Wal-Mart class action: So, what did the judge say in his opinion and what are some lessons for employers?

In ruling in favor of class certification in Dukes v. Wal-Mart Stores, the judge wrote an 84-page opinion. In this blog entry, I will summarize and comment on some highlights of this lengthy document.

Key points include: the role of subjectivity in promotion and pay decisions; criticism of Wal-Mart's diversity program; and several aspects of the lengthy discussion of statistics.

In reading this, please keep in mind that (1) I’m not an attorney; my interests lie with the implications of this decision for HR managers and expert witnesses, (2) my purpose is neither to support what the judge said nor is it to criticize what the judge said; (3) I’m basing these comments solely on the judge’s opinion; I have not read other documents in connection with this case, so I’m noting these points based on limited information; and (4) keep in mind that the judge was only ruling on class certification, not making a decision on whether discrimination had occurred.

I would appreciate any comments you might have regarding this commentary. You may use the comments feature below or email me at mharris@umsl.edu.

Role of Subjectivity in HR decisions

In some discrimination cases requesting class certification, the defendants have argued that where managers have substantial discretion in compensation or promotion decisions, it is not appropriate to certify a class because this would require the court or jury to consider the individual facts for each specific decision. Indeed, this appears to be a significant factor in certain courts’ decisions.

Subjectivity in Promotion Decisions. The judge in the Dukes v. Wal-Mart case emphasized two components that made for a lot of subjectivity here: minimal objective criteria in the promotion decisions and the failure to post most of the promotional opportunities. It would appear, from reading the decision, that the promotions at issue were made with almost no guidance from the company. Thus, standard HR practices, such as a job analysis, structured interview practices, tests, and so forth, were not used here.

LESSON: When making promotion decisions, employers should standardize and structure the decisions. Managers and supervisors simply cannot have complete discretion over these decisions. HRM experts have developed tools (e.g., structured interviewing) to help make good promotion decisions. Also, companies should have a mechanism for posting promotions. In the web-based culture in which we live, this should be easy to do.

Subjectivity in Pay Decisions. While there appeared to be more structure in pay determinations, the judge concluded that a great deal of discretion remained. For example, while there was a minimum starting rate for hourly workers, Store Managers were allowed to depart from the minimum within a two dollar range without "objective criteria and with limited oversight." Given the average (2001) pay for hourly workers was $18,000, the judge asserted that 2 hours an hour discretion was "significant." Well, assuming a 2,000 hour work year, this means that on average, hourly workers earned $9 per hour. A 2-dollar range would represent as much as a 22% increase over the minimum rate.

LESSON: There can be too much discretion in pay setting. While some discretion may be acceptable, more than 20% is too much. Companies need to consider more structure in how pay decisions are made (e.g., specify what factors should be considered), and more monitoring is needed.

Diversity Issues

Subjectivity in HR decisions is, by itself, probably not enough to assume that discrimination is occurring. The judge considered testimony by an expert witness to be relevant in this regard. In criticizing Wal-Mart’s diversity record, the judge noted that despite the fact Wal-Mart "earned national diversity awards and its executives discuss diversity and include it in company handbooks and trainings," the company had not identified possible barriers to women’s advancement, and the goals the company had set often seemed "ad hoc" (I think the term "subjective and arbitrary" might capture the judge’s thinking here). The judge also used the term "lip service" in connection with the company’s diversity attempts.

LESSON: There must be a system for assessing problems in minority and female representation in the workforce, and shortcomings must be identified and strategies and tactics identified for addressing problems. This must translate into outcomes. Numbers count!

The Numbers ...

As in most cases of this nature, the numbers or statistics play a central role. There are several interesting features in this case. I’ll separate my discussion of statistics into the following sections: (1) benchmarking analyses; (2) level of analysis; and (3) missing variables.

1. Benchmarking Analyses: The plaintiffs used an expert who compared Wal-Mart’s female representation to other major retailers. Despite the defendant’s criticisms of this technique, the judge looked favorably upon this expert’s conclusion that the shortfall at Wal-Mart "cannot be explained in terms of lack of qualifications, interest, or availability among female employees."

2. Level of Analysis: Both the plaintiffs’ experts and the company’s experts used a sophisticated statistical analysis called "multiple regression analysis." While the details of this technique are well beyond the present scope, I want to point out that different results may be found, depending on the level of the analysis. That is, in the present case, the plaintiffs analyzed pay differences at the regional level (there were 41 regions) and found that women on average earned less, even taking into account various factors (e.g., seniority, performance, etc. see below "Missing Variables"). The defendants analyzed pay differences at the store sub-unit level (of which there were about 7,500) and found that there were very few instances of women being paid less. Without going into the explanation, the judge ruled that the plaintiffs’ expert’s analysis was sufficient to pass initial muster.

3. Missing Variables: A major advantage of multiple regression analysis is that it enables one to remove the effect of other variables (e.g., in this case, tenure, performance, etc.), and isolate the relationship between one variable of key interest (e.g., in this case, employee gender) and the criterion (e.g., in this case, pay). But, in most lawsuits, it is probably impossible to get data on every possible variable that might affect pay. Thus, the defendant in the present case challenged the plaintiffs’ compensation analyses on the grounds that they left out and failed to include such variables as prior grocery experience and leave of absence. In refusing to reject the plaintiffs’ expert’s analyses, the judge made two key points. First, other courts did not require that all possible variables be included in the analyses. Second, the judge noted that the defendant was not able to prove that including those "missing variables" would change the results.

LESSON: Companies should proactively analyze their compensation and promotion systems for possible disparate impact against women and minorities. Analyses should examine data at various levels; where disparities exist, strategies and tactics should be implemented to solve the problems. Benchmarking against companies in similar industries should be conducted to make sure that there are not major discrepancies from other organizations.

CONCLUSION: I look to others to discuss the legal implications of this case (e.g., its impact on similar litigation in other courts). Of course, there are other points here, but I have chosen to comment on what I thought was most interesting.

As noted throughout, I think there are some clear lessons for companies here. Although the judge ruled only on class certification, not liability, his discussion provides considerable food for thought as to the potential vulnerabilities of Wal-Mart's practices when the issue of liability is considered.