dissenting.
Our review of this case leads me to several conclusions not shared by the majority of the panel, prompting this dissenting opinion, respectfully submitted.
I. PRELIMINARY.
There are no disagreements of the panel on several preliminary observations:
A. Equal Work Findings.
Appellant Treva Bohm failed to prove that her employer, respondent Hartz Wholesale, paid wages to her that were unequal to those paid to men who did equal work. Equal pay for equal work is required by the Equal Pay Act of 1963, 29 U.S.C. § 206(d)(1), and relief for a violation of the act can be obtained in a claim of sex discrimination under section 703(a), Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e-2(a), or the third section of the Minnesota Human Rights Act, Minn. Stat. § 363.03, subd. l(2)(c) (1980). Danz v. Jones, 263 N.W.2d 395 (Minn.1978). Appellant no longer quarrels with the trial court’s finding on this topic.
B. Analysis.
The pattern for analysis of discrimination cases has been repeatedly announced by the federal courts and the Minnesota Su*909preme Court. The supreme court has succinctly observed:
Because of the substantial similarities in the language and purposes of the two statutes, in construing the Minnesota Human Rights Act this court has applied principles developed in the adjudication of claims arising under Title VII of the 1964 Civil Rights Act, 42 U.S.C. § 2000e (1976). See, e.g., Fisher Nut Co. v. Lewis ex rel. Garcia, 320 N.W.2d 731, 734 (Minn.1982); Danz v. Jones, 263 N.W.2d 395, 398-99 (Minn.1978). The beginning point for discussion of cases brought under Title VII is generally the three-part analysis established by the Supreme Court in McDonnell-Douglas Corp. v. Green, 411 U.S. 792, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973), which consists of a prima facie case, an answer, and a rebuttal. Because McDonnell-Douglas concerned an allegation of discrimination based only on an employer’s refusal to hire, the specific elements of the McDonnell-Douglas court’s formulation of the plaintiff’s prima facie case must be modified for varying factual patterns and employment contexts.
Hubbard v. United Press International, Inc., 330 N.W.2d 428, 441-42 (Minn.1983) (footnotes omitted).
The United States Supreme Court has explained that a prima facie case is established when the plaintiff has offered evidence “adequate to create an inference” that a decision was based on a discriminatory criterion. International Brotherhood of Teamsters v. United States, 431 U.S. 324, 358, 97 S.Ct. 1843, 1866, 52 L.Ed.2d 396 (1977). The Minnesota Supreme Court has held that the plaintiff’s burden is to show the “bare essentials of unequal treatment based on * * * sex.” Danz, 263 N.W.2d at 399. In Teamsters, the Supreme Court said that the inference of discrimination is created by evidence that eliminates the most common legitimate reasons which the employer might have used to make its decision. Teamsters, 431 U.S. at 358 n. 44, 97 S.Ct. at 1866 n. 44.
The Eighth Circuit Court of Appeals has explained that a prima facie case establishes a “presumption of illegal discrimination,” and that this presumption “drops from the case” if the employer succeeds in proving a legitimate, nondiscriminatory reason for its action. Craik v. Minnesota State University Board, 731 F.2d 465, 469 (8th Cir.1984). If the presumption is overcome, the plaintiff has the ultimate burden of persuasion on claims of individual discrimination. Danz, 263 N.W.2d at 399.1
C. Scope of Review.
Our scope of review is mixed. Factual findings by the trial court will not be reversed on appeal unless they are clearly erroneous. Dodge v. Minnesota Mining and Manufacturing Co., 278 N.W.2d 97, 100 (Minn.1979). On the other hand, in cases of this kind the appellate courts commonly find no substantial dispute as to the operative facts. In such a case, “we may freely review the evidence and draw our own inferences.” Kaster v. Independent School District No. 625, 284 N.W.2d 362, 366 (Minn.1979).
Most of the facts discussed in this opinion involve no substantial dispute, so that we must interpret the facts in light of controlling principles of law. However, unless they are clearly erroneous, we must defer to several trial court findings on disputed facts, one on the skills of appellant, one on the significance of market information to explain Hartz Wholesale’s practices, and another on the significance of economic pressures in the choice of respondent to discharge Treva Bohm.
*910II. FACTS.
To complete this opinion, it is necessary to restate the facts of the case. On some subjects, such as the amounts of wages paid by the employer before raises given late in 1980, the statement here conflicts with the report of facts in the majority opinion. A careful study of the record, including business records received as exhibits at trial, confirms the accuracy of the facts stated here. Although many of these facts were not recited by the trial court, none conflict with trial court findings. Our review of the evidence was burdened, as the trial court’s must have been, by the fact that several critical witnesses were not called upon to elaborate on some important questions.
