This appeal tests whether an elected county prosecutor must retain the at-will confidential secretaries hired by the predecessor he defeated, who supported the predecessor politically.

Facts

The firings of Susan Hobler and Linda Southwell, at-will secretaries in the Adams County prosecutor’s office, occurred after a countywide election that replaced then-prosecutor David Sandhaus. Adams County is a lightly populated rural county in southeastern Washington. The county seat is in Ritzville, a little town southwest of Spokane, southeast of Wenatchie, and northeast of Kennewick. The county’s elected prosecutor works mainly in the Ritzville office, but has a branch office in Othello, an even smaller town to the southeast.

After Sandhaus was elected as the Adams County prosecutor, he hired Ms. Hobler and Ms. Southwell as support staff. The county prosecutor prosecutes criminal cases, collects child support, and advises the County Commissioners on legal matters and policy. In addition to the prosecutor and whatever assistants or deputies he had in the Ritzville office, that office ordinarily has had two support staff positions. The Othello office was devoted exclusively to collecting child support, and for a lengthy period had no resident attorney. Two non-attorney staff persons worked at the Othello office as well.

Sandhaus, the outgoing prosecutor, was evidently controversial. Prior to his electoral defeat, there had been a recall petition filed against him. The record establishes that relations were bad between his office and the county commissioners. Linda Southwell testified that “there were two schools of power, basically, that David Sandhaus was one and that the commissioners were the other. And there was frequent head butting going on there. And, as an employee in the office, you had to be walking rather gingerly.” The defendant in this case, Gary Brueher, defeated Sandhaus in the general election by a very wide margin, 71.3% to 28.3%. Evidently viewing this electoral victory as a mandate for change in the prosecutor’s office, Brueher came into office pledging a “new team.” Brueher fired Hobler and Southwell shortly after taking office.

We take the facts about the plaintiffs’ work responsibilities from their own testimony at their depositions, since this is review of a summary judgment granted in favor of the defendant.

At the time of their firings, the plaintiffs had risen to be Sandhaus’s right hand men, as it were, in their respective offices. Both had always been at-will employees. Hobler worked at the county seat office in Ritzville and Southwell worked at the smaller office in Othello.

Sandhaus used Hobler to sit in on interviews when he hired prosecutors and support staff, to advise him on whom to hire, and to give him confidential notes of what was said at the meetings she attended. He also used her as a witness to sit in on sensitive conversations that might need one, such as discussions with employees about personnel problems. She reported to Sandhaus confidentially on “performance issues” relating to the office’s attorneys. For example, when Sandhaus was out of the office and a deputy violated an office policy by filing a complaint based on what a policeman told her, rather than waiting for the written police report, Ho-bler reported the errant deputy to San-dhaus (to the deputy’s considerable annoyance).

As the administrator of the office, Ho-bler did payroll, so she knew how much money everyone made. Also, she worked with the county auditor on obtaining and reviewing expenditures for the prosecutor’s budget. She characterized herself as a “liaison individual between the elected official [Sandhaus] and the balance of the populace.” She regularly spoke with the people at the courthouse, the county commissioners, and the sheriff. Thus, Hobler often acted on behalf of Sandhaus with respect to the county’s other important officials. In addition, Sandhaus would frequently call Hobler into his office “to impart the day to day on goings in the office,” acting as his “eyes and ears,” as many confidential personal secretaries do.

Linda Southwell administered the Othello office when Sandhaus wasn’t there. He only came there once every couple of months, so Southwell ran the office for him independently and without supervision much of the time. The office did nothing but child support collection, no criminal prosecutions, so the work could be done largely by administrative staff. For a substantial period of time there was no resident attorney.

After Sandhaus lost his election, Linda Southwell was concerned about her office’s child support enforcement budget, and Sandhaus told her that he couldn’t do much about it because he was leaving office. So Southwell herself worked with Sandhaus’s deputy and developed “a plan to continue funding for this support enforcement program, which was about to come to an end.” The plan laid out “what moneys we needed, how they were to be used, how they were to be divided up.” She also made the plan for the “travel arrangements, motel arrangements, reimbursement for flying, etc.” for whenever prosecutors came to the Othello office. Southwell had “overall responsibility for the functioning of the office.” She said that “[i]f someone didn’t show up for work, I had to call [Sandhaus]. If something unusual happened in the office that might become something he would be responsible for, I would report to him.” Sandhaus used her to monitor the performance of the staff and the other people in the office and be sure the work was performed. She also managed the office’s purchasing.

