissatisfied with her impending transfer, circulated a questionnaire that asked a series of general questions about levels of office morale and the confidence employees had in their supervisors. The Court concluded that (with one exception not relevant here) the “questions reflected] one employee’s dissatisfaction with a transfer and an attempt to turn that displeasure into a cause célebre” and hence did not raise issues of public concern.
Id.
at 148, 103 S.Ct. 1684.
Connick
emphasized, however, that “[wjhether an employee’s speech addresses a matter of public concern must be determined by the content, form, and context of a given statement, as revealed by the whole record.”
Id.
at 147-48, 103 S.Ct. 1684. Moreover, it is clear that the presence of a personal motivation for an employee’s speech, although certainly a factor in the public-concern analysis, need not destroy the character of a communication as one of public concern.
See Tao,
27 F.3d at 639 (“The motivation of the employee is only one factor to be considered in assessing whether a statement is one of public concern.”). Even a disgruntled employee can raise a matter that is of considerable concern to the public.
See, e.g., Connick,
461 U.S. at 148, 103 S.Ct. 1684;
Tao,
27 F.3d at 640;
Stroman v. Colleton County School Dist.,
981 F.2d 152, 157-58 (4th Cir.1992) (finding that although a letter circulated by a teacher to fellow teachers was prompted by a “personal grievance,” and the letter’s substance, “in large part, seems to be limited to this grievance,” the letter raised questions of budget mismanagement, and hence was arguably on a subject of public concern);
Zamboni v. Stamler,
847 F.2d 73, 77-78 (3d Cir.1988) (finding that a police detective’s objections to a revised promotion scheme were on a subject of public concern, even though the employee was partly motivated by a desire to improve his own chances for a promotion). Indeed, it may be that those employees who are dissatisfied with their workplaces are precisely those who are likeliest to notice malfeasance, and be willing to speak up about it.
We find that there was a substantial public concern with the subject matter of both the letter to the editor and the conversation about the Ballou High School speech, and that the fact that O’Donnell may have acted partly on the basis of personal motivations makes no difference. O’Donnell’s letter to the editor raised two broad issues. First, by saying that Hennessy’s “transfer from the homicide branch not only was a tragic loss to the Metropolitán Police Department but a disservice to the citizens of the District,” it indicated that the Homicide Branch’s work had been impaired by Hennessy’s loss. There had already been some public debate about the effectiveness of the Homicide Branch (in the
60 Minutes
broadcast); O’Donnell’s letter contributed to this debate, and thus was on a subject of public concern.
Second, and more importantly, the letter contributed to the public’s knowledge of the Soulsby-Hennessy agreement, and thus of Soulsby’s fitness for office. The pact involved an attempt by Soulsby to prevent information that might have damaged his chances of confirmation from coming to light; members of the public might legitimately see this agreement as improper, or worse. Although, as we discuss below, there remain questions about precisely
what
information the letter contributed about the pact, it made at least an evaluative contribution. It was therefore on a subject of public concern. A communication that provides information that might help the public to “make informed decisions about the operation of their government mei'its the highest degi'ee of first amendment protection,”
Hall,
856 F.2d at 259; there are few more fundamental decisions that the public can make about its government than assessing the probity of a public official.
See Foster v. Ripley,
645 F.2d 1142, 1148 (D.C.Cir.1981) (observing that an employee’s interest in speaking out is
“entitled to more weight when the employee is ... acting as a whistleblower exposing corruption among public officials”);
cf. Connick,
461 U.S. at 148, 103 S.Ct. 1684 (noting that the plaintiff had not claimed “that the District Attorney’s Office was not discharging its governmental responsibilities,” or sought “to bring to light actual or potential wrongdoing or breach of public trust”).
Turning to O’Donnell’s exchange with Soulsby after the Ballou High School speech, O’Donnell says that he told Soulsby that it was not true that the 123 unsolved murders only came to light after O’Donnell and Hennessy were transferred, and reminded him that Soulsby himself had opposed an investigation. This conversation implicated a number of broader policy issues, including questions about whether Soulsby had been right to transfer Hennessy and O’Donnell to less responsible positions and about what caused the backlog and how it was detected. Thus, this conversation, too, was on a topic of public concern.
