Plaintiff-appellant Charlotte Dexter began working for Honeywell, Inc. ("Honeywell") in 1964. Plaintiff-appellant Howard Evers began working for Honeywell in 1979. In 1990, Honeywell spun off its defense business to Alliant Techsystems, Inc. ("Alliant"). Following the spin-off, both Dexter and Evers became Alliant employees. Dexter was laid off, effective April 22, 1993, in connection with a workforce reduction.2 At the time of her termination, Dexter was fifty-four years old. Evers was also terminated as a part of the reduction in workforce. His termination was effective May 14, 1993. Evers was sixty-two years old at the time.

Both Evers and Dexter filed suit against Alliant alleging that they were terminated because of their age in violation of the Age Discrimination in Employment Act ("ADEA"), 29 U.S.C. §§ 621-34, and the Minnesota Human Rights Act ("MHRA"), Minn. Stat. § 363.3 The district court granted summary judgment in favor of Alliant in each case. Plaintiffs filed timely notices of appeal, and the cases were consolidated for argument and submission.4

Dexter began working for Honeywell in 1964 as a switchboard operator. By 1976 or 1977, Dexter had risen to the position of Labor Relations Analyst. After being promoted to Labor Relations Representative, Dexter made a lateral move in 1983 to become a Cost Estimating Administrator in Honeywell's Underseas System Department. From 1989 until her layoff, Dexter's principal responsibility was to develop cost estimates for the sale of spare parts for the Mark 46 torpedo to the U.S. and foreign governments.

On December 8, 1992, Alliant CEO Toby Warson asked top administrators to prepare proposals for a reduction in administrative personnel. Each department head then received a memo providing a "projected headcount reduction" for his or her department. A short time later, Theresa Haugan, the manager to whom Dexter reported, was informed that she needed to terminate two employees. Haugan proceeded to assign ratings to her employees based on Alliant's Workforce Reduction Criteria, which contained the following five express criteria: (1) performance rating, (2) performance ranking, (3) critical skills, (4) cross-functional capabilities, and (5) leadership.

In making her evaluations under the Workforce Reduction Criteria, Haugan relied in part on the results of self-assessments that had been administered to the Cost Estimating Administrators under Haugan's supervision in the summer of 1992. These self-assessments were created with input from managers, human resource personnel, and the Cost Estimating Administrators themselves. Haugan compiled the information from the assessments and provided each Cost Estimating Administrator with a copy of the results in August 1992. Dexter's self-assessment resulted in a total score of 67, the lowest of all the Cost Estimating Administrators in Haugan's department. The next lowest point total was 101. In making her workforce reduction ratings, Haugan also relied on her own observations of each employee's work, information contained in each employee's personnel file, and input from other managers.5 After completing her ratings, in December 1992, Haugan met with all of her Cost Estimating Administrators and distributed copies of the ratings, "to show the Administrators where they ranked." Dexter did not approach Haugan with any concerns about her rating or ranking.

Based on her rating, Haugan ranked Dexter last of the nine employees in her job category and, as a result, selected Dexter and Pam Christopher, age thirty-five and the second lowest ranked Cost Estimating Administrator, for termination. Haugan's lay-off decisions were approved without change by higher management and Human Resources personnel.

Evers began working for Honeywell in November 1979 as a Liaison Engineer. In 1981, Evers became a Senior Design Engineer in the Underseas Systems Division, where he was part of the Torpedo Design Group. Evers remained in that position without promotion until his termination.

In late 1992, Mitch Erikson, the head of the Mechanical Design Unit, which included the Torpedo Design Group, instructed his managers to rank their employees consistent with the Workforce Reduction Criteria and to make layoff recommendations. Scott Lenberg, Evers' manager, completed Evers' rating. After the ratings were completed, Evers' ranked last among engineers in his grade, EN22, receiving only nine of twenty-five possible points. Jim Belling, age thirty-three, ranked last among EN23 engineers.6 Based on these rankings, together with a discussion by the managers, Evers and Bellings were selected for termination.

We review the district court's grant of summary judgment de novo. Fisher v. Pharmacia & Upjohn, 225 F.3d 915, 919 (8th Cir. 2000). Summary judgment is appropriate when the evidence, viewed in the light most favorable to the non-moving party, indicates that there is no genuine issue of material fact and the moving party is entitled to judgment as a matter of law. Id.