There may be a genuine issue of fact as to whether Brueher actually fired Hobler and Southwell because they supported Sandhaus against him in the county election. The way he put it was that, regardless of their competence, he didn’t feel he could trust them. In his deposition testimony Brueher stated he thought that “if I would have kept Sue Hobler on, that the attorneys would have left and the office would have been in complete disarray.” This fits together with what Ms. Hobler testified to, about reporting to Sandhaus on attorney failures to follow policy, and how upset an attorney became when Ho-bler reported on her. Brueher’s testimony shows the firings may have been motivated by a concern about the interaction between the prosecutor’s office and the courthouse. He said, “Sue wasn’t a real popular person at the courthouse ... the whole courthouse was against her.” As for Ms. South-well in the Othello office, Brueher stated that “I needed somebody in Othello that could operate independently and that I could trust. And because of her involvement in that e-mail incident and other things I just didn’t think I could trust her.” Brueher testified that “I was elected by a landslide and that the people wanted change and I wanted to change and get the office going in the right direction.”

But there was also evidence from which a jury could conclude that Mr. Brueher fired Ms. Hobler and Ms. Southwell simply because they supported Mr. Sandhaus politically in his election campaign. Because there may be an issue of fact as to the motivation for the dismissals, we take the facts most favorable to the plaintiffs for purposes of summary judgment. Thus, we presume for the purposes of this appeal that Hobler and Southwell met their burden of establishing that Brueher fired both of them for expressing their political support for his election rival.

After their termination, Hobler and Southwell brought this 42 U.S.C. § 1983 action against Adams County in Spokane County Superior Court, claiming that their dismissals violated their First Amendment rights. The county removed the case to the United States District Court for the Eastern District of Washington pursuant to 28 U.S.C. § 1441 and on the basis of federal question jurisdiction. In federal court Hobler and Southwell amended their complaint, dropping their claim against the county, and substituting a claim against Gary Brueher individually. Brueher moved for summary judgment. The district court granted summary judgment, characterizing Hobler and Southwell as confidential secretaries in very small offices, whom the prosecutor had to be able to replace. The district court went on to say that Brueher “probably would have been derelict had he not terminated their employment” and that he could do so under Branti v. Finkel. We have appellate jurisdiction over this matter pursuant to the district court’s final judgment on Brueher’s motion for summary judgment.

Analysis

In Elrod v. Bums a new Democratic sheriff fired all the Republicans in the sheriffs office who weren’t protected by civil service rules. The Supreme Court held that wholesale patronage dismissals going beyond policymaking and confidential positions encroached on employees’ First Amendment rights. The plurality opinion in Elrod was clarified by a majority opinion in Branti. In Branti, an incoming public defender sought to fire all the Republican assistant public defenders because they were members of the wrong party. The Supreme Court held that firing all the Republicans was unconstitutional, because the First Amendment protection against government retaliation for speech applied as well to political beliefs and affiliation.

But the Court took note of an exception to this First Amendment protection, where the “position is one in which political affiliation is a legitimate matter to be considered.” For example, a state university football coach couldn’t be fired for political affiliation, but the “various assistants” whom a governor uses to “write speeches, explain his views to the press, or communicate with the legislature” could have their jobs conditioned on sharing the governor’s political beliefs and party commitments. This is often characterized as the “policymaker” and “confidential employee” exception. But the Court described the proper inquiry in Branti as “not whether the label ‘policymaker’ or ‘confidential’ fits a particular position; rather, the question is whether the hiring authority can demonstrate that party affiliation is an appropriate requirement for the effective performance of the public office involved.” The Branti rule applies not merely to party affiliation, but more broadly to political beliefs, expression, and support. Therefore, although the partisan affiliations of Sandhaus, Brueher, Hobler and Southwell have no bearing on this case, their political activities fall within Branti’s (and its progeny’s) consideration.