2.
Governmental Interest
In the second step of the
Pickering
analysis, we consider whether; the government’s interest in “promoting the efficiency of the public services it performs through its employees,”
Pickering,
391 U.S. at 568, 88 S.Ct. 1731, outweighs the employee’s interest in speaking out and the public’s interest in hearing what he has to say — or, in the case of speech made
in
private (like some of the speech here), the public’s interest in having the employee make himself heard within his organization.
We have already outlined the nature of the interests on the employee’s side of the equation. As to those on the government’s side, the Court explained in
Rankin v. McPherson,
483 U.S. 378, 388, 107 S.Ct. 2891, 97 L.Ed.2d 315 (1987), that we must consider “whether the statement impairs discipline by superiors or harmony among coworkers, has a detrimental impact on close working relationships for which personal loyalty and confidence are necessary, or impedes the performance of the speaker’s duties or interferes with the regular operation of the enterprise.” And, just as we consider the content,-manner, time and place of the speech in deciding whether it is of public concern at all, we take these factors into account in weighing- the governmental interest in regulating the speech.
See Connick,
461 U.S. at 152-53, 103 S.Ct. 1684.
Before we launch into the governmental-interest analysis, we will dispose of two
preliminary -issues. First, the District rightly argues, and the district court correct-, ly agreed, that because of the special degree of trust-and discipline required in a police force there may be a stronger governmental interest in regulating the speech of police officers than in regulating the speech of other governmental employees. The Supreme Court has found that a police department may legitimately regulate its officers’ personal grooming in order to make its officers readily recognizable and promote “esprit de corps,”
Kelley v. Johnson,
425 U.S. 238, 248, 96 S.Ct. 1440, 47. L.Ed.2d 708. (1976), and a number of courts of appeals have extended this principle to regulation of speech.
See, e.g., Tyler v. City of Mountain Home,
72 F.3d 568, 570 (8th Cir.1995) (finding that a police department permissibly demoted a sergeant for sending a confrontational letter to another police department on official letterhead without first clearing the letter with the chief of police, as required by police department policy). We note, however, that the right to regulate an officer’s speech is not absolute,
see Biggs v. Village of Dupo,
892 F.2d 1298, 1303 (7th Cir.1990) (“[Fjreedom of speech is not traded for an officer’s badge”); in particular, when a police officer speaks out on an issue that he is uniquely qualified to address, we must be cautious in accepting the claim that the public interest demands that he be silent.
See id.
(citing
Pickering,
391 U.S. at 572, 88 S.Ct. 1731).
The District also observes that it is especially disruptive for the high-level employees of a governmental agency to express public disagreement with the agency’s policies. This is true; as we said in
Hall,
“[h]igh-level- officials must be permitted to accomplish their organizational objectives through key deputies who are loyal, cooperative, willing to carry out their superiors’ poli
cies, and perceived by the public as sharing their superiors’ aims.” 856 F.2d at 263. An employee may be dismissed who “expresses views on matters within the core of his responsibilities that reflect[ ] a policy disagreement with his superiors such that they could not expect him to carry out their policy choices vigorously.”
Id.
at 265. In
Hall,
for example, the athletic director at the University of the District of Columbia was dismissed after repeatedly and publicly disagreeing with his superiors about the proper response to violations of university and National College Athletic Association rules.
In order to decide whether a public employee is within the “narrow band” of “key deputies” who are subject to this analysis,
id.
at 263, we consider whether the employee’s job relates to policy matters, and whether the employee has “broad responsibilities with respect to policy formulation, implementation, or enunciation.”
Id.
at 264. It is quite clear that O’Donnell’s initial position as head of the Investigative Services Bureau was in this category; in that position, he was in charge of numerous key operational offices of the Police Department, and one of seven members of the Police Department’s Executive Committee, its principal policy-making body.
But when O’Donnell was transferred to the Property Division, he moved outside of the “narrow band” we discussed in
Hall.
His job seems to have had some policy responsibilities; he was in charge of a number of Police Department facilities, and his complaint describes several reforms he instituted in their operations, seemingly on his own initiative. But we doubt that O’Donnell’s policy responsibilities were broad enough, or sufficiently central to the Department’s mandate, to make him one of the “key deputies” discussed in
Hall.
