In this appeal of an employment discrimination, disparate treatment case, Defendant-Appellant Anthony M. Frank, the Postmaster General (the Service), asserts that the district court erroneously found that Plaintiff-Appellee Elzie D. Odom was discriminated against on the bases of race and age in his bid for a promotion.1 Finding that the district court clearly erred in the factual determinations on which its conclusion of discrimination was founded, we reverse and render.

* FACTS AND PROCEDURAL HISTORY

Odom began working for the Service in October 1950. He served as a postal inspector from 1967 until his retirement in 1987. In November 1983, he sought but did not receive a promotion to "level 24 Prevention Team Leader," a position which had recently been created in his home division of Fort Worth. At the time he applied, Odom was fifty-four years old and thus within the class of persons protected by the Age Discrimination in Employment Act.2 He is also African American and thus protected against racial discrimination as well.

Postal inspectors generally serve in one of two capacities: 1) auditors, or 2) criminal violations specialists. For the most part, Odom served as an auditor throughout his tenure as a postal inspector. He did, however, perform miscellaneous criminal assignments from time to time between 1968 and 1971. He also served on a "security assignment," which was a quasi-criminal violations position, for a time during the period 1981-1982.

Unlike most of his prior experience, the position of Prevention Team Leader which Odom sought in 1983 would involve almost entirely criminal work. The new leader would concentrate in four work areas, each projected to account for a specified percentage of his time: thirty percent in "external crimes"; thirty percent in "security"; twenty-five percent in "fraud"; and fifteen percent in "internal crimes." According to the announcement of the vacancy, the position would also require the applicant to possess "highly developed written and oral communication skills" and "well developed human relations skills."3

Odom submitted his application in September 1983. He was one of sixteen persons from several different divisions who applied for the position. Seven of the sixteen were forty years old or older; three of the sixteen were African American and one was Hispanic. Of the total number of applicants, seven including Odom were from the Fort Worth Division. Of those seven, two--Odom and Peay--were African American; a third--Herrera--was Hispanic; and the remaining four--Horton, Jennings, Nichols, and Price--were Caucasian.

The application, PS Form 991, comprised several sections, some of which were to be filled out by the applicant. When the applicant finished his sections, the form was to be forwarded to his immediate supervisor for completion of a recommendation section. When that was accomplished, the form was to be forwarded to the Inspector in Charge to fill in an additional recommendation section, thereby completing the form.

The Inspector in Charge of the Fort Worth Division was D.C. Strader, a native American. Three of the seven applicants from the Fort Worth Division--Herrera, Jennings, and Price--worked directly under Strader at the time, so none had an immediate supervisor other than Strader. Consequently, the applications for those three contained only one supervisory recommendation--Strader's.

As noted, sixteen persons applied for the subject position.4 After all application forms were complete, they were to be forwarded to the Southern Regional Office of the Postal Inspection Service in Memphis, Tennessee, for further processing.

Under Service regulations, completed applications for a position such as the one involved in the instant case are initially screened by a review panel. The members of the panel involved in the instant case were selected by Robert N. Moore, the Regional Chief Inspector for the Southern Region, and the ultimate decisionmaker for the subject position. The panel members were "required to be at or above the rank of the position [at] issue and to have, as a group, functional knowledge of the position[ ] [at] issue."5

The three persons selected to constitute the instant panel were Michael Gump (the designated Chairman), George Hicks, and Hubert Smith. All three were white males, "despite the fact that Postal Service Guidelines specified that 'every effort will be made to designate at least one woman or one minority group member to serve on each review committee.' "6 And, although it is apparently contrary to Service regulations for a supervisor of a worker to sit on a review panel considering that worker's application for a promotion such as the one involved here, Hicks was selected for the review panel despite being the supervisor of R. Hurlbut, one of the applicants who is white and, like Odom, is over forty years old.

On October 19, 1983, the review panel met to select applicants to be interviewed. At that time, all sixteen of the applications for the position had been completed, but inexplicably two applications from the Fort Worth office had failed to be forwarded to the panel. One of the two was Odom's.7 Unaware that two applications were missing, the panel selected ten of the applicants for interviews. Five of those ten were from Fort Worth.

After the review panel made its selection of the applicants to be interviewed, Chairman Gump received Odom's application.8 The panelists then conferred about Odom's application but decided not to add him to the group to be interviewed.9

When Odom learned that he had not been selected for an interview, he complained to Strader, asserting the belief that the decision not to interview him was racially based (no mention of age). After hearing Odom's allegations, Strader discussed the matter with Gump who again conferred with the other panelists. As a result of Strader's intercession, the panel decided to grant Odom an interview.