A. Hartz Wholesale Wage Practices.
1.In November 1980, Treva Bohm was one of 13 women who were employed by Hartz Wholesale and were classified by the employer as “unskilled clerical.” Bohm and seven others were paid an hourly wage of $6.53, the highest wage rate for those in this class.
Bohm is a high school graduate and successfully completed a one year secretarial course at an area vocational-technical school. She began work for Hartz Wholesale as a keypunch operator, entering information into a computer.
Beginning in 1976, her second year at Hartz, Bohm worked in the computer and billing department, also called the data processing department. Her duties included filing, telephone answering, recording hours and mileage of truck drivers, assisting in preparing and handling invoices, and assisting Walter Lund in preparing daily master sheets, documents which showed deliveries to be made, trucks to be used, and load weights. The evidence does not permit a finding that Bohm’s work required no skills or qualifications.
The 13 “unskilled clerical” workers were employed in five different departments. Eight received the $6.53 hourly wage in 1980, and there is no testimony that their duties were the same.
2. Hartz Wholesale’s records show employment of approximately 125 men at the end of 1980. About 20 men were salaried; no women held salaried jobs for Hartz. About 90 men earned wages of over $7.60 per hour, and more than 30 of those earned $8.00 per hour or more.
The trial court identified six male employees of Hartz who were paid less than $6.53 per hour in 1980. All were part-time employees, and two were hired in 1980. They worked in three departments where the hourly wage of full-time non-salaried men averaged about $7.80.
3. Two male employees of Hartz Wholesale, both assigned with appellant to the computer and billing department, had clerical responsibilities. These two employees were paid $8.99 per hour in November 1980; there is evidence that appellant’s work was not equal to theirs. Each of the men, however, was classified without reference to clerical duties; one was classified “merchandise shipment,” and the other “merchandise receivings.”
4. Hartz Wholesale employed four women in 1980 who were not classified as “unskilled clerical.” One was a junior buyer, two were computer operators, and one was classified as “clerical.” The hourly wages of the four women averaged $7.33.
These four women had unusual job qualifications. Each did work similar to male employees whose earnings were higher. The two computer operators were paid less than the lowest paid full-time male employees of Hartz, those who received $7.60 per hour for warehousing work.
5. In December 1980, Hartz department heads met to set employee wages for 1981. Clerical employees, the “unskilled clerical” workers, received a 5.97 percent wage increase, 39 cents per hour.2 Other employees (except for three salaried men and two of the janitors) received wage increases of up to 20 percent.
*911The majority opinion describes a wage survey conducted by a Hartz official late in 1980. This officer testified that an analysis of the survey provided the “reason” for wage decisions at that time. The raise for female clerical workers, this officer explained, was due to the observation that Hartz had “reached a ceiling” for that class of employees.
The 1981 Hartz wage increases were announced to the seven computer department employees at a December meeting in the office of Norman Olson, head of the department. Treva Bohm commented on the unequal effect of the wage increases among department employees. In fact, two men who had clerical duties, and who were already paid $2.46 per hour more than the women in clerical jobs, received increases of 89 cents per hour, a 9.9 percent increase, more than twice the amount of the women’s raises.
B. Appellant Laid Off.
1. At the December 1980 department meeting, Treva Bohm vigorously challenged Norman Olson about the wage increase plan. She told him: “You can’t do that.” She commented about the effect of the plan in the department. Olson testified at trial that the tenor of Bohm’s complaint was that female employees were not being treated fairly. Olson told her at the meeting that the plan wouldn’t be changed.
Bohm talked to Olson about the raises three or four times, principally in the two or three weeks after the department meeting. Olson said she should talk to the board of directors if she had a complaint. Olson, Bohm says, told her he would let her know when the board would meet; when she asked Olson later about the board meeting, she says he told her the meeting had already been held.
2. Early in 1981, Bohm’s job was cut back to half days. Olson, to whom appellant made complaints about the 1981 raises, decided on the cutback, and he had exclusive authority to make decisions of the kind for his department. About two months later, in late February or early March, Olson told Bohm she was laid off, effective in two weeks. She was not replaced.
Olson testified that the cutback in hours was due to a shortage of work, and this was neither explained nor documented.
3.When Bohm was laid off, Olson told her the decision was based on a need of the company to “cut back due to loss of some accounts.” At trial, respondent reiterated that an expected loss of some gross receipts prompted termination of Bohm’s employment.
a. The company “cutback” in 1981 was said to be due to a loss of three customers, effective April 1, 1981. The corporate secretary said the company anticipated a 7 percent sales reduction; other evidence indicates that a 7 percent reduction would lower receipts by about $245,000. The secretary said company officers decided to cut about 5 to 7 percent of its expenses, but the total amount of expenses was not shown.