O’Donnell’s work, although important, primarily involved support functions. And, as he describes his status within the Department, his policymaking role was decidedly marginal. Soulsby could safely dispense with answering his memos, and O’Donnell was relegated to the audience at meetings of high-level Police Department officials. On these facts, O’Donnell cannot be labeled a “key deputy.”
With these principles in mind, we turn to an analysis of the speech in this case.
a.
Memoranda.
O’Donnell claims that after he wrote Soulsby a memo about crime-fighting policy in the District, Soulsby transferred him to the Property Division to block discussion of crime in the District until after Soulsby was confirmed. He also avers that Soulsby ignored his memos about policy changes at the Property Division in order to discourage him from uncovering problems at the Department, and that these memos later helped prompt his demotion.
From their descriptions in the record, O’Donnell’s memos consisted entirely of innocuous, practical suggestions about Police Department policy. If O’Donnell’s memos played any role in his transfer to the Property Division (and we do not know for sure what led to that transfer), it might have been in revealing policy differences between O’Donnell and Soulsby of a kind that, under
Hall,
could legitimately prompt a high-level deputy’s transfer out of the inner circle. The same is true of the memos O’Donnell wrote while at the Property Division; it is hard to see how routine policy memoranda could have prompted any adverse response from Soulsby, but, if they did, it would have been the result of disagreements about policy.
b.
Conversations.
O’Donnell also avers that Soulsby transferred him to the Property Division in retaliation for the conversation in which O’Donnell argued that
the 123 deaths should be investigated further, and that his demotion was partially prompted by the later conversation in which O’Donnell reprimanded Soulsby for misleading the public about the investigation of those same deaths.
We begin with the first of these conversations. The District argues that, under
Hall,
Soulsby was entitled to transfer O’Donnell on the basis of the policy differences this (asserted) conversation revealed.
Hall
gives employers considerable leeway to ensure that high-level officials toe the party line, but it does not give them unchecked power to silence them. In some cases, the public interest in a high-level official’s speech will outweigh any interest in that official’s bu-' reaucratic loyalty. Hall permits adverse action on the basis of speech that implicates “the government interest in accomplishing its organizational objectives through compatible policy level deputies,” 856 F.2d at 264; but sometimes what an employee has to say will signal not policy differences but potential serious violations of the public trust. For instance, had the athletic director in Hall been punished merely for suggesting to the university’s board of directors that one of his subordinates had committed a grave act of embezzlement, the result in that case might have been different.
At first glance, O’Donnell’s account of his conversation with Soulsby seems troubling. Supposedly, O’Donnell was punished for disagreeing with Soulsby’s order that nothing should be done to investigate allegations that 123 murders of African-American women had gone unsolved, even though a serial killer might be at large — suggesting that O’Donnell was challenging a severe breach of the public trust. A closer examination, though, reveals the matter is more complicated. O’Donnell gives only the vaguest account of Soulsby’s reasons for opposing an investigation. His complaint does not attribute any dishonorable motive to Soulsby (such as a desire to cover up the Department’s errors), and on this appeal O’Donnell cites no reason to believe that Soulsby had any such motivation. Moreover, O’Donnell says that Soulsby called the article about the backlog “untrue”; although it is unclear what this means, one reading would be that Soulsby believed that there was no unusual backlog because the number of unsolved murders of African-American women was no larger (taking populations into account) than the backlog of unsolved murders generally. If so, the conversation turned on a legitimate disagreement about statistics, rather than on a potential dereliction of duty.
The vagueness of O’Donnell’s complaint, however, renders this reading, and any other, speculative. If it is O’Donnell’s claim that Soulsby retaliated against him for seeking to dissuade Soulsby from committing a grave breach of the public trust (or for some other impermissible reason) he should make that claim more explicitly. In the absence of any such allegation, we must assume that any adverse action taken against him as a result of his conversation with Soulsby occurred because of a legitimate disagreement about policy, and so was permitted by Hall. We thus affirm the district court’s dismissal of this element of O’Donnell’s claim.
