In the past decade new sociological research challenges the assumption that employers are mere wage takers when determining compensation. In some cases internal institutional constraints and politics play a greater role than we (and the courts) are willing to acknowledge. Although employers are taking into account prevailing market wage rates for the relevant positions at issue42, they are also influenced by internal power dynamics and patterns of conflict vis-à-vis different groups of employees. This may result in managerial decisions to compensate some groups or occupations beyond the rate warranted by external market constraints.
Although this research targets gender-based occupational segregation, it is also relevant to the general discussion about the validity of the market defense. The market defense draws its strength from the assertion that wages are determined by external forces. If this premise is put in question by empirical analysis, employers can no longer claim that they are not actively participating in the setting of wages. If in fact they have some input into the process of determining wages, and are consciously deciding to adhere to demands of specific groups of employees by compensating them above the market rate and this decision results in pay disparity between men and women, then this could be conceptualized as sex discrimination. At this point pay disparity would not be the product of adhering to the external market valuation, but rather following some internal process of institutional decision-making. If this is the case, the market cannot be blamed for the wage disparity across individuals or groups of employees.
In “Legalizing Gender Inequality: Courts, Markets and Unequal Pay for Women in America”43 sociologists Robert Nelson and William Bridges, conduct a thorough qualitative investigation, which they term “critical empiricism” into four high profile comparable worth cases.44 They set out to examine whether market forces, as argued by the defendants and accepted by the courts, were the cause of the pay disparity across gendered occupations. As students of organizational behavior they are interested in understanding how organizations unconsciously effectuate gender inequality.
In one of the cases, Christensen v. Iowa45, female clerical workers filed a pay discrimination lawsuit against the University of Northern Iowa. Their claim was that they were underpaid in comparison to male physical plant workers, even though both occupations were assigned identical pay grades in an internal job evaluation report. Specifically, the report stated that physical plant workers were overpaid relative to clerical workers, and that pay to the physical plant workers was inflated in comparison to external market pay rates for similar jobs. The university chose not to implement the recommendations. It sincerely feared that it would not be able to attract physical plant workers at the wage rate it was paying the predominantly female clerical workers.46 The market justification was not presented as an excuse for conscious sex discrimination, a deliberate decision to pay male occupations more. But rather an unconscious process in which internal power structures affected how management perceived market rates for the gendered jobs.47 The university’s perception that it could not lower the current wages of physical plant workers was based on the organizational strength of that group of workers. “The Physical plant workers, informally known as the “meatpackers”, because of their identification with unionized workers engaged in self-conscious collective bargaining. The clerical workers, in contrast, were content to participate in amicable “committee” meetings with university officials.”48 This led the university to “worry about the union guys but not the women”.49 The authors argue that while the employer was speaking in terms of the market to justify its pay scheme “the market did not compel the university’s decision. Organizational politics compelled the university to give selective attention to the demands of workers in predominantly male jobs”.50 This study, as well as others,51 demonstrate that even if we generally embrace the market defense, i.e., accept wage disparities resulting from external market forces, we may encounter circumstances where what initially is deemed an external factor, after close examination, turns out to be an internal institutional decision unrelated to external market wage rates. Since discerning which cases carry a true market component and which reflect institutional politics and structures is a tricky matter, we should be extremely suspicious of any market claim presented.
D. Current Treatment of Market Justifications is Inconsistent Paying women less than men just because the employer realizes he can hire women for less pay than men is discriminatory.52 This type of behavior falls neatly within the traditional framework of antidiscrimination theory. The employer is basing employment decisions, in this case about compensation, on the applicant’s or employee’s gender, offering less pay because an applicant or employee is a woman. The causation requirement is clearly met, despite the employer’s ability to differentiate compensation based on factors external to his operation. On the other hand, when an employer is basing compensation decisions on factors other than sex, but these factors are not related to ability or productivity, the law is ambiguous. In that situation the implicit instruction of the courts is that as long as the employer did not resort to sex based classifications he is on safe grounds. The employer is free to construct whatever wage structure he desires, extracting profit from the fact that reservation wages, compensation expectation, and negotiation skills vary among the pool of similar job applicants.
But relying on a market justification can result in both scenarios in gender wage disparities. In Both Corning Glass Works v. Brennan53 and Christensen v. Iowa54 the employer explained his wage structure in terms of external market constraints or opportunities (to pay women less than men). In both cases external market valuations were blamed for generating the internal gender wage disparity. The reason the University of Iowa prevailed where Corning failed has to do with the source of the employer’s own classification. Corning confessed it classified employees based on sex, while the University of Iowa admitted only to classification based on occupation. Usually, employers arguing that the source of the pay disparity is classifications or factors other than sex can successfully defend their market explanation.