All interviews, including Odom's, were held in Memphis in October 1983. Each panelist independently rated the persons interviewed and came up with his own "top five" list. Those lists were then compared and discussed, ultimately producing the review panel's consensus "top five" list. Odom was on neither the consensus list nor on any panelist's top five list. "Each panel member's individual top five list included the names of Inspectors Herrera, Horton, Nichols and Price. There was initially some disagreement over whether to include Inspector Jennings or Inspector Hurlbut, but Inspector Hurlbut was ultimately agreed upon."10

The panel's top five list was then sent to Regional Chief Inspector Moore to make the actual selection. In addition to considering the applications, recommendations of supervisors, and the review panel's selections, Moore telephoned Strader and specifically asked for his "pick." In response Strader indicated candidly that, given a choice, he would select Price. "On November 4, 1983, Moore announced that Inspector Price had been selected for the ... position."11

After the decision on the promotion was announced, Odom filed a formal charge of age and racial discrimination with the Service. The Equal Opportunity Employment Commission (EEOC) conducted an administrative hearing on Odom's charge. The EEOC's hearing officer concluded that Odom had been discriminated against. The Service, however, disagreed with the EEOC examiner and found that Odom had not been discriminated against. After more than one hundred eighty days passed without any action being taken on his appeal from the Service's final agency decision, Odom filed the instant lawsuit.

The district court held a bench trial and concluded that Odom had been discriminated against on the bases of both age and race. The court stated that it was "persuaded by the totality of the circumstances surrounding the promotion process that the offered explanation for the denial of the [subject] position is a pretext for discriminatory actions."12 The court listed the facts which it found Odom to have established in demonstrating discrimination:

a. Plaintiff was clearly better qualified than the selectee for the Prevention Team Leader position in terms of his performance and experience;

b. The method of completion and submission of Plaintiff Odom's PS Form 991 is evidence of pre-selection due to the deliberate omission of the highlights and successes of Plaintiff's career with the agency, when similar highlights were detailed by the supervisor on the applications of other white applicants.

c. The review panel appointed to make the selection, which was made up solely of white males, was improperly constituted under the agency's own regulations. The inclusion of Mr. Hicks on the review panel in violation of internal regulations which prohibited supervisors of applicants from serving on the review panel is especially probative since Defendant excluded eligible and available minorities from serving on the review panel because they had supervised some of the applicants or came from the same division as some of the applicants.

d. Evidence was adduced that the agency had an unwritten policy discouraging promotion of persons over forty to upwardly mobile positions.

e. The statistical data introduced by Plaintiff indicates [sic] exclusion of blacks and persons over the age of forty as selectees for higher level positions within the Postal Inspection Service.13

Finding the Service liable for discrimination, the district court awarded Odom $8,707.04 in unpaid backpay, $8,796.15 in "unpaid annuity," and $6,212.68 in pre-judgment interest.14 The court also awarded Odom $33,646.50 in attorneys' fees.15 The Service timely appealed.

As noted above, this is an appeal from a bench trial. We review the factual findings of such a proceeding under the clearly erroneous standard of review.16 Issues of law are reviewed de novo.17

Observing that the district court expressly based its decision on the "totality of the circumstances surrounding the promotion process" to which Odom was exposed, we have exhaustively reviewed the evidence contained in the record that either supports or fails to support the findings of the district court. Left with the unmistakable impression that the district court's direct or implied findings of discrete facts were either not supported by sufficient evidence or simply wrong, we find that the court clearly erred in making its ultimate factual determination of age and racial discrimination. We shall discuss each key fact in turn.

1. Strader's Completion and the Submission of Odom's Application

a. Strader's Completion of the Forms

The district court found that the manner in which Odom's application was completed by Strader evidenced negative pre-selection "due to the deliberate omission of the highlights and successes" of Odom's career when "similar highlights were detailed by the supervisor on the applications of ... white applicants."18 The phrase "completion ... of Plaintiff Odom's PS Form 991" as used by the district court refers to Strader's recommendation, which was the last requirement for the completion of Odom's (and any other) application from the Fort Worth division. The crux of Odom's argument on this point, which the district court accepted, is that Strader provided more extensive and complimentary recommendations on the application forms of white employees than on those of African American employees, and thus discriminatorily "pre-screened" the positions, i.e., sent an implicit message to the panel as to whom Strader did and did not want the panel to consider seriously for the position.