We have repeatedly considered patronage dismissal for policymakers and whether certain classes of employees were policymakers for Branti purposes. For example, in Fazio v. City and County of San Francisco, we considered a district attorney’s decision to fire a high-ranking assistant because the assistant ran against him. We affirmed summary judgment against the fired assistant, holding that “an employer may fire a public employee for purely political reasons if the employer can demonstrate that political considerations are ‘appropriate requirement[s] for the effective performance of the job.’ ” Our holding recognized that some positions must be subject to patronage dismissal for the sake of effective governance and implementation of policy. In so doing, we cited a decision of the Seventh Circuit with approval, saying that “[a] public agency would be unmanageable if its head had to ... retain his political enemies ... in positions of confidence or positions in which they would be ... exercising discretion in the implementation of policy.” Thus, we held that Assistant District Attorney Fazio was a “policymaker” for purposes of the Branti exception.

Similarly, we held in Biggs v. Best, Best & Krieger that “an employee’s status as a policymaking or confidential employee would be dispositive of any First Amendment retaliation claim.” Biggs held that the “policymaker” exception was broader than “one who makes policy,” and held further that if an employee falls within the Branti exception, that is the end of the analysis. Thus, where the Branti exception applies the employee can be fired “for purely political reasons” without any Pickering balancing.

We carefully considered the mode and standard of review for political dismissal cases in Walker v. City of Lakewood. In Walker we reiterated that “ ‘an employee’s status as a policymaking or confidential employee [is] dispositive of any First Amendment retaliation claim.’ ” Likewise, Walker stated that we only consider whether Pickering protects the employee’s speech, activities, affiliations or beliefs in cases where the employee is not in a poli-cymaking or confidential position. We held in Walker that the question whether the employee held a policymaking or confidential position is “a mixed question of fact and law.” This is because “[dietermin-ing the particular duties of a position is a factual question, while determining whether those duties ultimately make that position a policymaking or confidential question is a question of law.”

Walker’s important holding means that the question whether the Branti exception applies is properly determined by summary judgment or occasionally a motion to dismiss rather than by trial, at least where the duties of the position, insofar as they are material, are not genuinely at issue. Insofar as the material “duties of a position” are not at issue, the question of “whether these duties ultimately make that position a policymaking or confidential [position] cannot properly be submitted to a jury because it is a question of law.” In the case at bar, there is no genuine issue of fact as to the duties of Ms. Hobler and Ms. Southwell, insofar as they are material, so the district court properly did not submit the question of whether they fell within the Branti exception to a jury. “As with all mixed questions, we conduct a de novo review.”

Our patronage dismissal cases have involved the “policymaker” branch of the Branti exception rather than the “confidential employee” branch, e.g. Fazio, Walker, and Diruzza. This case, though, has been briefed entirely on the question whether Ms. Hobler and Ms. Southwell were “confidential employees.”

As with “policymakers,” the question cannot properly be answered by determining whether their titles fit into the pigeonhole, because neither “confidential employees” nor “policymakers” is, under Branti, a pigeonhole at all. The question for us on appeal is “whether the hiring authority can demonstrate that [politics] is an appropriate requirement for the effective performance of the public office involved.” Thus, the question is not whether Ms. Hobler or Ms. Southwell held the title “confidential secretary” or could properly be denominated as such, but whether the work they did made their support for the outgoing prosecutor against the incoming prosecutor “an appropriate requirement for the effective performance of the public office involved.”

This clears out some underbrush from both sides’ arguments. It isn’t determinative that neither held the job title of “confidential secretary.” Nor could a public employer legalize wholesale patronage dismissals simply by titling all the support staff as “confidential staff.” Likewise, it isn’t determinative that Southwell signed a “confidentiality agreement” or that both had access to highly confidential information, such as arose in paternity cases in Othello or regarding who was about to be indicted in Ritzville. Many public employees have access to and work with confidential information yet are shielded from wholesale patronage dismissals under Branti. As Branti stated regarding mere access to confidential information, “although an assistant [public defender] is bound to obtain access to confidential information arising out of various attorney-client relationships, that information has no bearing whatsoever on partisan political concerns.”

