I am pleased to share my latest post to Bloomberg Law.

n the context of sexual harassment, we all are aware of the hashtag #timesup. Well, the same is true for pay inequity and pay discrimination.

In 2018, a majority of states have considered bills to address pay inequity. A number of these bills passed and we can expect even more activity in 2019. The number of lawsuits alleging pay discrimination has also increased. For example. in 2018 alone, the Boston Symphony Orchestra, Five Guys, Nike, and Spotify were among the many companies sued for pay discrimination.

For legal, business, and fairness reasons, employers need to get ahead of the curve and conduct an analysis to make sure they are paying fairly. But a pay equity analysis is deceptively complex.

Below are eight of the more salient mistakes/misconceptions I have observed when companies do pay equity evaluations and ways to address them.

1. Initial Assessment Not Under Privilege

An analysis to identify potential pay inequity later may be used against your organization to prove bias. So, consider analyzing your pay parity analysis under the attorney-client privilege.

However, even if structured and administered properly, a privileged investigation does not necessarily mean everything is privileged. For example only, the fact and scope of the analysis may not be privileged.

So don’t do an analysis, unless you are prepared to act on it.

2. Not Using an Expert

Many organizations have individuals who are technically qualified to do a multi-regression analysis, the heart of the pay equity analysis. But you want someone with deep expertise to guide you on what is a deceptively-complex process.

For example only, you will want to look at not only total compensation but also each element separately, including variable compensation where inequity may be more likely to exist. The variables to be held constant will differ depending on the element of compensation. Plus, there is the matter of “sequencing.” Not sure of the nuances of sequencing? I rest my case on the value of an expert. Plus, if there is litigation, relying on a strong expert has obvious benefits.

3. Looking Only at Jobs That Are Exactly the Same

Unlike the federal Equal Pay Act, many recently-enacted state pay equity laws do not focus only on pay equity where the jobs are the same. They require comparisons of jobs that are substantially similar, comparable or some other similar standard. This is true, for example, in California, Maryland, Massachusetts, New Jersey, Oregon, and Washington.

But what about jurisdictions where there is no pay equity law that goes beyond looking at pay relative to the “same” job, at least not yet?

To minimize claims (even if you prevail) and to maximize true equity, employers may benefit from looking at some jobs that are substantially similar or comparable, even in jurisdictions where it is not yet required. At a very minimum, consider looking at single incumbent positions, where employees may make the comparisons on their own.

4. Focusing Only on Gender

Federal and state anti-discrimination laws cover pay bias. And, these laws apply to a broad range of protected groups, not just gender. Further, the thee most recently-enacted pay equity laws—New Jersey, Oregon and Washington—apply to all protected groups under the state’s applicable non-discrimination law.

So, don’t focus only on gender. At a very minimum, employers should consider race, national origin, and age, too.

5. Relying on Defenses That Are Not Defenses

The fact that, as a result of the multi-regression analysis, there is a statistically-significant disparity does not mean that there is unlawful bias. It means the employer must evaluate whether there are legitimate, non-discriminatory reasons for the “outliers” (both high and low, not just low).

Some possible “defenses” include performance, seniority, and experience. But employers cannot simply pluck a defense listed in the applicable law.

Employers must ask: was the defense a factor considered, and applied consistently, in the process of determining this aspect of the employees’ compensation?

6. Assuming Good Statistics Means No Bias

It is dangerous to assume that no statistical problems mean there are no pay parity problems. A deeper dive is recommended.

For example only, there may be a perfect correlation between merit and performance evaluations, but that does not mean the performance evaluations are not tainted by conscious or implicit bias. How to address bias in performance evaluations and other factors that may include implicit or conscious bias is also complex.

Employers may be inclined to centralize and institute strong guardrails to limit the discretion that may lead to bias with regard to performance appraisals, discretionary bonuses, etc. But the centralization and guard rails may create the commonality of which class actions are borne. Employers need to walk the razor’s edge to avoid individual bias without creating an easy argument for certification of the class action.

7. Making Unnecessary Admissions in Taking Corrective Action

Let’s assume there are some low outliers that cannot be explained by legitimate reasons and need to be increased. The disparity could have been caused by bias. It also could have been a good faith mistake, a bad manager, etc. The reality is that, in most cases, you really won’t know the cause for sure; you will “know” only there is a problem to be corrected.

So make corrections without admissions that may not be true and invite back pay claims. Focus on the need for a change without speculating as to the cause of the inequity.

8. Documenting Outside of Privilege

If there is a pay equity challenge, you may want to rely on your analysis. If your initial analysis was conducted under privilege and you waive the privilege, how far does the waiver go? A court has discretion to hold that the desired discrete waiver of privilege results in a waiver over the entire subject matter of the analysis.

How do employers mitigate this risk? When the analysis is done under privilege, employers should consider re-running the final analysis and document the final decisions outside of privilege so that the employer can rely on non-privileged information in defending any challenge. This critical step, so important in protecting the privilege, is often missed.