The company succeeded in cutting $200,-000 in expenses, five and three-fourths percent of its predicted revenue losses, through changes in its rental of computers and purchase of insurance. No other data was given to substantiate the claim of a need for expense cutbacks.
b. Only Treva Bohm and one other employee were permanently terminated early in 1981. One month after appellant was terminated, the company laid off a part-time warehouse employee who earned wages of $4.50 per hour. One other employee was laid off for six weeks, and another for three months. The weekly hours of one worker were reduced from 40 to 25 for eight months.
The layoff of Treva Bohm saved Hartz Wholesale wages of no more than $13,800 per year. Her wages were the only substantial wage savings shown by respondent.
c. A company officer testified that decisions on reducing expenses early in 1981 included instructions to Olson that he terminate an employee. Olson testified otherwise. He said he was told to “cut back,” *912and that he had to look over his “full operation” to decide how cuts might be made. Olson was the company employee who rearranged computer rental arrangements early in 1981 to produce an annual savings of about $100,000.
III. ISSUES, WAGE DISCRIMINATION.
The trial court’s findings should “reflect the shifting burdens of going forward.” Danz, 263 N.W.2d at 400. Here the court found that evidence did not create an inference of unequal pay for equal work, that appellant failed to prove her work was equal to the work of any employee who received higher wages.
The trial court also found that Hartz practices “are not discriminatory towards women.” This is the only finding bearing on the question whether appellant succeeded in making a prima facie showing of wage discrimination independent of a theory of unequal pay for equal work. It is difficult to identify the trial court’s rationale.
The court may have failed to recognize a lawful basis for any discrimination claim not supported by evidence of equal work. This conclusion is erroneous. According to controlling judicial decisions, prohibited wage discrimination includes intentional inequities among men and women with different job descriptions. I conclude that the evidence here, discussed below, establishes an unrebutted prima facie case of unlawful discrimination.
Alternatively, the decision of the trial court may rest on its opinion, discussed later, that any patterns of inequality in Hartz’s wage practices were legitimately explained by looking at market information. This conclusion is also in error. Market factors do not neutralize the forms of discrimination proven here. In fact, controlling decisions demand a conclusion that applies here, that an employer cannot follow market patterns where there is no evidence of a business need to do so. Moreover, respondent’s use of market information enlarges rather than diminishes proof of discrimination.
IV. UNLAWFUL DISCRIMINATION.
Based on the following review of controlling appellate decisions, it is fundamental error to conclude that there is a lone legitimate cause of action for wage discrimination based on sex, a claim of unequal pay for equal work.
A. Discrimination Law.
In discrimination eases under Minn.Stat. ch. 363, the Minnesota Supreme Court has applied principles developed in court decisions under Title VII of the Civil Rights Act. Danz, 263 N.W.2d at 398-99. Since Danz was decided in 1978, the United States Supreme Court has concluded that Title VII covers sex-based wage discrimination in spite of the fact that an employer deals with employees in different job categories, those who are not engaged in equal work. County of Washington v. Gunther, 452 U.S. 161, 101 S.Ct. 2242, 68 L.Ed.2d 751 (1981). This is also the holding in a precursor to Gunther, International Union of Electrical, Radio and Machine Workers v. Westinghouse Electric Corp., 631 F.2d 1094 (3rd Cir.1980), cert. denied, 452 U.S. 967, 101 S.Ct. 3121, 69 L.Ed.2d 980 (1981).
Gunther dealt with the Bennett Amendment, a sentence added to the Civil Rights Act to exempt from its provisions wage differentiation on the basis of sex that is “authorized by” the Equal Pay Act of 1963. See 42 U.S.C. § 2000e-2(h). The Supreme Court decided that the amendment did not incorporate the “equal work” condition stated in the Equal Pay Act. Instead, the court said that the amendment exempted conduct “authorized by” provisions of the Equal Pay Act that establish four affirmative defenses to a claim of wage discrimination on the basis of sex; these defenses have to do with setting wages according to seniority, merit, or productivity. See 29 U.S.C. § 206(d)(1). Thus, except for the specified affirmative defenses, wage discrimination on the basis of sex is treated no differently than wage discrimination on the *913basis of race or religion.3 Id. 631 F.2d at 1105.
The effects of the Gunther decision are the subject of continued debate. Some courts suggest that the decision has not significantly expanded the availability of pay equity remedies. See Spaulding v. University of Washington, 740 F.2d 686 (9th Cir.), cert. denied — U.S. -, 105 S.Ct. 511, 83 L.Ed.2d 401 (1984). See also Power v. Barry County, 539 F.Supp. 721 (W.D.Mich.1982). These cases are discussed later in this opinion. Even these courts, however, recognize that Gunther establishes the legitimacy of a claim of intentional wage discrimination on the basis of sex, without proof that affected employees do equal work. Spalding, 704 F.2d at 699-700; Power, 539 F.Supp. at 726.