As to O’Donnell’s later conversation with Soulsby about the Ballou High School press conference, this incident merits little discussion. There is little evidence of any causal link between this conversation and O’Donnell’s later demotion, which was tied much more closely to O’Donnell’s letter to the editor. And, although the conversation raised some issues about Soulsby’s veracity, it did not touch on the more serious issues surrounding the Soulsby-Hennessy pact. Finally, O’Donnell’s personal motivations for this conversation were especially strong, as he was responding to a speech in which Soulsby had blamed him for the failure to detect the backlog of unsolved murders.
c.
The Letter to the Editor.
We now turn to O’Donnell’s letter to the editor. The letter appears to be in that small category of speech by public employees that has a high potential both to cause disruption and to inform the public. Because both elements of the governmental-interest analysis are thus at their height, we must perform our analysis with especiál caution, to avoid shortchanging the interests on either side of the equation.
We begin with the governmental interests. As the District points out, O’Donnell’s letter presented a substantial risk of harming his relationship with Soulsby and injuring Police Department morale and discipline. And, although (as we have found) O’Donnell was not in Soulsby’s inner circle when he wrote the letter, he was a fairly senior police official, so that it was of some importance that he appear loyal to his superiors. But the circumstances of this case also weaken the governmental interests to some extent. Although it no doubt would have been difficult for Souls-by and O’Donnell to work together after O’Donnell’s letter was written, on the record as it stands their working relationship seems to have been very limited in any case: O’Donnell occasionally wrote Soulsby memo-randa (and received no response), and the two occasionally saw each other at meetings, but there appears to have been little other working relationship.
See Rode v. Dellarciprete,
845 F.2d 1195, 1202 (3d Cir.1988) (noting that the plaintiff and the supervisor she criticized did not work closely together). And, as to the discipline and morale of the Police Department, the frequent media criticism of the Department’s (and Soulsby’s) performance in the weeks preceding O’Donnell’s letter had doubtless already taken its toll. Although this did not leave O’Donnell entirely free to kick the Department while it was down, -the fact that his criticism was cumulative does diminish the harm it caused. Moreover, O’Donnell’s letter followed a
60 Minutes
broadcast which, according to O’Donnell, “documented that senior detectives in the Homicide Branch were firmly convinced that defendant Soulsby’s reassignment of Capt. Hennessy was unjustified and had damaged the MPD,” and in which Hennessy acknowledged the existence of a tape of his confrontation with Soulsby. There is no indication on the present record that either the “senior detectives” or Hennessy were ever punished for these statements, which were presumably also quite injurious to Department discipline and morale; this weakens the District’s claim that O’Donnell’s letter could not safely be left unpunished.
To turn to the public-concern side of the equation, O’Donnell’s letter makes two principal points: it observes that Hennessy’s transfer out of the Homicide Branch hurt the Branch, and it endorses the credibility of Hennessy and questions that of Soulsby. The second comment overshadows the first both in its potential value to the public and in its potential disruptiveness; we will therefore focus our analysis of that aspect of the letter.
As we have said, information about the pact was clearly of great concern to the public; concrete additional information about the circumstances of the pact might well have overcome any governmental interest in silencing O’Donnell. But it is unclear on the present record what O’Donnell’s letter really contributed to the public’s knowledge about the Soulsby-Hennessy pact. The value of a communication to the public hinges on the extent to which a reasonable member of the public would give it weight in deciding issues of public importance. For instance, the public would reásonably give little or no weight to expressions of personal animus or spleen, but great weight to highly probative evidence bearing on problems of public importance.
The Federal Rules of Evidence provide a useful guidepost for assessing the probative value of O’Donnell’s comments. (We emphasize that we. use these rules only as a framework for our analysis; by no means do we purport to import the rules of evidence wholesale into the First Amendment.) O’Donnell’s letter is cast in terms of “credibility,” but it is unclear how the public would have interpreted this term (or indeed what he meant by it). If he meant merely to comment generally on the character of Hennessy and Soulsby, then the letter is of little value to the public. As the Advisory Committee Note to Federal Rule 404 observes, “[cjharacter evidence is of slight probative value and may be very prejudicial,” as it
“tends to distract the trier of fact from the main question of what actually happened on the particular occasion.” But if the letter is read as a commentary on the believability of recent statements by Hennessy and Soulsby, it might have more significance. As Rule of Evidence 608 indicates, evidence as to the credibility of a witness can at times have substantial probative value.