The Service counters that, in crediting Odom's argument, the district court gave short shrift to Strader's legitimate, non-discriminatory explanation that it just happened that he was either the immediate supervisor of, or had a much closer working experience with, the white and Hispanic applicants. According to the Service, this, coupled with the fact that there was no other "immediate supervisor" to write recommendations for those applicants whom he directly supervised, explained why Strader's recommendations of the non-black applicants were more detailed than were those of Odom and the other black applicant, Peay.

Strader provided recommendations for the seven applicants from the Fort Worth Division. Three of the applicants--Herrera, Jennings, and Price--had no immediate supervisor other than Strader. Strader thus served a dual role for those three applicants, so his longer, more detailed, and more complimentary recommendations might thus be explained. Additionally, Strader testified that, even though he was not Nichols's immediate supervisor, he knew Nichols well and had worked closely with him in the past. Strader thus explains his more detailed and longer recommendation of Nichols as well.

The final non-African American that Strader recommended was Horton. Odom's attorney failed on her direct examination of Strader to adduce any evidence regarding either his professional or personal relationship with Horton. Nonetheless, one of her comments during direct examination set the tenor of things to come. At one point during this examination Strader commented: "My basic knowledge of Mr. Odom at the time was--I didn't have that much knowledge of him." Odom's attorney replied: "You mean to tell me you had contact with every other inspector who applied for this position but Mr. Odom?"

On cross examination, Strader testified that he gave longer and more detailed evaluations of some applicants because he had worked with them in the past and was more familiar with them. In the course of identifying those persons with whom he had worked in the past, Strader testified that he thought Horton too had reported directly to him for a period.

Q. So your testimony is that you're certain that Mr. Price reported directly to you prior to this promotion?

A. Yes.

Q. And Mr. Horton may also have to a certain extent; is that correct?

A. Yes.

Despite the somewhat equivocal nature of the Service's attorney's second question, Odom's attorney did not return to the subject of Strader's past experience with Horton when she conducted her redirect examination of Strader.

Although Strader's testimony about his prior direct supervision of Horton was less than absolute, it constituted at least some evidence. More significantly, it was uncontradicted. As Odom adduced no evidence favorable to his position regarding supervision of Horton, we are compelled to accept Strader's legitimate, non-discriminatory explanation that he was either the immediate supervisor of, or had a closer working experience with, each white or Hispanic applicant. Consequently, we hold that the district court clearly erred in finding that the manner in which Odom's application was completed by Strader evidenced negative pre-selection.

b. Late Submission of Odom's Application

As noted, Odom's application did not reach the panel until after the panel's initial screening had been completed. Inexplicably, that application was not transmitted with the original batch of five from Fort Worth. Like one other among the seven applications from Fort Worth, Odom's apparently was omitted from the package. Importantly, however, nothing in the record--and nothing urged by Odom--reflects any evidence of conscious or intentional delay, much less racial or age animus.

When Odom's application was received, the panelists conferred by telephone about its merits. Moreover, when Odom was not granted an interview, the panel reversed itself at Strader's behest and interviewed Odom. Yet from the palpable innuendo in the district court's opinion, we cannot help but infer that during the trial the court came to believe that some nexus existed between age and racial discrimination and the delayed submission of Odom's application. The unfairness of such an implication is demonstrated, however, in the court's statement early in its opinion:

Inspector Odom's application was one of the two that, for some unknown reason, either had not been received by the review panel with the rest of the Fort Worth Division applications or was not properly completed by Mr. Strader, despite the fact that Inspector Odom had submitted his application well in advance of the deadline.19

As the district court stated, there is no evidence concerning why Odom's application was not timely submitted. Neither is there any evidence that it "was not properly completed by Mr. Strader." The implication, by the court's repeated references to that late submission in context with other practices questioned by the court, that discriminatory animus against Odom produced the delivery glitch--is wholly baseless. Even Odom's counsel, at oral argument, would not claim that the late delivery was an intentional or deliberate act by Strader or anyone else. We cannot help but wonder at the court's failure to mention Strader's successful mitigating efforts on Odom's behalf to get the review panel to reconsider his application when it finally arrived in Memphis after the panel initially failed to grant him an interview--reconsideration that resulted in the panel's decision to reverse itself and grant Odom an interview. We find the district court's implied finding of discrimination in the isolated fact of late delivery of Odom's application to the review panel to be clearly erroneous.