B. Hartz Practices.
I conclude that appellant established a prima facie case of intentional wage discrimination based on sex. The majority apparently reaches a similar conclusion. I also conclude that respondent failed to state or prove any legitimate rationale for its wage practices. Respondent offered only one explanation for its practices, an annual survey of wages paid in Northern Minnesota, and this special topic is the subject of a later part of this opinion.
Eight observations about Hartz Wholesale wage practices demonstrate intentional sex-based discrimination:
1. The trial court erred, I believe, in disregarding the significance of the classification of appellant and others as “unskilled clerical” employees. There is no evidence suggesting that these people had no skills. Appellant had special education and experience that qualified her for numerous clerical tasks. The category of “unskilled clerical” is a specious device that serves to limit the wages for women; it is “transparently” discriminatory. See Gunther, 452 U.S. at 179, 101 S.Ct. at 2252. We have not found an appellate decision overlooking the discriminatory effect of a fallacious category of jobs that is this prominent in a business.
The trial court may have suggested a finding of fact on this topic. Its memorandum attached to a post-trial order states that Treva Bohm’s duties “clearly fall within” the category of “unskilled clerical.” On the evidence of record, that finding is clearly erroneous.
2. The category of “unskilled clerical” was more than misleading. It was over-broad. It was, as the Eighth Circuit Court of Appeals said of certain inter-departmental groupings, “too broad to be meaningful.” Craik, 731 F.2d at 479. An expert witness called by respondent said that the “clerical” category was “all encompassing,” and that it included the jobs of “clerk-typist[s], secretaries, accounting clerks, bookkeepers.” At Hartz Wholesale, it in-*914eluded women with different skills and duties who worked for five different departments.
The evidence indicates that the category of unskilled clerical was specifically used for wage discrimination purposes. The only women who escaped the category were also paid less than their male counterparts. The only men whose duties might have put them within the category were given special classifications.
3. Appellant has given unchallenged evidence that a 1981 wage increase plan affected men differently than women within the computer department. Raises more than twice as large were given to men who had clerical responsibilities not greatly different than women in the department. This evidence of discrimination was not explained by the market survey taken by a Hartz Wholesale employee in 1980, and it was not otherwise explained.
4. Treva Bohm challenged the wage practices at Hartz Wholesale, and she was fired. While that topic receives further attention below, the point here is to note that it shows a discriminatory approach of the employer. Thus, the discharge is relevant to the prima facie case of wage discrimination.
5. The record shows patterns of wage discrimination other than those directly affecting Treva Bohm. There is evidence to permit findings that every female employee at Hartz Wholesale was paid less than male counterparts. That kind of a pattern is relevant evidence of intentional discrimination toward one employee. Teamsters, 431 U.S. at 339-40 n. 20, 97 S.Ct. at 1856-57 n. 20; Craik, 731 F.2d at 470; Raster, 284 N.W.2d at 365-66. In fact, as to inferences of a pattern or practice of discrimination, the Craik court suggests that the prima facie case casts the ultimate burden of persuasion on the employer.4
6. Hartz intentionally chose to put a “ceiling” on the wages of female clerical workers. This had to do with the use of a market survey, and the topic is given more attention in a later part of this opinion.
7. Respondent used suspect job classifications to set wages, as discussed above. The employer also used market information for wage decisions. The evidence shows no other standards or objective criteria used by Hartz management to set wages. As the Minnesota Supreme Court has recognized, “[i]t is precisely this type of situation maximizes the possibility” of a discriminatory choice. Kaster, 284 N.W.2d at 366.
Does the demand for scrutiny of wage-setting mechanisms constitute judicial adoption of “comparable worth” proposals? See Gunther, 452 U.S. at 166, 101 S.Ct. at 2246.
Of course, to use standards for wage decisions is to engage in comparisons of work. In this effort, it is necessary to gauge skills, effort, responsibilities, and working conditions; these are the factors introduced in the Equal Pay Act of 1963, 29 U.S.C. § 206(d)(1), and the general standards utilized in analyzing whether wages are discriminatory. Danz, 263 N.W.2d at 400. Nevertheless, this opinion does not call for judicial formulation of job classes or other specific criteria for respondent’s future wage decisions.
It is significant, here and in other cases, when an employer uses no objective criteria. Raster, 284 N.W.2d at 366. Although this is evidence supporting an inference of discrimination, the absence of controls alone does not establish the inference.