Age discrimination may be established under either a disparate impact or a disparate treatment theory.7 Evers and Dexter bring claims under both theories. We turn first to the disparate impact claims. Under a disparate impact approach, the plaintiff does not need to prove intentional discrimination. Watson v. Fort Worth Bank & Trust, 487 U.S. 977, 987 (1988). "[T]he necessary premise of the disparate impact approach is that some employment practices, adopted without a deliberately discriminatory motive, may in operation be functionally equivalent to intentional discrimination." Id.

A plaintiff in a disparate impact case must first establish a prima facie case of disparate impact by identifying a specific employment practice and then presenting statistical evidence of a kind and degree sufficient to show that the practice in question caused the plaintiff to suffer adverse employment action because of his or her membership in a protected group. Watson, 487 U.S. at 994. If the plaintiff succeeds in making this prima facie showing, the burden then shifts to the employer to produce evidence demonstrating a legitimate business reason for the challenged practice. Id. at 997-98.8 If the employer successfully establishes a business justification, the plaintiff may still prevail by demonstrating that a comparably effective alternative practice would produce a significantly smaller adverse impact on the protected class. Id. at 998; Minn. Stat. § 363.03, subd. 11.

Appellants point to three facially-neutral employment practices in an attempt to establish a prima facie case. Specifically, appellants challenge (1) Alliant's lay-off guidelines and ranking process, (2) Alliant's budgeting process, and (3) the overall reduction in force.

With respect to the lay-off guidelines and ranking process, both Evers and Dexter challenge the guidelines and ranking system in general. We assume without deciding that appellants can establish a prima facie case of disparate impact. Alliant points to the need to keep its best employees as a business justification for its criteria and ranking process. Appellants counter that Alliant had no legitimate reason for using subjective criteria and for failing to include such considerations as seniority and historical performance in rating its employees. These arguments, however, fall under the less discriminatory alternative prong, because appellants are actually asserting that a less discriminatory rating system could be devised. Appellants point out that Alliant's "original ranking procedures" included "protections for older employees" and argue that, instead of eliminating these protections from its lay-off criteria, Alliant could have left these protections intact or replaced them with other similar protections. It is clear that a system with built-in protection for older employees would result in less of an adverse impact on older workers. However, appellants provide no evidence that such a system would be comparably effective in achieving Alliant's legitimate goal of retaining the best employees. Without such evidence, appellants' claims fail.

Dexter further maintains that the ranking process was faulty because Haugan relied in part on the "largely-inapplicable self-assessment[s]," failed to consult with Dexter's prior manager, and gave Dexter ratings that directly contradicted her prior evaluations. This claim fails at the prima facie case stage because Dexter cannot show a statistical disparity resulting from Haugan's practices, including the use of the self-assessments. Haugan was responsible for ranking only the nine Cost Estimating Administrators who worked below her. In Haugan's ranking, three of the top four rated Cost Estimating Administrators were over forty years old. The bottom five rated employees, with the exception of Dexter, were all under forty years old. There is no evidence that Haugan's application of the rating procedures resulted in a disparate impact on older employees.

Appellants next argue that Alliant's budgeting process was discriminatory because Alliant directed its supervisors to reduce their departmental budgets by a certain dollar amount and gave them head count reductions as well. Appellants contend that "the obvious alternative would have been for Alliant to set a monetary budget, but not a 'suggested headcount' reduction." However, before reaching alternatives, appellants must show a causal connection between the budgeting process and the statistical disparity. Appellants argue that "[i]t stands to reason that if a supervisor were directed to reduce the department budget by $100,000 and to reduce the headcount by one employee, the supervisor would eliminate one employee who earned $100,000." Appellants further contend that "there is a correlation between salary and years of service, and a correlation between years of service and age." Appellants support this contention with the affidavit of former Alliant employee Vatsal Munshi. Munshi, however, speaks only in generalizations. In response, Alliant points to statistics, which appellants do not contest, showing that, at all times relevant, the average salaries of the individuals who were laid off were lower than the average salaries of those who were retained. Given this fact, appellants fail to show that the budgeting process resulted in the layoff of older, more highly paid employees.

Finally, with respect to the overall reduction in force, we assume, without deciding, that appellants can establish a prima facie case. It is undisputed that Alliant experienced a sharp decline in revenue, prompting the overall reduction in force.9 Alliant, therefore, has established a business justification for the reduction in force. Appellants broadly contend that other, less discriminatory means of reducing costs were available to Alliant. However, to support this assertion, appellants point only to a document created by Alliant's Human Resources Department referred to as the "White Paper." The White Paper was created in 1994, after both Evers and Dexter had been terminated, by a member of the Human Resources Department who had been assigned the task of developing creative methods for reducing Alliant's costs. The White Paper sets forth a laundry list of possible options including, among other things, across the board salary reductions, buyouts into retirement, time off without pay, extended leave with benefits, forced vacations, elimination of overtime, discontinuation of the relocation bonus program, and a hiring freeze. The White Paper merely lists these possibilities and does not address their viability. Appellants have presented no evidence that any of these alternatives would result in cost savings comparable to those realized by the overall reduction in force or that the alternatives would be less discriminatory. Without this showing, appellants cannot overcome the legitimate business reason proffered by Alliant. The district court correctly granted summary judgment in favor of Alliant on appellants' disparate impact claims.