What matters is not any of these sorts of bright-line rules. What does matter is whether Ms. Hobler’s and Ms. Southwell’s actual duties in the Ritzville and Othello offices and their relationship to the elected official made them “confidential employees” in the Branti sense that their political conduct was “an appropriate requirement for the effective performance of the public office involved.”

Plainly, it was. Most offices have certain key personnel who aren’t policymakers in the Branti sense but who are critical to effective policy implementation, and whose loyalty and confidentiality are nec essary. It is hard to run any sort of office without certain employees who work so closely with the outgoing boss that any incoming boss must have the option of picking his or her own people for that position. The outgoing Adams County prosecutor, Sandhaus, did have just such key personnel in each of the two offices, Ritzville and Othello, who functioned as his conduit for the most sensitive information. Sandhaus depended on Hobler and South-well to tell him when other staff, including lawyers, weren’t doing their jobs the way he wanted them done. Sandhaus’s replacement, Brueher, quite reasonably felt that he couldn’t patch things up with the staff attorneys and take effective control of the office himself if Hobler and Southwell were still in between.

Moreover, Hobler and Southwell functioned as Sandhaus’s communications conduit to the public and other elected officials like the governor’s assistants in the Branti example. Because the county commissioners and the people at the courthouse were upset with the prosecutor’s office, Brueher could not carry out his program of making peace with them if Hobler and Southwell remained. Brueher needed precisely the same type of relationship with his key employees that Sandhaus had with Hobler and Southwell in order to effectively implement his policies. Requiring Brueher to keep on persons that his predecessor and political enemy worked with so closely would simply have stymied him.

The same language we quoted from the Seventh Circuit in Fazio relating to “policymakers” applies in this context to “confidential employees.” The Seventh Circuit noted, as we agreed, that “[a] public agency would be unmanageable if its head had to ... retain his political enemies ... in positions of confidence or positions in which they would be ... exercising discretion in the implementation of policy.” Without a confidential secretary the official can trust to carry out his views, funnel communications in and out according to his priorities, and represent him in a way that enhances rather than damages his reputation, an elected official cannot effectively perform his office. Thus, in the case at bar political loyalty is “an appropriate requirement for the effective performance of the public office involved.” It is striking that in this case, Hobler so much represented the prosecutor’s office that the personnel at the courthouse were upset specifically with her, not just with Sandhaus, and likewise the attorney staff in the prosecutor’s office. Similarly, it is significant that Southwell, not Sandhaus, took the initiative on working out the Othello office’s budget with the county auditor, with the great influence that budget planning always has on the circumstances of those who work in a government office.

The general rule in our sister circuits, which we adopt, is that a confidential secretary to a policymaker may, consistent with the First Amendment, be replaced by the policymaker’s successor for political reasons. The circuits vary in such details of application as whether it is a mixed question of law and fact and whether applicability of the Branti exception completely obviates the need to perform a Pickering analysis. But the general rule, that a confidential secretary, though not a policymaker, may fall within the Branti exception and may be dismissed for political reasons, appears to pertain everywhere.

For instance, the First Circuit held in Vazquez Rios that the executive secretary and the cultural attaché of a governor were confidential employees, while janitorial staff were not. They rightly recognized that “[i]t would strain credulity to read the First Amendment or Elrod to require an elected official to work in constant direct contact with a person viewed as a political enemy.” The First Circuit test for determining who is a confidential employee in this sense is “public employees who occupy positions of such unusually intimate propinquity relative to government leaders that, despite their nonin-volvement with partisanship and policy-making, political loyalty could be deemed an appropriate requirement of the job.” In the case at bar, the duties of Hobler and Southwell and the required working relationship with Brueher was of the necessary propinquity to make those positions fall under the Branti exception for confidential employees.

Similarly, the Sixth Circuit has held that “a secretary to a policymaker may be fired for political reasons,” as these positions “involve!/] access to confidential and political material, and political loyalty, whether partisan or personal, is an essential attribute of the job.” The Tenth Circuit has also affirmed the proposition that “political association and allegiance were appropriate requirements for the performance of[the employee’s] job as administrative assistant to the city manager.” The Seventh Circuit, in recognizing the principle that a secretary to an elected official often falls under Branti’s exception, explained that “[i]f Rosalynn Carter had been President Carter’s secretary, President Reagan would not have had to keep her on as his secretary.” Finally, although presented in a different procedural posture, the Second Circuit has held that considering political affiliation for a confidential secretary position is appropriate “when there is a rational connection between shared ideology and job performance.”