Also, the use of objective criteria may defeat a showing of discrimination suggested by other evidence. It is a worthwhile business practice, and in some situations it may be critically needed to demonstrate an equitable setting of wages. However, Title VII does not mandate the practice for all employers; this mandate is not given by the Equal Pay Act, nor by the Minnesota Human Rights Act; and this opinion does not suggest otherwise.
8. In sum, the court erred by failing to look at an important topic, the subject of *915intentional sex-based wage discrimination outside the situation of unequal pay for equal work. Evidence on that topic shows a prima facie case of unlawful discrimination. It is appropriate to count as part of that prima facie case the fact that the employer identifies no objective criteria it uses in setting wages.
The employer did not rebut evidence of discrimination in its wage decisions; it follows that appellant needed no additional proof to meet her ultimate burden of persuasion. The employer’s only attempted rebuttal, reference to a market survey, is considered below.
C. Legal Consequences.
I conclude that appropriate findings and conclusions would lead the trial court to an award of damages for inadequate wages. Further, the findings would substantiate evidence of the employer’s intent to engage in reprisal when discharging Treva Bohm after she opposed company wage practices; the reprisal topic appears to involve damages that are more substantial than those for unfair wages. Appellant’s claim of retaliatory discharge will be discussed later.
V. EMPLOYMENT MARKET FACTORS.
Responding to appellant’s claims of discriminatory wage patterns at Hartz Wholesale, the trial court concluded:
The wages defendant pays, and its decision with regard to 1981 wages, reflect the market place, the law of supply and demand. There is a rational basis in the record for defendant’s wage policy and it is not based upon a policy of sexual discrimination.
I dispute this finding. Moreover, because of appellate decisions that should influence us, this decision of the trial court was based on an erroneous view of law. As noted earlier, these errors may explain the trial court’s failure to consider inferences of discrimination among employees who had different jobs at Hartz Wholesale.
A. Christensen, The Facts.
To support regard for market facts, the trial court and the majority cite Christensen v. State of Iowa, 563 F.2d 353 (8th Cir.1977). There the Eighth Circuit Court of Appeals examined whether an employer showed a legitimate basis for wage inequities, and the court concluded that an employer’s regard for local market information was proper, that any other view “ignores economic realities.” Id. at 356.
First, it is important to note that the facts here are not the same as those in Christensen. There the alleged discrimination was in the form of a single decision to modify a wage system so that one group of employees would have the advantage of “advanced step starting pay.” Id. at 354. The disputed decision was made because of market wages for that kind of employee.
Here the suspect actions of the employer are numerous, including the formulation of classifications, the firing of an objector, and wage disparities for all women. In addition, the evidence shows that market information had nothing to do with job classifications and the firing decision. Likewise, the market does not explain disparate 1981 raises for female and male clerical workers. Finally, the employer in Christensen, unlike Hartz, had adopted and used equitable wage-setting standards.
Thus, I conclude the market information evidence here does not amply answer appellant’s prima facie case of discrimination, and that the trial court’s contrary finding was clearly erroneous. See Danz.
B. Christensen, The Law.
Even more important, there is reason to conclude that the approach of the court in Christensen is obsolete and incomplete.
A concurring judge in Christensen made a telling point. The majority view was correct, the judge said, only if it was true (because of the Bennett Amendment, discussed earlier) that Title VII prohibited only sex-based inequities between those who do equal work. 563 F.2d at 357. In other words, an enlarged prohibition, dealing with unequal pay for different work, *916could not be undone solely by reference to market information. This view is validated by an important consideration: Any wage inequality among those in different job categories can normally be explained by reference to the market.
As the concurring opinion suggests, I conclude that significance of the 1977 Eighth Circuit decision in Christensen was tied to the view of Title VII rejected by the United States Supreme Court four years later. See Gunther, and the discussion of the case earlier in this opinion. The same analysis is prompted by study of a similar decision of the Tenth Circuit Court of Appeals in 1980, another holding cited by the majority. Lemons v. City and County of Denver, 620 F.2d 228 (10th Cir.) cert. denied, 449 U.S. 888, 101 S.Ct. 244, 66 L.Ed.2d 114 (1980). There the primary opinion of the court intermixes its regard for a local market factor with its belief that the Civil Rights Act does not deal with inequities that “cross job description lines,” but only with unequal pay for “equal work.” 620 F.2d at 229-30.
C. Craik.
An alternative approach utilizes the rationale of a disparate impact claim. Using this analysis, it is allowed that the employment market should be seen as a neutral factor that an employer may use to set salaries, regardless of discriminatory impact, so long as the employer needs to look to the market for valid business reasons.5 Thus, an employer could offer high, inequitable salaries to enhance longevity or recruitment in a field where supply is low and the market wage is high.