We turn now to appellants' disparate treatment claims. "The ultimate question in every employment discrimination case involving a claim of disparate treatment is whether the plaintiff was the victim of intentional discrimination." Reeves v. Sanderson Plumbing Prods., Inc., 120 S. Ct. 2097, 2111 (2000). Under both the ADEA and the MHRA, disparate treatment claims based on indirect evidence are governed by the burden-shifting framework set out in McDonnell Douglas Corp. v. Green, 411 U.S. 792 (1973). See Snow v. Ridgeview Med. Ctr., 128 F.3d 1201, 1208 (8th Cir. 1997) (citing Bevan v. Honeywell, Inc., 118 F.3d 603, 613 (8th Cir. 1997); Feges v. Perkins Rest., Inc., 483 N.W.2d 701, 710 (Minn. 1992)).10 Under this framework, a plaintiff must first establish a prima facie case of age discrimination. Fisher, 225 F.3d at 919. If the plaintiff establishes a prima facie case, the burden then shifts to the employer who must "articulate a legitimate, nondiscriminatory reason for any adverse employment action taken against [the plaintiff]." Id. "If [the employer] puts forth such a reason, [the plaintiff] must then present evidence sufficient to raise a question of material fact as to whether [the employer's] proffered reason was pretextual and to create a reasonable inference that age was a determinative factor in the adverse employment decision." Id.

We assume that appellants are each able to establish a prima facie case of age discrimination. Alliant asserts that Evers and Dexter were selected for termination because business conditions necessitated the implementation of a reduction in workforce and both Evers and Dexter received low rankings under the Workforce Reduction Criteria. Appellants challenge this explanation, arguing that there are material issues of fact as to its truth in two respects. See Reeves, 120 S. Ct. at 2108 ("In appropriate circumstances, the trier of fact can reasonably infer from the falsity of the explanation that the employer is dissembling to cover up a discriminatory purpose.") First, appellants contend that "there are factual issues as to the need for Alliant to undertake such massive workforce reductions." However, "when a company exercises its business judgment in deciding to reduce its work force, 'it need not provide evidence of financial distress to make it a 'legitimate' RIF.'" Regel v. K-Mart Corp., 190 F.3d 876, 880 (8th Cir. 1999) (quoting Hardin v. Hussmann Corp., 45 F.3d 262, 265 (8th Cir. 1995)). "A company's exercise of its business judgment is not a proper subject for judicial oversight." Id. (internal quotations and citations omitted).

Appellants further challenge their low ratings. Dexter argues that Haugan's rating was faulty in three respects. Dexter received a rating of two out of five possible points in the area of critical skills. Dexter argues that this rating was unsupported because Haugan relied on the self-assessments that had been completed in the summer of 1992.11 Dexter contends that the self-assessments were faulty because they addressed many computer skills that were inapplicable to the Cost Estimating Administrator position and failed to include "the vast majority of the skills" that Dexter actually used in her position. However, there is nothing inherently discriminatory in the skills that are included in the self-assessments. Eight of the nine Cost Estimating Administrators who were evaluated as a part of the 1993 reduction in force took part in the 1992 self-evaluations. The highest score on the self-assessments was a 157. The two administrators besides Dexter who were over forty received scores of 143 and 142, placing them fourth and fifth in the self-assessment ranking. These scores were within 15 points of the top score and more than double Dexter's score of 67. The sixth and seventh place scores, 108 and 101, belonged to employees who were thirty-nine and thirty years old, respectively.

Dexter next argues that Haugan gave her a "bogus" score in the cross-functional capabilities category. Haugan rated Dexter a two in this area which corresponded to a finding that Dexter had a "demonstrated ability to absorb small increments of additional/new responsibilities within the same function."12 Dexter argues that her low score was unsupported because (1) Haugan gave two secretaries higher ratings in this area, (2) Dexter's last performance evaluation stated that she consistently met expectations "outside of Cost Estimating," and (3) Dexter demonstrated cross-functional capabilities through her previous work in labor relations. First, it is undisputed that the secretaries were rated based on their own job family, which is separate and distinct from the Cost Estimating Administrator position. The fact that Haugan believed that two of the secretaries demonstrated skills outside of their secretarial duties is irrelevant to her determination regarding Dexter's ability to work outside of the Cost Estimating Administrator position.