2. Comparative Qualifications for the Position

The district court found that "Odom was clearly better qualified than Price for the Prevention Team Leader position in terms of his performance and experience."20 Generally, a court's belief that an unprotected applicant who has been promoted is less qualified than a protected applicant who has been passed over, will not in and of itself support a finding of pretext for discrimination. If, however, the passed over applicant who is protected against discrimination is clearly better qualified for the position in question, a finding of pretext masking discrimination can be supported by the promotion of the less qualified person.21 In considering the relative qualifications here, the district court placed those of Odom and Price side-by-side and found:

45. Inspector Odom was better qualified for the position than Inspector Price when comparing them according to the stated qualifications and attributes sought in the vacancy announcement and as stated by the review panel members. These qualifications and attributes included leadership abilities, oral and written communication skills, and technical knowledge or skills of the position.

46. Inspector Odom had previously led task forces, had been assigned the training of junior inspectors, and had previously served as a team leader. Additionally, Inspector Odom had several commendation letters in his personnel file for various work that he had performed in the criminal arena, some of which he had supervised.22

When we conduct a like comparison of the two applicants' qualifications, however, we are led to the conclusion that the district court clearly erred in finding that Odom was clearly (as distinguished from merely ) better qualified for the position than was Price. Rather, as readily conceded by the Service's attorney at oral argument, the record supports the conclusion that Odom and Price were similarly qualified for the position. We simply do not find support for the district court's conclusion that Odom's qualifications were so greatly or significantly superior to Price's to make Odom "clearly better qualified."

Price had significant recent experience in several of the criminal areas that were most relevant to the new position. As noted above, Odom's experience had not been primarily on the criminal side of the inspection service; to the contrary, his had been almost entirely on the audit side. Most of the work for which Odom had been primarily responsible during the several years preceding to the application process simply was not relevant to the new position.

Odom contends that the fact that he was a career audit specialist did not matter, and that the Service's assertion of the irrelevance of his specialty and his experience was merely a pretextual explanation. We disagree. The review panelists' testimony demonstrates what follows naturally to us: The fact that Odom's primary experience did not match the position sought was legitimately relevant and significant to the panel's determination.

Additionally, Price had a college degree while Odom did not.23 Based on raw numbers, Odom had participated in more instructional courses and programs than had Price; but that was to be expected, given the greater length of time that Odom had worked for the Service. Their respective statements of "specific qualifications" are quite different, but neither is particularly more impressive than the other. A careful and objective comparison of Price's and Odom's applications reveals no glaring distinction that would support a finding that Odom was "clearly better qualified than [Price] for the Prevention Team Leader position."24

We also remain cognizant of the fact that the evaluation of applicants (and applications) for high level positions in any discipline--business, industry, government, military, or education--involves both objective and subjective elements. We also recognize that subjectivity has a potentiality for abuse by those evaluators who would use it to shield improprieties in the selection process, possibly even as a pretext for discrimination. On the other hand, as a general rule judges are not as well suited by training or experience to evaluate qualifications for high level promotion in other disciplines as are those persons who have trained and worked for years in the field of endeavor for which the applicants under consideration are being evaluated.

Therefore, unless disparities in curricula vitae are so apparent as virtually to jump off the page and slap us in the face, we judges should be reluctant to substitute our views for those of the individuals charged with the evaluation duty by virtue of their own years of experience and expertise in the field in question. We cannot here disabuse ourselves of the conclusion that the district court clearly erred in substituting its comparative evaluation of the two candidates for that of the review panel to find that Odom's credentials were so obviously and substantially superior to Price's that Odom was "clearly better qualified" for the job than was Price. We find that neither singly nor collectively do Odom's qualifications leap from the record and cry out to all who would listen that he was vastly--or even clearly--more qualified for the subject job than was Price. The district court's finding to the contrary was clear error.