The Eighth Circuit Court of Appeals has now used this approach. Craik, 731 F.2d at 468 n. 4. There the court permitted market-factor increases for all male job categories identified as “scarce market areas,” based on a magistrate’s finding of fact, not clearly erroneous, that the wage increases were “necessary.” Id. at 480.
The significance of the need factor noted in Craik is understood by examining Griggs v. Duke Power Company, 401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971), where the concept of disparate impact was first pronounced. Griggs involved the racial impact of testing requirements for job promotions. The supreme court said:
The Act proscribes not only overt discrimination but also practices that are fair in form, but discriminatory in operation. The touchstone is business necessity. If an employment practice which operates to exclude Negroes cannot be shown to be related to job performance, the practice is prohibited.
In the context of this case, it is unnecessary to reach the question whether testing requirements that take into account capability for the next succeeding position or related future promotion might be utilized upon a showing that such long-range requirements fulfill a genuine business need. In the present case the Company has made no such showing.
Sfc # ⅝! Sfc ⅜! #
[G]ood intent or absence of discriminatory intent does not redeem employment procedures or testing mechanisms that operate as “built-in headwinds” for minority groups and are unrelated to measuring job capability.
*917Congress has placed on the employer the burden of showing that any given requirement must have a manifest relationship to the employment in question.
Id. at 431-32, 91 S.Ct. at 853-54 (emphasis added).
I conclude that the difference in approaches of the Eighth Circuit Court of Appeals in Christensen (1977) and Craik (1984) traces to an intervening event, the 1981 Gunther decision. The Christensen approach is no longer valid, as a concurring judge preordained. Under current law, market-factor wage setting is not always appropriate, as Christensen suggested, but is appropriate only when necessary.
Here, the evidence shows that use of market information perpetuated and enlarged inequities among classes of employees that were completely or dominantly male or female. The trial court found that Hartz wages “reflect the market place.” There is no finding of fact that this practice was needed for recruitment, employment stability, or some other business purpose.
There is very little evidence in the record tending to explain why Hartz used market information. A corporate officer said that Hartz had an interest in being competitive “to hold our employees.” It was not shown that Hartz had a turnover problem due to wages for any category of workers. It was not claimed that there was a risk of such a problem. There was no evidence of an employee shortage for any kind of job.
D. Gunther.
Respondent’s witnesses repeatedly testified that market information led them to put a “ceiling” on the wages for female clerical workers. There was no evidence showing a legitimate business purpose for setting this ceiling. Two salaried employees were also said to be subject to a wage ceiling at the end of 1980; they were both among the top paid corporate employees.
The ceiling device, displacing other measures of fair wages, is a matter of fact like the wage-setting practice which prompted the decision of the United States Supreme Court in Gunther, 452 U.S. at 180-81, 101 S.Ct. at 2253-54. There the employer disregarded its own wage standards when setting the salaries of female matrons. Because of the testimony that market information was used to shape a ceiling, an inequitable lid on the wages for one class of employees, the market factor enlarges rather than diminishes the appearance of wage discrimination.
E. Other Ideas, including “Comparable Worth.”
The issue discussed here is complex, and additional points need to be explored.
1. Ninth Circuit. The conclusion here conflicts with some appellate decisions. For example, the Ninth Circuit Court of Appeals has concluded since Gunther that the market-factor in wage setting is a “given” for an employer, not part of its policy-making, so that the employer cannot be held responsible for its disparate impact. Spaulding, 740 F.2d at 708. A concurring judge discounted this analysis. As stated above, a different view has been announced in the Eighth Circuit, one that is reasonable, and I would decide this case according to that view.
2. Human Rights Policy. Limited acceptance of market factors is consistent with the broad purposes of Title VII and the Minnesota Human Rights Act. See Westinghouse, 631 F.2d at 1097. See also Griggs, 401 U.S. at 431, 91 S.Ct. at 853. It is a deception to aim at ending discrimination but to concede that all are exempt who can show that their decisions “reflect” the social or economic practices in the community.
In Danz, the Minnesota Supreme Court demonstrated a thorough and demanding approach to sex-based wage discrimination claims. It required, among other things, that a trial court “carefully scrutinize” reasons for action posed by an employer when wage partiality for male salaries is apparent. 263 N.W.2d at 402 n. 3. The approach of the Danz court signals the thorough judicial analysis that is so commonly needed to properly handle pay equity litigation.
3. “Comparable Worth.” According to the preceding analysis, we must disregard unnecessary sex-based market factors. *918Without deferring to the market, do we necessarily engage in judicial measuring of “comparable worth”? That would overextend the immediate effects of Gunther. See 452 U.S. at 166, 101 S.Ct. at 2246.