This doctrinal inconsistency in addressing market arguments is driven by the centrality of the causation model. Only in cases, such as Corning Glass, where evidence that a causal link between the sex of the plaintiffs and the compensation decision exists are the courts willing to dismiss the market defense.55 In other cases, resulting in similar pay discrepancies between men and women, the market justification is given credence.56 After all, no causal link between the compensation practice and the sex of the plaintiffs was articulated, and the courts respect the employer’s apparently nondiscriminatory business judgment.57 The similarity between these two scenarios warrants parallel treatment of the market defense. In both cases the employer is not driven by animus toward women, but rather profit maximization considerations. In both cases women’s wages are adversely affected by a compensation policy, contributing to the gender wage disparity. The fact that in the latter case one cannot pinpoint the causal process that connects the decision to gender should not serve as a strong basis for sanctioning market arguments. The rationale of Corning Glass Works, that an employer cannot pay women less than men for performing the same work, should govern all circumstances of market driven wage disparities.
III. The Difficulty of Using Traditional Discrimination Models to Confront Market Justifications A. The Traditional Discrimination Framework: Emphasizing Causation The market-defense controversy presents a challenge to employment discrimination theory due to the centrality of causation in conceptualizing discrimination in American law. Fundamentally discrimination is understood to be the action of treating people differently on the basis of some prohibited group classification.58 In the case of sex discrimination it is taking action based on gender. The causal link between the scrutinized action and group membership has to be articulated.
Title VII embraces the causation requirement both in its disparate treatment model and disparate impact model. The causation requirement is salient in disparate treatment theory. The ultimate question in any disparate treatment litigation is whether a plaintiff was able to meet her burden of persuasion and demonstrate that sex (or any other regulated group membership category) was a motivating factor in the employment decision.59 There are a couple of ways to meeting this burden. A plaintiff could resort to direct evidence or circumstantial evidence (including statistical evidence).60 But at the end of the day, she must convince the fact finder that her gender was a motivating factor in the final decision regarding compensation, denial of promotion, termination or sexual harassment.
In disparate impact theory, the centrality of the causation requirement is covert, but nonetheless present. Under disparate impact theory61 a plaintiff argues that an employment practice adversely impacts members of a protected group. This is a statistical claim. The claim is that there is a strong statistical correlation between group membership and the employment practice. If no data on statistical correlation are offered, the employer is not required to defend his decision making process.62 Since disparate impact is based on providing statistical data on the correlation between employment practices and group membership, it also entails a causation requirement. Although the plaintiff is relieved of the requirement to demonstrate that employment actions were based on group membership, she is still obliged to demonstrate the statistical correlation (impact) between employment practices and group membership.
B. Limitation of the Causation Model in Addressing Gender Wage Disparities i. Disparate Treatment Law Disparate treatment law is ill suited to deal with gender pay disparity resulting from market considerations. An employer can avoid liability if he can persuade the fact finder that pay decisions were solely based on other factors than the employee’s gender. Offering market justifications for the decision-making severs the causal link between the sex of the plaintiff and the lower compensation level. A genuine external market justification does not fit within the causation model described. It is not because of the plaintiff’s sex that she was paid less than another co-worker performing the same work, but rather due to another factor unrelated to sex, such as the fact she did not demand annual raises, did not negotiate her initial salary or any other factor which is not regulated by law.
As discussed above, only in the limited cases in which plaintiffs can establish causation between their sex and wage determination will they prevail in disparate treatment litigation. Corning Glass is of limited application. Presently most employers do not intentionally pay women less than men, and if they consciously take advantage of the fact that women are willing to work for lower wages than comparable men, they take extra care not to reveal their motivation. Arguments that employers basing compensation decisions on factors such as prior salaries are engaging in intentional sex discrimination because those prior salaries reflect sex discrimination in the labor market have also failed under disparate treatment law, unless a plaintiff can offer specific evidence that the previous employer engaged in discriminatory practices is setting wages.63 Ambiguity of the market justification and obstacles in verifying its existence hinder the success of sex discrimination claims based on disparate treatment. The sociological studies revealing salience of intra institutional structures and internal firm politics to wage determination could possibly aid prospective plaintiffs. If the claim is that the employer is responsible for setting the lower pay for women performing equal work of male co-workers and that this process is independent, or only loosely dependent, on external market constraints, one can possibly meet the causation requirement that the employer is basing his decision on sex (and not the market). But as the empirical studies illustrate, providing the background information for such a factual claim is an onerous, time consuming and expensive task. In individual claims this may not be worth the cost. In group-based claims, such as a class action, the investment in collecting information, hiring expert witnesses, and laying out the argument of intentional internal practices rather than market driven disparities may be an economically sensible decision.