It is significant that in Elrod, the dismissal of all Republicans was in an office of 3,000 people, where there could be no intimate relationship of trust between all the dismissed employees and the person who dismissed them. And in Branti, the dismissal of all Republican assistants was in a Rockland County, New York public defender’s office with nine assistants. Because the assistant public defenders’ re sponsibilities were primarily or solely to defend individuals charged with crimes, “it would undermine, rather than promote, the effective performance of an assistant public defender’s office to make his tenure dependent on his allegiance to the dominant political party.” But trust and loyalty are factors in the relationship between a policymaker and confidential secretary necessary to promote the effective implementation of policy.

Several factors make the case at bar a particularly strong one for applying the “confidential employee” branch of Branti The offices were very small, so the elected official, Brueher, would have a continual and close working relationship with his support staff. The dismissals were retail, of two specific individuals whose personal activities in the office mattered a great deal to the incoming official, as opposed to the wholesale terminations in Branti and Elrod of all persons not of the right political party. Both Hobler and Southwell continually interacted on behalf of the prosecutor with the courthouse personnel who affected how successful he was and the elected county commissioners who determined his budget and who looked to him for legal advice. The confidential secretaries were the public face of the individual who fired them, not, as in Branti, the representatives of clients other than that official.

That Southwell worked in the Othello office and saw the prosecutor personally only once every couple of months does not militate against her “confidential employee” status. If anything, it magnified it. Brueher would have had to rely on her even more than if he’d been there, to tell him what was going on, to run the office, to represent him to the public and office holders in Othello, to put together budget information, and to keep him informed on personnel matters. His knowledge of what was going on was limited to what she perceived and told him. Physical propinquity is not required to make political loyalty “an appropriate requirement for the effective performance of the public office involved.”

We refuse to convert the Branti test— whether political affiliation is an appropriate requirement for the effective performance of the public office involved — into something else. An alternative test would invite the pigeonholing that the Court held was impermissible, and a multifactor test would necessarily list so many factors as to be no more determinative of the outcome than the Branti formulation. The test for whether someone is a “confidential employee” in the Branti sense is simply whether political loyalty, or the absence of political adverseness, is “an appropriate requirement for the effective performance of the public office involved.” This case suggests a few factors that bear on the question, but neither a multifactor “test” nor an exhaustive and exclusive list of factors is appropriate. Among the factors that suggest themselves in this case are: (I) how closely does the person work with the official? (2) does the person’s job require personal loyalty to the official? (3) is the office so small that the relationship is necessarily close, or so large that it isn’t? (4) does the official rely on the person for information about delicate matters within the office or communications with the public or other officials on behalf of the official? (5) would the official’s ability to manage relationships with office staff or persons with whom the office deals be impaired if the persons are politically loyal to an adversary or not loyal to him? (6) were the dismissals of only one or a small number of employees who worked most closely with the policymaker, or were they wholesale dismissals? (7) do the individu als speak to other employees, the public and to other policymakers on behalf of the official? In other cases, other factors may enter the balance.

Plaintiffs also brought a claim under state law, for the tort of wrongful discharge in violation of public policy. They concede that in this case, the only public policy at issue is the First Amendment, so we affirm the summary judgment on the state tort claim on the same basis as on the section 1983 claim as no violation of Hobler’s or Southwell’s First Amendment rights occurred.

Conclusion

Because plaintiffs were “confidential employees” in the Bmnti sense, the First Amendment did not protect them from dismissal because of their political loyalty to the defendant’s political adversary, so the summary judgment for defendant is

AFFIRMED.