We do not know which reading a reasonable member of the public would have adopted, or to what extent the credibility of statements by Hennessy and Soulsby was important in the public’s. mind when O’Donnell wrote his letter. It is unclear on the present record to what extent Soulsby contested Hennessy’s account of the pact. O’Donnell says that Soulsby at some point publicly apologized for entering into his pact with Hennessy, but we do not know whether this apology occurred before or after O’Donnell wrote his letter. It may also be that the state of the debate was such that an informed reader would infer from O’Donnell’s letter that his comments on Hennessy’s and Soulsby’s veracity were based not only on a general knowledge of their character, but also on personal knowledge of some of the events surrounding the Soulsby-Hennessy agreement. The allegations surrounding the pact are exceptionally intricate (one expects a Maltese falcon to enter into the story at any moment). In the eyes of an informed reader, some of the statements in O’Donnell’s letter, such as his discussion of Hennessy’s illustrious service at the Homicide Branch, might have taken on a significance that is invisible to us now;
The district court dismissed O’Donnell’s First Amendment claims before any record had been developed, and indeed before the District filed an answer. We conclude that it is not possible to decide what value O’Donnell’s letter would have had to the public without a more complete record as to the state of the public debate at the time the letter was written. If the public’s view of the Soulsby-Hennessy pact would reasonably have been influenced by evidence as to Hennessy’s (and Soulsby’s) credibility,' and if O’Donnell’s words on that subject would have carried weight, then O’Donnell’s comments might have made a sufficient contribution to the public debate to outweigh the governmental interest in his silence. There are also some uncertainties in our analysis of the nature of the governmental interest; it is unclear, for instance, whether we are correct to assume that a number of other Police Department officials had criticized Soulsby in public with impunity. The district court should permit the parties to develop these issues further on remand.
B.
Due Process Clause
O’Donnell’s complaint attempts to plead two distinct due process claims. His first is based upon the stigma and loss of “rank, status and pay” associated with the District’s reassigning him to the Property Division, blaming him for the failings of the Department, and then demoting him. His second is based upon the harm to O’Donnell’s employment opportunities associated with these acts.
These claims are not really distinct under applicable due process law, and they are in any event too broad. A loss of rank, status and pay alone is not actionable under the Due Process Clause. The Supreme Court found in
Board of Regents v. Roth,
408 U.S. 564, 573-78, 92 S.Ct. 2701, 33 L.Ed.2d 548 (1972), that an at-will employee has no liberty or property interest in continued employment, and it is clear that D.C. law creates no such interest.
See
D.C.Code § 4-104 (permitting the Mayor to return Deputy Chiefs to the ranks of captain “when the Mayor so determines”). Nor may O’Donnell sue purely on the basis of the stigma associated with being fired; the Court found in
Paul v. Davis,
424 U.S. 693, 96 S.Ct. 1155, 47 L.Ed.2d 405 (1976), that stigma alone is not actionable, without a showing that a “right or status previously recognized by state law” has been “distinctly altered or extinguished.”
Id.
at 711, 96 S.Ct. 1155.
The
Roth
Court did cite two circumstances in which a governmental employment decision might be actionable. First, it observed that, although the plaintiff could not sue for the nonrenewal of his contract alone, “[t]he State ... did not make any charge
against [the emplbyee] that might seriously damage his standing and associations in the community,” and that “[i]t did not base the nonrenewal of his contract on a charge, for example, that he had been guilty of dishonesty, or immorality.”
Roth,
408 U.S. at 573, 92 S.Ct. 2701. Second, the Court noted that “there is no suggestion that the State, in declining to reemploy the respondent, Imposed on him a stigma or other disability that foreclosed his freedom to take advantage of other employment opportunities,” and that it did not, for instance, “invoke any regulations to bar the respondent from all other public employment in state universities.”
Id.
O’Donnell argues that his claim fits within each of these two exceptions.
1.
Reputation-Phis
The first category of claim, in which the plaintiff points to the conjunction of official defamation and adverse employment action, is usually termed a “reputation-plus” claim.