Plaintiffs could present social framework evidence64 to support their claim that underlying sex discrimination was a motivating factor in compensation decisions. Such evidence was successfully utilized in high profile discrimination cases such as Price Waterhouse v. Hopkins65 and Dukes v. Walmart Inc.66 In social framework testimony an expert witness can explain how “general research results are used to construct a frame of reference or background context for deciding factual issues crucial to the resolution of a specific case.”67 In the context of employment discrimination litigation it usually is offered to “educate fact-finders about the conditions under which gender stereotypes and prejudices are likely to influence impressions, evaluations and behavior in social and organization settings.”68 When using social framework testimony to explain how employers are stereotyping women according to gender roles, the argument is clear: women are not expected to behave like men, and those that do are punished for this cross gender behavior.69 This satisfies the causation requirement that the disparity in treatment was “because of sex”. When we enter the domain of internal firm processes affecting terms and conditions of employment we are on less stable grounds.70 While the theory that the employer is partially insulated from external market forces is easy to articulate, providing evidence in specific cases that this amounts to sex based discrimination is more difficult.
Take for example Christensen v. Iowa71 where an expert witness explained that the university was basing its compensation scheme on an unfounded perception that it could not hire physical plant workers at lower rates than it was currently paying. But since this perception was based on the internal pressure this group of employees was exerting on management, this still does not necessarily lead to the conclusion that it was paying them more because of their sex. Perhaps it was just their organizational power, and the fact that they were men was incidental. The expert testimony could only refute the employer’s offered explanation that external market forces warranted the higher wages for the male dominated occupation. This does not amount to proving the causal link between sex and lower wages for women.72 Disparate treatment theory places the ultimate burden of persuasion on the plaintiff.73 The fact that the plaintiff was able to discredit the external market defense with expert testimony does not mandate liability.74 The fact finder must be convinced that sex was a motivating factor in the compensation decision.75 Not all internally driven compensation schemes that result in lower wages for women performing the same work as men satisfy this condition.
ii. Disparate Impact Law At first blush disparate impact law seems an adequate means of handling the market defense. Disparate impact law is all about identifying and then scrutinizing the business relevancy of neutral employment practices that adversely impact a protected class.76 In theory this is the vehicle to examine whether market-based practices are harming women, and if there is an adverse impact to assess whether there is a legitimate management interest in continuing such a market-based practice. However, disparate impact proves to be an unfaithful servant to wage equality because of doctrinal and practical issues.
The first problem is that it is unclear whether disparate impact theory is available for plaintiffs claiming gender based pay discrimination. The Bennett Amendment, a coordination clause between the EPA and Title VII, states that there will be no Title VII violation if the compensation differentiation is “authorized by the provision of section 206 (d) of title 29,”77 the EPA. There are two possible interpretations to the Bennett Amendment. A broad interpretation would preclude a finding of violation in a case of gender based pay discrimination, unless the practice would have violated the Equal Pay Act as well, including the restrictive condition of “equal work”. A narrower reading of the amendment would incorporate only the four affirmative defenses of the EPA78 In Washington v. Gunther79 the Supreme Court opted for the narrow interpretation, enabling plaintiffs basing their wage discrimination claims on Title VII to avoid proving the jobs compared involved “equal work”. But the Court did explain the amendment could have significant consequences for Title VII litigation, on account of the fourth affirmative defense to an EPA claim. This defense states the employer can justify pay disparity resulting from “any other factor other than sex” (AFOTS defense). The court intimated but did not decide whether the AFOTS defense undermines Griggs type disparate impact analysis under the EPA, and by the Bennett Amendment, under Title VII as well.80 The Court did not conclusively decide this issue in Gunther.81 On this judicial intimation several jurisdictions have interpreted the EPA and Title VII as restricting sex based compensation discrimination claims to disparate treatment type analysis.82 Other courts resisted, continuing to apply disparate impact analysis to both EPA and Title VII claims.83 In Smith v. City of Jackson84, an age discrimination case that looked into the applicability of disparate impact to the Age Discrimination in Employment Act (1967)85, the Supreme Court noted in a footnote “if Congress intended to prohibit all disparate impact claims, it could have certainly done so. For instance, in the Equal Pay Act, Congress barred recovery if pay differential was based “on any other factor” – reasonable or unreasonable – “other than sex”.86 Again, there is a strong suggestion, albeit in dictum, that the EPA’s fourth affirmative defense effectively rules out disparate impact87.