Furthermore, Dexter points to a portion of her 1992 performance evaluation to support the assertion that she consistently met expectations "outside of Cost Estimating." An examination of the 1992 evaluation reveals that the quoted portion of the evaluation is included under the "Teamwork and Involvement" section, which provides:

2. Teamwork and Involvement - Involves others as needed to get the job done in a timely and efficient manner. Solicits and provides constructive feedback, exchanges viewpoints objectively, and shares information.

Within the department, Char seems to operate very well. She solicits input either from her work director or other Cost Estimators [to] aid in accomplishing estimates. Outside of Cost Estimating Char interfaces with individuals in Procurement, Material Evaluation Engineering, and Subcontracts who need constant surveilance [sic] to assure timely submission of their inputs. Consistently Meets

Clearly, this portion of the 1992 evaluation deals with communication with other departments, not with the development of cross-functional skills. Haugan's rating of Dexter's cross-functional capabilities in connection with the reduction in force is not inconsistent with Dexter's 1992 performance evaluation which did not address cross-functional skills. See Hutson v. McDonnell Douglas Corp., 63 F.3d 771, 780 (8th Cir. 1995) (finding no evidence of pretext when plaintiff failed to identify any actual conflicts between assessment in question and prior evaluations). Also, with respect to Dexter's work in labor relations, it is not the role of this court to sit as a "super-personnel department" to second guess the wisdom of a business's personnel decisions. See Hill v. St. Louis Univ., 123 F.3d 1114, 1120 (8th Cir. 1997) (citing Hutson, 63 F.3d at 781). It is not unlawful for a company to make personnel decisions based on erroneous evaluations. Id. While Dexter characterizes her cross-functional skills as "demonstrated and acknowledged," she fails to point to any evidence supporting this assertion. Haugan testified in her deposition that, in her opinion, the fact that Dexter worked in another department did not necessarily mean that Dexter had cross-functional capabilities. There is no evidence to support an inference that Haugan believed that Dexter had cross-functional skills at levels beyond a two, yet chose to give her a lower rating.

Finally, Dexter contends that her rating of two in the leadership category was inconsistent with the 1992 performance review in which she had received a rating of "consistently exceeds expectations" in a category entitled "Innovation - assumes initiative to propose and implement new and promising ideas. Shows good judgment in evaluation alternative approaches for improvement." Dexter argues that a rating of "at least a '3' would have been consistent." Under the Workforce Reduction Criteria, a score of two in leadership corresponds to a finding that the employee "[h]as potential to lead in either a technical or team environment." A score of three would be appropriate if the employee "[d]emonstrates leadership skills in affecting desired change in either technical or team environment." Once again, we find no inconsistency between the 1992 evaluation and Haugan's rating. The 1992 evaluation noted that "[c]reative innovation is one of Char's strong points. She does excellent work in developing customized spread sheets on the PC." The evaluator cited a recent project as an example of Dexter's excellent customized spread sheets. However, the evaluation did not address leadership skills in connection with the assessment of Dexter's innovation.

Evers raises only a general challenge to his low rating, asserting that Alliant produced no evidence to justify the low scores that Lenberg assigned to Evers. Evers attempts to challenge these low scores, relying only on his own testimony that he "performed as well as many of the retained employees." Evers, however, fails to produce any affirmative evidence that Lenberg manipulated the rating system. See Walton v. McDonnell Douglas Corp., 167 F.3d 423, 428 (8th Cir. 1999) (stating that, in order to defeat summary judgment, a plaintiff must present affirmative evidence rather than simply contend that a jury might disbelieve the defendant's evidence). Furthermore, "a comparison that reveals that [the plaintiff] was only as qualified as the retained employees would not raise an inference of discrimination." Lewis v. Aerospace Comm. Credit Union, 114 F.3d 745, 749 (8th Cir. 1997).

We find sufficient evidence in the record to support Alliant's proffered reasons for both terminations. As a result, "any presumption of discrimination 'drops out of the picture.'" Bogren v. Minnesota, 236 F.3d 399, 404 (8th Cir. Dec. 22, 2000) (citing St. Mary's Honor Ctr. v. Hicks, 509 U.S. 502, 511 (1993)). Appellants, therefore, bear the burden of producing sufficient evidence from which a reasonable trier of fact could conclude that Alliant's explanations are pretextual. Id. at 404.