The majority concludes that this opinion adopts a comparable worth approach. The merit of the dissent rests heavily on a careful statement about this concept.
It is presently established, based on existing statutory law,6 that appellant could not recover on a demand for compensation based on the “intrinsic worth” of her work and the work of males. That is indicated by Gunther. Id. It is helpful and appropriate to designate that concept as one of “pure comparable worth.” Power, 539 F.Supp. at 723.
Gunther made one other thing clear. It established that there is a territory between comparable worth, at one pole, and equal pay for equal work, on the other pole. That territory has not been fully explored and charted, but it is certainly there. An earlier part of this opinion ends in the conclusion that appellant showed cause for a remedy that falls in the field of cases recognized in Gunther. Significantly, one of the appropriate indicators of such a cause is a practice of setting wages without use of objective criteria.
Into this picture enters the specter of market factors. It is asserted that wages can be set according to a sex-slanted market. If that were true, the territory newly recognized in Gunther is fenced off. Legitimate remedies are denied as firmly as are pleas for compensating comparable worth. The result is to disregard Gunther. More to the point, the devastating result is the erroneous assessment of valid causes of action by clients, lawyers and judges.
The analysis here has meaning for both sides of these disputes. An advocate for pay equity cannot come to court with nothing more than a banner calling for equal pay for comparable worth. See Power. The advocate must identify and prove the telling signs of any claimed discrimination.
On the other hand, an employer’s defender cannot take cover under a banner warning the court about comparable worth, and it is no different to rest alone on evidence of a market survey.
While further issues will need to be decided before the law is fully clarified in these difficult cases, it is abundantly clear that employers and courts have responsibilities to address wage discrimination when it exists, responsibilities that cannot be avoided by the simple expedient of referring to a local wage survey.
VI. RETALIATORY DISCHARGE.
• Claims of retaliatory discharge are governed by Minn.Stat. § 363.03, subd. 7 (1984). The courts are to analyze these cases with the same three-part approach utilized in other discrimination cases under Title VII. Hubbard, 330 N.W.2d at 444. The Hubbard court described the ingredients of a retaliatory discharge:
In order to establish a prima facie case where an alleged retaliatory discharge is involved, an employee must establish: (1) statutorily-protected conduct by the employee; (2) adverse employment action by the employer; and (3) a causal connection between the two.
Id. at 444.
A. Prima Facie Case.
The trial court erroneously identified the statutorily-protected conduct of Treva Bohm. Thus, the court failed to consider a causal connection between that conduct and appellant’s discharge.
In a memorandum, the trial court said: She has not made a showing that she was discharged from her position with the defendant in retaliation for her filing of charges with the Equal Employment Opportunity Commission. Defendant was not even aware of plaintiff’s complaint when the decision to discharge her was made.
Treva Bohm’s adamant opposition to the salary practices of her employer was made *919known over a period of weeks beginning in December 1980. She initiated a formal complaint in April 1981, after she was discharged.
The statute provides:
It is an unfair discriminatory practice for any employer * * * to intentionally engage in any reprisal against any person because that person:
(1) opposed a practice forbidden under this chapter or has filed a charge * * under this chapter.
Minn.Stat. § 363.03, subd. 7(1) (1984) (emphasis added). It was error to address the reprisal issue with regard only for the filing of a complaint. The trial court had to make ultimate findings on the relationship between Bohm’s discharge and her opposition to company wage practices.
A review of the record indicates that appellant succeeded in making a prima fa-cie case of retaliatory discharge.
It is undisputed that Bohm actively opposed company wage practices, and that she made known that her opposition had to do with the effect of those practices on females.
During her pattern of opposition, or within a very short time thereafter, her hours were reduced. In Hubbard, the Supreme Court said:
It is recognized, however, that this causal connection may be demonstrated indirectly by evidence of circumstances that justify an inference of retaliatory motive, such as a showing that the employer has actual or imputed knowledge of protected activity and the adverse employment action follows closely in time.
Id., 330 N.W.2d at 445. Appellant was laid off within three months after she first expressed her opposition to company wage practices, and the evidence shows that her opposition continued for at least several weeks after it began. Her hours were halved earlier, apparently within the first month after her opposition began.
B. The Employer’s Rationale.
The trial court found that appellant’s hours were reduced “because she could no longer be kept busy.” That finding is clearly erroneous. It is based on the unsupported testimony of a witness who merely said that there was a “shortage of work.” Nothing was said regarding the work available for Treva Bohm. The witness’s statement was not explained nor documented.