Paul
explained this passage of
Roth
as meaning that defamation alone is not actionable under the due process clause, but that defamation “in the course of the termination of employment” is. 424 U.S. at 710, 96 S.Ct. 1155. This circuit has since said that a demotion may also suffice to trigger a due process claim. “For a defamation to give rise to á right to procedural due process, it is necessary — we need not say when it is sufficient — that the defamation be accompanied by a discharge from government employment
or at least a demotion in rank and pay.” Mosrie v. Barry,
718 F.2d 1151, 1161 (D.C.Cir.1983) (emphasis added). Although the conceptual basis for reputation-plus claims is not fully clear, it presumably rests on the fact that official criticism will carry much more weight if the person criticized is at the same time demoted or' fired. Requiring a demotion or firing to trigger a defamation claim also helps to limit the scope of permissible due process claims to a small set of truly serious claims, thus limiting the con-stitutionalization of tort law.
Although O’Donnell' can point to isolated defamatory statements and to a demotion, he cannot demonstrate that the two occurred' together; only defamation that-is
“accompanied
by a discharge from government employment or at least a demotion in rank and pay” is actionable.
Mosrie,
718 F.2d at 1161 (emphasis added). First, O’Donnell’s transfer to the Property Division did not involve any loss of rank. A lateral transfer, even to a job that is regarded as a “dumping ground,” cannot form the basis of a reputation-plus claim.
Mosrie,
718 ,F.2d at 1156, 1161-62. O’Donnell also cites Soulsby’s criticism of him in'the Ballou High School speech; but O’Donnell’s demotion was clearly triggered by his letter to the editor, and had no obvious link, temporal or logical, to the speech. Finally, on the record as it stands, there is no sign that the District made any defamatory statements about O’Donnell at the time of his demotion.
See Orange v. District of Columbia,
59 F.3d 1267, 1274 (D.C.Cir.1995) (concluding that a firing without any associated public statements is not actionable).
2.
Stigma or Disability
The second
Roth
exception turns on the combination of an adverse employment action and “a stigma or other disability that foreclosed [the plaintiffs] freedom to take advantage of other employment opportunities,”
Roth,
408 U.S. at 573, 92 S.Ct. 2701; it differs. from the first in that it does not depend on official speech, but on a continuing stigma or disability arising from official action. We applied this exception in
Kartseva v. Department of State,
37 F.3d 1524 (D.C.Cir.1994). In that case, the plaintiff was a Russian translator who had been denied a security clearance by the State Department, and therefore fired by the consulting company that employed her. We distinguished two ways in which this 'denial could effect a “stigma or other disability” within the meaning of
Roth.
First, it might
have “formally or automatically exelude[d] Kartseva from work on some category of future State contracts or from other government employment opportunities.”
Id.
at 1528. Second, it might, without this “binding effect,” still have “the broad effect of largely precluding Kartseva from pursuing her chosen career as a Russian translator.”
Id.
(emphasis omitted). O’Donnell claims that the various measures that the District took against him effectively precluded him from pursuing his profession, and so bring him within the second prong of
Kartseva.
The Constitution protects an individual’s “right to follow a chosen trade or profession” without governmental interference.
Kartseva,
37 F.3d at 1529 (quoting
Cafeteria Workers v. McElroy,
367 U.S. 886, 895-96, 81 S.Ct. 1743, 6 L.Ed.2d 1230 (1961)). Government action that has the effect of “seriously affect[ing], if not destroy[ing]” a plaintiffs ability to pursue his chosen profession,
Kartseva,
37 F.3d at 1529 (quoting
Greene v. McElroy,
360 U.S. 474, 492, 79 S.Ct. 1400, 3 L.Ed.2d 1377 (1959)), or “substantially reducing] the value of his human capital,”
Taylor v. Resolution Trust Corp.,
56 F.3d 1497, 1506 (D.C.Cir.1995), thus infringes a liberty interest. There are, however, some limitations on this principle. As the Court made clear in
Siegert v. Gilley,
500 U.S. 226, 111 S.Ct. 1789, 114 L.Ed.2d 277 (1991), a showing of reputational harm alone cannot suffice to demonstrate that a liberty interest has been infringed. Thus, a plaintiff who (like O’Donnell) seeks to make out a claim of interference with the right to follow a chosen trade or profession that is based exclusively on reputational harm must show that the harm occurred in conjunction with, or flowed from, some tangible change in status. One such change in status, of course, is an adverse employment action.