. 445 U.S. 507 , 100 S.Ct. 1287 , 63 L.Ed.2d 574 (1980). . 28 U.S.C. § 1291 . . 427 U.S. 347 , 96 S.Ct. 2673 , 49 L.Ed.2d 547 (1976). . Id. at 357 , 96 S.Ct. 2673 . . Branti, 445 U.S. at 518-20 , 100 S.Ct. 1287 . . Id. at 519-20 , 100 S.Ct. 1287 . . Id. at 518 , 100 S.Ct. 1287 . . Id. . Id. . See, e.g., Savage v. Gorski, 850 F.2d 64 (2d Cir.1988); Faughender v. City of North Olmsted, 927 F.2d 909 (6th Cir.1991); Soderbeck v. Burnett Cty., 752 F.2d 285 (7th Cir.1985); DiRuzza v. County of Tehama, 206 F.3d 1304, 1310 (9th Cir.2000); Fazio v. City and County of San Francisco, 125 F.3d 1328 (9th Cir.1997). . 125 F.3d 1328 (9th Cir.1997). . Id. at 1332 (alteration in original) (quoting Branti, 445 U.S. at 518 , 100 S.Ct. 1287 ). . Id. at 1333 (alteration in original) (citing Wilbur v. Mahan, 3 F.3d 214, 217 (7th Cir.1993)). . Id. at 1334. . 189 F.3d 989 (9th Cir.1999). . Id. at 994-95. . Id. at 995 (quoting Fazio, 125 F.3d at 1331 (holding that Pickering balancing is not reached if the Branti exception applies)). But see Barker v. City of Del City, 215 F.3d 1134, 1139 (10th Cir.2000) (holding that an employee's status as a policymaker under the Branti exception does not obviate the Pickering analysis). . Pickering v. Board of Ed. of Tp. High School Dist. 205, Will County, Illinois, 391 U.S. 563 , 88 S.Ct. 1731 , 20 L.Ed.2d 811 (1968). Under Pickering , we generally balance "the interest of the [employee], as a citizen, in commenting upon matters of public concern and the interest of the State, as an employer, in promoting the efficiency of the public services it performs through its employees." Id. at 568 , 88 S.Ct. 1731 ; Fazio, 125 F.3d at 1331 . . 272 F.3d 1114 , 1131-32 (9th Cir.2001). . Id. at 1131 (quoting Biggs, 189 F.3d at 994-95). . Id. at 1132. . Id. at 1131. . Id. at 1132. . Id. . Id. (citing United States v. City of Spokane, 918 F.2d 84, 86 (9th Cir.1990)). . Branti, 445 U.S. at 518 , 100 S.Ct. 1287 . .Id. at 519, 100 S.Ct. 1287 ; see also Vazquez Rios v. Hernandez Colon, 819 F.2d 319, 324-26 (1st Cir.1987) (holding that access to confidential information didn’t transform janitorial staff into confidential employees under Branti). . Fazio, 125 F.3d at 1333 (alteration in original) (citing Wilbur, 3 F.3d at 217 ). See also Weisbuch v. County of Los Angeles, 119 F.3d 778, 784 (9th Cir.1997) (“A district attorney whose priority is violent crimes may need an intake supervisor who agrees, not one who believes and gives speeches to the effect that the number one priority should be narcotics cases. The alternative would be a policy carried out inefficiently or not at all, and constant departmental strife as the [employee] was perceived to be spying on or micromanaging the renegade director’s department to see whether it was carrying out official policies.”). . Compare Walker, 272 F.3d at 1131, recognizing an existing circuit split on the issue and applying the mixed question of law and fact analysis, with Soderbeck, 752 F.2d at 288. . Compare Fazio, 125 F.3d at 1331 , with Barker, 215 F.3d at 1139 . . Vazquez-Rios, 819 F.2d at 323 . . Id. at 324 . . Faughender, 927 F.2d at 914 (citation omitted). . Barker, 215 F.3d at 1138 . The Tenth Circuit, in contrast to our circuit, holds that the Branti exception does not obviate Pickering balancing. Id. at 1139. See also supra at n. 16. . Soderbeck, 752 F.2d at 288. The Seventh Circuit differs only in that they do not follow Walker on the mixed question of law and fact issue. . Savage, 850 F.2d at 66-70 (reversing the district court’s grant of a preliminary injunction in favor of employees including several confidential secretaries). . Branti, 445 U.S. at 519-20 , 100 S.Ct. 1287 .