For individual claimants, the disparate impact course of action may prove too expensive. It is not sufficient to argue that specific practices disparately impact women compared to men, the plaintiff has to offer statistical data supporting the claim which requires collecting data and hiring experts to conduct regression analysis. Take for example, the argument presented in section II.B that variance between the sexes in negotiation skills and styles can adversely impact the compensation of women. General reference to the negotiation literature on this topic will not satisfy the requirement of statistical proof of disparate impact.88 Thus, a successful plaintiff will have to engage in at least a three step process to establish his prima facie case of disparate impact: (1) identify the particular “market” practice he deemed discriminatory89, (2) collect data from within the organization on how this practice affected the wages of individual employees, and (3) run the statistical regression analysis to show the required disparity between men and women. It is highly unlikely that individuals will find it worthwhile to invest the resources to pursue such analysis.
But even if we focus on class action cases, where investment in preparing a disparate impact claim may prove economically worthwhile, the theory will encounter doctrinal hurdles. Statistical disparity resulting from application of a neutral employment practice establishes only a prima facie case of discrimination. At this point the employer has the opportunity to demonstrate that the challenged practice is “job related for the position in question and consistent with business necessity”.90 This is a relevancy test. In other contexts of disparate impact law the business necessity defense has been interpreted quite broadly.91 Although cost-saving justifications have been rejected in the context of disparate treatment92, usually when an employer raises a cost-saving or profit-enhancing argument for his disparate impact practice, it will qualify as “business necessity”.93 Structuring pay levels in a manner that takes into account employee individual wage expectations or demands may meet the applied standard of Title VII business necessity. The business necessity defense is not restricted to productivity or ability arguments. It is rather a loose-reviewing mechanism, which engages in a balancing act between the interest of protected group members and the interest of the employer. When it comes to profitability or labor cost saving claims, the scale is skewed toward the employer’s interest.94 Under this standard of review, proponents of the market justifications can argue that negotiation techniques which may disparately impact women are lawful.95 If the techniques are aimed at promoting profitability by extracting more of the contractual surplus, they may meet the threshold of the business necessity standard.
Ian Ayers offers the following guidelines for crafting the contours of “business necessity” with relation to the increasing profitability argument96:
“Policies that exploit a firm's market power to extract supra-competitive profits from employees or consumers should not fall within the limits of the business necessity defense in disparate impact litigation. Even though such policies can substantially enhance a firm's profitability, profits that are the byproduct of market failure are less justified than those that are a byproduct of competition. By enjoining employment and consumer policies that extract supra-competitive profits disproportionately from racial minorities and other protected classes, disparate impact law can help make markets both more competitive and less racially discriminatory.”97 Ayers’s contribution is in noticing “not all increments to profitability deserve equal judicial respect”.98 He sketches this vignette:
“An employer pays high-school graduates an amount equal to their marginal productivity but institutes a new policy of paying non-high school graduates less. Imagine that the policy has a disparate impact against African Americans, who, in this hypothetical, are less likely to have a high school diploma. The employer might justify the pay difference by arguing that non-graduates tend to be less productive than high-school graduates. I will call this the productivity defense. In the alternative, the employer might try to justify paying non-graduates less, not because the non-graduates are less productive, but solely because these employees have fewer employment alternatives than high-school graduates. For example, imagine that non-graduates were more tied to their hometown than graduates and hence had fewer work alternatives. I will call this the market power defense, because the lower pay is a function of the employer's greater market power over the non-graduates. Note that both of these defenses are, at core, about profitability. The productivity defense in essence says that the policy enhances the firm's profitability because the employer will be more profitable if it is not forced to pay workers more than their marginal productivity. A firm that pays less productive workers the same as more productive workers will tend to be unprofitable. But the market power defense is also about profitability, because finding a group of workers who will work for a sub-competitive wage is also an effective way for a firm to increase its profits.”99