The trial court found that discharge of appellant occurred “as a result of Defendant losing three major grocery accounts.” This finding was also clearly erroneous. The imprecise decision of the employer for cutbacks, based on anticipated revenue losses, was abundantly satisfied by major savings in the rental of computers and the purchase of insurance. Treva Bohm was the only full-time employee who was permanently laid off, saving the company .004 percent of its revenues. She was laid off a month before predicted revenue losses. This is not adequate evidence of a legitimate cause for discharge of appellant independent of the prima facie connection between her discharge and her opposition to company wage practices.
The evidence compels a finding that the company’s expense cutbacks early in 1981 were “pretextual” regarding the termination of Treva Bohm’s employment. Hubbard, 330 N.W.2d at 445; Danz, 263 N.W.2d at 402-03. The evidence requires a finding that the discharge of Treva Bohm was more likely than not a reprisal for her opposition to discriminatory wage practices.
CONCLUSION
Appellant succeeded in establishing a pri-ma facie case that her wages were inequitable due to discriminatory practices toward women in her job category, and patterns of discrimination in the wage setting practices of her employer. Respondent’s only attempt to prove legitimate reasons for its suspect practices was its proof about a local market survey. The evidence on use of market information does not rebut appellant’s prima facie case; in fact, it enlarges the inference of discrimination. Further, the employer’s practices are not excused by showing its use of market in*920formation, where it cannot be shown that the use is necessary. In addition, appellant succeeded in proving a retaliatory discharge, independent of the pretended cause for her layoff.
Summarizing, Treva Bohm witnessed an unlawfully inequitable wage arrangement. She saw raises occur that enlarged the inequities. She vigorously objected. As a result, she lost her job. She is entitled to compensation.
The matter should be remanded for determination of damages, principally including those appropriate for loss of employment.
. Over the dissent of one judge, the Craik court also held that an employer has the ultimate burden of persuasion, to show "that it is more likely than not” that the employer did not unlawfully discriminate, in those cases where the plaintifffs] make a prima facie showing of a "pattern or practice of discrimination.” Id. at 470. Appellant’s case is strengthened by evidence of a pattern of discrimination, but the conclusions stated in this opinion do not depend on a decision that respondent had the ultimate burden of persuasion.
. The trial court found that "clerical" employees received a 5.97 percent wage increase. The evidence shows this rate only for the women in the unskilled clerical category.
. Appellate decisions have infrequently noted the fact that the Bennett Amendment dealt only with sex-based discrimination, and that the equal pay/equal work doctrine never existed for cases of discrimination on the basis of race or religion. The importance of this point was highlighted by the Third Circuit Court of Appeals in Westinghouse, 631 F.2d 1094, as follows:
Under the applicable law it is clear that Westinghouse could not create job classifications whereby different wages were paid to one group solely because of considerations of religion, race or national origin. The statutory issue here is whether Congress intended to permit Westinghouse to willfully discriminate against women in a way in which it could not discriminate against blacks or whites, Jews or Gentiles, Protestants or Catholics, Italians or Irish, or any other group protected by the Act.
Id. at 1096-97 (footnotes omitted).
As an example, it is clear that Title VII prohibits an employer from paying more per hour to welders than plumbers if the reason for the employer paying higher wages to the welder is that the majority of the welders are Protestants and that the majority of the plumbers are Catholics. In such a case an employer would be "classify[ing] his employees ... in [a] way which would deprive any individual of employment opportunities [high wages] .. because of such individual’s ... religion." 42 U.S.C. § 2000e-2(a). While Westinghouse presumably would not challenge the illegality of the scheme outlined above, it asserts that the scheme would be permissible if the reason for the wage disparity is that the majority of welders are men and the majority of plumbers women.
id. at 1100 (deletions, additions and emphasis in original).
. See footnote 1.
. The question discussed here is whether a device which operates to discriminate, which has disparate impact, is nevertheless justified. The analysis must also focus on the proof of improper impact; here it is necessary to reflect on the earlier discussion of factors showing discrimination. To evaluate disparate impact, those factors may be significant even if they do not prove an intent to discriminate. Heagney v. University of Washington, 642 F.2d 1157, 1163 (9th Cir.1981). On the other hand, the evidence must be sufficient to prove disparate impact, not merely to create an inference of it. Johnson v. Uncle Ben's, Inc., 657 F.2d 750, 753 (5th Cir.1981), cert. denied, 459 U.S. 967, 103 S.Ct. 293, 74 L.Ed.2d 277 (1982). The discrimination factors here, which I have said are sufficient to establish a prima facie case of disparate treatment, also suffice to ultimately prove discriminatory results of the market survey approach, both as to wages paid and the termination of appellant’s job.
. This case is not affected by Minnesota's 1984 equitable compensation statute, which benefits employees of political subdivisions of the State. Minn.Stat. §§ 471.991-.999 (1984).