Without labeling it discriminatory, the district court expressly considered the improper make up of the Service's review panel as a key element in the "totality of the circumstances surrounding the promotion process" that led to the court's finding of discrimination against Odom. As we observed earlier, the court stated:

The review panel appointed to make the selection, which was made up solely of white males, was improperly constituted under the agency's own regulations. The inclusion of Mr. Hicks on the review panel in violation of internal regulations which prohibited supervisors of applicants from serving on the review panel is especially probative since Defendant excluded eligible and available minorities from serving on the review panel because they had supervised some of the applicants or came from the same division as some of the applicants.25

In making this determination, the district court rejected and did not discuss the only evidence in the record--testimony by one of Regional Chief Inspector Moore's assistants--concerning the composition of the review panel. Apparently ignoring the Service's witnesses, the court implicitly found that the composition of the panel evidenced a discriminatory intent toward Odom. We cannot accept either the district court's unexplained disregard of the facially benign explanation given by the Service or the inference of discrimination the court made from the fact that the panel was comprised of three white males in derogation of a provision in Service guidelines for the inclusion of at least one woman or minority on every panel.

The district court found the constitution of the review panel "especially probative," because minorities who were supervisors of applicants for positions and were of appropriate rank were not appointed to the panel while Hicks, a white male, was included on the panel even though an employee whom he supervised (Hurlbut) was under consideration by the panel. Yet if the district court even considered the uncontradicted testimony of Walker Liner, the Regional Personal Officer and one of Moore's assistants, that fact is not apparent from its opinion.

Liner testified that the fact that Hicks was the supervisor of one of the applicants did not come to anyone's attention when the recommendations for the review panel assignment were being made. By the time it was discovered that Hurlbut had applied and that he was supervised by Hicks, the panel was simply too far into the process for anything to be done about the problem. Moreover, Liner's uncontradicted testimony was that when on previous occasions panels had been constituted in technical violation of the guidelines, it was not uncommon for the panelists to serve nonetheless. His experience had been that it was difficult (if not impossible) to constitute every panel in a technically correct manner, and that the fact that it occurred here had nothing to do with Odom's case.

The district court also ignored, without comment, the reasonable, benign explanations for there having been no minority member or woman on the instant review panel. At the time the panel was appointed, there were no women supervisors at the appropriate level in the Southern region. As for potential minority participation on the review panel, Liner provided reasonable (and uncontradicted) explanations concerning why each potential qualified minorities supervisor had to be excluded.26

The district court found that the bending of the Service's guideline by allowing Hicks to serve on the panel while excluding potential minority panelist evidenced discrimination against Odom. As discussed, however, the record does not contain any evidence whatsoever of a nexus between the panel's make-up and age or racial discrimination towards Odom--certainly none sufficient to support a finding of discrimination on this point. The inclusion of Hicks on the panel was at most one of oversight regarding Hurlbut's application. Besides, the unavailability of qualified female or minority supervisors to serve on the panel made the appointment of some white male unavoidable. Further, the relevant service guideline does not impose an absolute duty to include a woman or minority group member on each review committee. Instead, the guideline dictates that "every effort will be made" to designate one such person to each review committee. This language implicitly recognizes that it will not always be possible to compose a review committee in accordance with this standard.

Odom's argument, and the reliance of the district court in this regard, encounter another problem: There is simply no evidence that the inclusion of Hicks on the panel had any producing causal connection with Odom's failure to make the panel's top five list. All three panel members were white males, so there is no basis for singling out Hick's appointment as evidence of race or age bias. That Hicks was a supervisor of one of the applicants who made the consensus top five list may show cronyism, but cannot be classified as reflecting race or age bias. After all, not one of the three panelists included Odom in his "top five" list. Moreover, the only difference in the panelists' individual top five lists was the inclusion of either Nichols or Hurlbut as the last of the sixteen applicants to make the consensus top five list. If anyone should be heard to complain about the composition of the panel, it could only be Nichols--not Odom. For even if Hicks' position on the panel had been filled by a non-white, non-male, non-supervisor who eventually included Odom on her top five list, he still would not have been on the top five list of either of the other two panelists, and thus presumably would not have made the "cut." To the extent that the district court found the evidence of discrimination in the makeup of the review panel, the court clearly erred.

4. Evidence of an Unwritten Policy to Discriminate Against Persons Over Forty

The district court next stated that "[e]vidence was adduced" that the Service "had an unwritten policy discouraging promotion of persons over forty to upwardly mobile positions." We assume from its inclusion of this oblique statement in its findings that the court accepted as fact that such a "policy" actually existed. The record, however, does not support such a conclusion.