O’Donnell claims that after his demotion he was unable to find employment in his chosen field, as a chief of police in a major city or municipality. After his complaint was filed, he became police chief in Brunswick, Maryland, whose population is 6,000. (Because this fact did not appear in O’Donnell’s complaint, the district court could not consider it in ruling on the District’s motion to dismiss; it therefore granted summary judgment as to Count III of O’Donnell’s complaint.)
O’Donnell may well have suffered some stigma in the job market as a result of his demotion. We conclude, however, that the evidence in this case falls short of that required to make out a claim under
Roth
and
Kartseva.
To begin with, the stigma associated with O’Donnell’s demotion was less severe than that attached to Kartseva’s disqualification. Kartseva was disqualified, for unknown but potentially serious reasons, from an uncertain and perhaps broad range of government employment; we found this to be potentially sufficiently stigmatic to render her unemployable in much of her field. By contrast, the reason for O’Donnell’s demotion was fairly clear, and the stigma itself less severe. Although it is difficult to put a good face on the denial of a security clearance, it may be that some employers would welcome a police official who (as they might see it) places the public’s interest above bureaucratic pressures.
Furthermore, O’Donnell’s new job demonstrates that the stigma he suffered cannot have been too disabling. Although he has moved from a large city to a town (Washington’s population is over 500,000, as against Brunswick’s population of 6,000), he has moved up in the ranks; he is now chief of police. O’Donnell’s overall responsibilities have certainly dwindled (for instance, he supervises far fewer officers), and a still greater diminution in his employment opportunities might well have sufficed to show that the value of his “human capital,”
Taylor,
56 F.3d at 1506, has dropped enough to infringe his liberty interest in his chosen profession. But O’Donnell has presented no concrete evidence (and the circumstances do not demonstrate) that he would have done dramatically better on the job market in the absence of his demotion. Certainly, we cannot be sure that he could have obtained, as was his stated wish, a job as chief of police in a “major city or municipality.” Even if the demotion did set him back a step on his career path (which may or may not be true), that would fall markedly short of showing that his ability to pursue his chosen profession has been
“seriously affected, if not destroyed.”
Kartseva,
37 F.3d at 1529 (quoting
Greene,
360 U.S. at 492, 79 S.Ct. 1400). Summary judgment as to this claim was therefore appropriate.
C.
Qualified Immunity
The District argued below, and argues again here, that Soulsby, the only individual defendant (Barry was sued only in his official capacity), is entitled to qualified immunity. The district court did not reach this question; we do, and we agree with the District. An official is entitled to qualified immunity unless she has violated a clearly established right. “The contours of the right must be sufficiently clear that a reasonable officer would understand that what he is doing violates that right. This is not to say that an official action is protected by qualified immunity unless the very action in question has been held unlawful ... but it is to say that in the light of preexisting law the unlawfulness must be apparent.”
Anderson v. Creighton,
483 U.S. 635, 639-40, 107 S.Ct. 3034, 97 L.Ed.2d 523 (1987). As our analysis of the status of O’Donnell’s letter demonstrates, it is by no means clear whether that letter is legally protected, so that any unlawfulness in Soulsby’s acts in (allegedly) securing O’Donnell’s demotion was not “apparent.” Soulsby is thus entitled to qualified immunity-
ill. Conclusion
In summary, we affirm the district court’s dismissal of O’Donnell’s First Amendment claims, except for its dismissal of his claim based on the letter to the editor. As to this claim, we remand to permit the district court to determine, on the basis of a more complete record, what the significance of O’Donnell’s references to the “credibility” of Hennessy and Soulsby would have been to an informed member of the public at the time O’Donnell wrote his letter, and in turn to decide to what degree the letter touched on a topic of public concern and hence weighed against the government’s interest in O’Donnell’s silence. We affirm the district court’s disposition (by dismissal and summary judgment) of O’Donnell’s due process claims. Finally, we find that Soulsby is entitled to qualified immunity in the proceedings on remand.
So ordered.