Appellants argue that they have created a material issue of fact as to pretext, pointing to statistical evidence, Alliant's rating process, and written and oral comments as support. As circumstantial evidence of pretext, appellants point to company-wide statistical studies conducted by their expert, Alliant, and the EEOC which they claim prove the disproportionate lay-off of older employees. See, e.g., Bevan v. Honeywell, Inc., 118 F.3d 603, 611-12 (8th Cir. 1997). One summary of the statistical analysis of the Alliant layoffs prepared by the EEOC reads as follows, "The evidence shows that at the end of 1990, employees age 40 and over constituted 42% of the salaried workforce, yet 58% of the employees laid off over the next five years were age 40 or over, and more than half of those were age 50 and above."13 Appellants do not dispute, however, Alliant's assertion that from 1990 to 1995, the average age of Alliant's workforce rose from 39.5 years to 43.25 years. Alliant further asserts without contradiction by appellants that, assuming no hiring and no departures except normal retirement, the projected increase in average age during that period would have been from 39.5 to 43.27 years.

Moreover, in order for statistical evidence to be probative of pretext, it must analyze the treatment of comparable employees. Brown v. McDonnell Douglas Corp., 113 F.3d 139, 142 (8th Cir. 1997). Appellants' expert, Dr. David Peterson, admits that his investigation into "the extent to which employee performance may account for the disparities . . . has been limited." Appellants point out that Dr. Peterson's report states that "older employees generally received lower ratings than other people in the pools with which they were compared." However, Dr. Peterson recognized, both in his deposition and in a treatise of which he was a co-author, that because successful employees tend to be promoted, low performers at any level will tend to be older than others at the same level. While we recognize that company-wide statistics can be probative as to the presence of a general climate of age bias, see, e.g., Bevan, 118 F.3d at 611, we conclude that, given the evidence in this case, appellants' statistics are insufficient to create a material issue of fact as to pretext.

Appellants contend that "Alliant's layoff decisions were subjective, uninformed, and reeked of favoritism," citing testimony that managers often made changes to lay-off lists at their own whims and that one administrator ordered his friends' names removed from lay-off lists. Appellants further challenge Alliant's decision to change its rating process, which stripped away built-in protections for older employees and instituted a "snap-shot" approach of evaluating performance based only on recent performance. "[T]he presence of subjectivity in employment evaluations is itself not a grounds for challenging those evaluations as discriminatory." Hutson, 63 F.3d at 780. Moreover, "'[t]here is nothing inherently discriminatory in an employer choosing to rely on recent performance more heavily than past performance in deciding which employees to terminate during a [reduction in force].'" Dammen v. Unimed Med. Ctr., 236 F.3d 978, 982 (8th Cir. Jan. 17, 2001) (quoting Brown, 113 F.3d at 142). As we have noted, it is not unlawful for an employer to make "employment decisions based upon poor job performance, erroneous evaluations, personal conflicts between employees, or even unsound business practices," Hill, 123 F.3d at 1120, as long as these decisions are not the result of discrimination based on an employee's membership in a protected class. Appellants have not produced any evidence that Alliant made changes to the lay-off lists based on employee age.

Finally, appellants identify several comments and written statements as evidence of pretext. However, all of this evidence is too remote in time from the layoffs of Evers and Dexter to be probative. See EEOC v. McDonnell Douglas Corp., 191 F.3d 948, 953 (8th Cir. 1999); Walton, 167 F.3d at 428 (holding that a remark by a decisionmaker made two years before termination was too remote in time to support a finding of pretext for intentional discrimination). None of these comments or statements were causally related to Evers and Dexter's terminations. Stray remarks, without more, do not give rise to an inference of discrimination. Fisher, 225 F.3d at 922. A careful review of the record reveals that the facts of these cases do not support a reasonable inference of discriminatory intent. Appellants' disparate treatment claims therefore fail. As we stated in Hill, it must be remembered that anti-discrimination laws, including the ADEA and the MHRA, "are not meant to transform 'at will' employment into perpetual employment where equal treatment is guaranteed to all employees and termination is legal only 'for cause.'" Hill, 123 F.3d at 1120. The statutes instead "serve the narrow purpose of prohibiting discrimination based on certain, discreet classifications," including age. Id. We have previously recognized the unfortunate fact that "even capable employees are released when an employer is down-sizing." Hutson, 63 F.3d at 779. Because appellants fail to point to any evidence which would support a finding of age discrimination, the district court correctly granted summary judgment in favor of Alliant.

AFFIRMED.