The only evidence that Odom produced on this point consists of four statements, which together cannot overcome the clearly erroneous standard to support a finding of an "unwritten policy" of age discrimination. First, Inspector Smith, one of the review panelists, mused that, at one point in his career, he may have been denied both a promotion and a lateral move because of age. Second, Smith said that, in his opinion, persons over fifty did not have the same chance for advancement as younger workers. Third, Inspector Gump, the chairman of the review panel, had written "potential for advancement" in his notes concerning general factors that the panelists should consider; and he testified that he had intended to use this factor only in the event a tie-breaker became necessary--which it never did. Finally, there are Odom's self-serving but otherwise unsupported assertions concerning his belief in the existence of such an unwritten policy.

All of the evidence adduced to demonstrate the existence of such a "policy" is at most anecdotal and bare speculation. We thus hold that the district court's finding that the Service maintained such an unwritten policy--and the inclusion thereof among the "circumstances" considered by the court in concluding that discrimination occurred--was clear error.

Finally, the district court's last finding among the totality of circumstances supporting a determination that the Service had discriminated against Odom was that "[t]he statistical data ... indicates [sic] exclusions of blacks and persons over the age of forty as selectees for higher level positions." In our meticulous combing of the record of this appeal, we find no statistical evidence introduced by Odom that supports this finding.

The only items of evidence introduced in the trial that could be viewed (even erroneously) as statistical evidence are found in the Service's responses to Odom's interrogatories--responses which the Service entered into evidence at trial. They are nothing more than raw data of the age, race, and location of persons promoted to level 24 positions in the inspection service from 1980-1983. Odom introduced no analysis of this data. The "statistical evidence" that was presented to the district court, without more, is not competent to prove anything. It is simply impossible to discern from the record what (if anything) the Service's responses to the interrogatories are supposed to mean--much less to determine that they "indicate[ ] exclusion of blacks and persons over the age of forty as selectees for higher level positions within the Postal Inspection Service."27 The district court's finding that this so-called statistical evidence demonstrated discrimination was thus clearly erroneous.

Although it does not demonstrate age-based or racial discrimination, the evidence adduced in the instant case does appear to confirm something else of an untoward nature: the perpetuation of what, for lack of a better term, is frequently labeled the "good old boy" network. Much of the evidence in this case--e.g., the subliminal messages in Strader's recommendations and the fact that after the panel completed its work, Moore (who was supposed to be the final decisionmaker) called Strader and, in effect, allowed him to pick the person he wanted to fill the position--gives us the impression that the decisionmakers were merely "going through the motions" of the required procedure while in fact ensuring that Strader would get the person he wanted to fill the new position. But even if that impression is correct, it does not amount to racial or age-based discrimination. Again, misfeasance, malfeasance, or nonfeasance--without nexus to age or race--is not actionable here.

The essence of the Service's explanation concerning why Odom did not receive the subject promotion is that, at the conclusion of its promotion process, Odom was simply not the top-rated candidate. Even though the conduct of the Service's process might not pass the "smell test," and might well raise eyebrows, Odom has failed to adduce forth sufficient evidence to demonstrate that discriminatory intent motivated the acts or omissions of anyone involved in the promotion process.

In St. Mary's Honor Center v. Hicks,28 the Supreme Court recently discussed, at length, the burden of proof applicable to a claim of racial discrimination.

The factfinder's disbelief of the reasons put forward by the defendant (particularly if disbelief is accompanied by a suspicion of mendacity) may, together with the elements of the prima facia case, suffice to show intentional discrimination. Thus, rejection of the defendant's proffered reasons, will permit the trier of fact to infer the ultimate fact of intentional discrimination,4 and the Court of Appeals was correct when it noted that, upon such rejection, "[n]o additional proof of discrimination is required." [Hicks v. St. Mary's Honor Center ], 970 F.2d at 493 [ (8th Cir.1992) ] (emphasis added). But the Court of Appeals' holding that rejection of the defendant's proffered reasons compels judgment for the plaintiff disregards the fundamental principle of Rule 301 that a presumption does not shift the burden of proof, and ignores our repeated admonition that the Title VII plaintiff at all times bears the "ultimate burden of persuasion."

Odom has failed as a matter of law to demonstrate race or age discrimination, even though his evidence may be sufficient to bring into question the objectivity of the Service's selection process as administered. But that alone cannot carry the day.

Odom failed to produce sufficient evidence that the reasons given by the Service for his not receiving the promotion were a pretextual smokescreen masking racial or age-based discrimination. We have found the district court's express or implied factual findings, which supported its ultimate finding of discrimination, to be clearly erroneous. We therefore REVERSE the judgment of that court and RENDER judgment in favor of the Service, dismissing Odom's action.