Article | Jul 2020
Does Your Organization Have Hidden Pay Bias?
Actionable steps companies can take to uncover any existing systemic bias in their compensation administration.
A quick Google news search on “pay equity” reveals the breadth and depth of a theme that has captured the public’s attention. On any given day, multiple high-profile publications are likely to offer stories on the gender and racial pay gap, pay equity and discrimination, and diverse representation at the management and board level. Companies cannot remain insulated from this external pressure on gender and racial issues and there is clear risk from inaction. But given the vast array of topics and nuance, what is it exactly that should be of primary concern and where can human resources get involved and make a difference?
While all of the topics have weight, we suggest HR teams begin by focusing on compliance with the principles of pay equity, which has been the subject of renewed emphasis with recent legislation at the state level. Understand if and where you may have pay differences that are—from a statistically significant standpoint—potentially attributable to gender or race and should be remediated.
The various federal laws mandating equal pay for equal work have been in place as early as 1963 and are well known, particularly among HR professionals. While most states have had some form of pay equity legislation for some time, some recent state legislation has broadened the definition from equal pay for equal work to equal pay for comparable work.
While this broader definition clearly applies to companies with operations in certain states, it may be a precursor to actions in other states. As employers address pay equity within the context of a broader definition, they face questions including:
- What exactly is comparable work or in other words, what is the definition of comparable?
- What are acceptable factors that allow the company to pay differently?
How do you determine if there’s a problem?
The move from equal pay for equal work to equal pay for comparable work somewhat complicates comparisons and any assessment exercises must broaden in scope. Previously, you would ensure the pay of a male Accountant II and a female Accountant II or a Black Accountant II and a White Accountant II is the same, assuming all other factors such as education, career tenure, performance, level of responsibility, etc. are equal. Under the new legislation in many states, now you may need to also assess whether a janitor and a food service worker are on equal footing, all other factors being equal, and this requires more thought about the nature of the job duties, the requirements for holding the position and being successful in the position, the variations in skillset, etc.
To determine if your company has an issue with pay equity, it might seem reasonable to aggregate all employees in a certain pay grade, average the salaries for sub-groups (male versus female; white versus underrepresented minority), and calculate the variance. You may find that one sub-group is paid less, but this simple calculation doesn’t explain why and can’t accurately tell you if it’s attributable to gender, race, or simply coincident with these factors. Unfortunately, an accurate analysis is not so straightforward.
What does work well is a statistical model rather than a mathematical one. Multi-variate regression analysis allows you to take many factors into account at the same time. For example, such an analysis can examine a range of employee-specific factors, such as gender, race, tenure with the company, related job experience, performance, and specialized skills, as well as job-related factors such as level of responsibility, functional area of focus (e.g., accounting, software engineering), and level of management responsibility.
Understanding the role of each of these underlying factors taken as a whole will provide a much more accurate and reliable understanding of inevitable pay variances. The regression model will determine which factors (or independent variables), including gender and/or race, have a statistically significant impact on pay (or dependent variable). Regression results provide an R-square, which is an indication as to how much of the variation in pay can be explained by those factors used in the model. For instance, an R-square of 0.90 indicates that the factors used in a regression model are able to explain 90% of the variance in pay—which is significant. R-square results that are 0.70 or less suggest that a meaningful amount of the variance in pay cannot be explained by the identified factors, which places more burden on an organization to rationalize pay variances.
The regression model will also identify any stark outliers, which can then be further examined for potential bias. These outliers provide you with further opportunity to ensure pay levels are appropriate, which usually requires further assessment of such things as the identified employee’s job-related experience, total years of work experience, record of performance, and specialized skills. It is important to note that these outliers often consist of females, under-represented minorities, as well as men.
How often should you assess your company’s position?
If your company has not already conducted a statistical analysis, it is a wise risk mitigation endeavor to do so now. Having a clear understanding of your current position, valid reasons to explain any anomalies, and a well thought out (if only reactive) communication plan will help shield the company from legal and reputational risk. To that point, be sure to bring in your legal team, general counsel, or external counsel for protected internal communication as you plan, conduct, and analyze your study.
In general, organizations should conduct a pay equity analysis every three years. However, there may be factors that would suggest more frequent analyses are advisable. For example, any significant corporate activity like a merger, spin-off, or significant organizational changes such as new business units or a reduction in force would necessitate a new round of modeling. It is best, however, to do this after the dust has settled and pay systems are integrated or updated to reflect the organizational changes.
What if there’s an issue?
After helping numerous companies prepare and execute their assessments, not surprisingly we have found a direct relationship between the extent of unexplained pay variations and the state of an organization’s compensation program and processes. For instance, we have found that those with a significant portion of pay variances that can be explained by relevant job and incumbent factors (e.g., higher R-squares) usually have well defined jobs and career progressions, up-to-date pay ranges, and clearly defined pay management policies. Those with a higher percentage of pay variances that are not explained (e.g., lower R-squares) typically have programs that are not as well maintained. The good news is that adjustments can be implemented as appropriate and issues with your compensation administration can be addressed.
The fact is we don’t see systemic bias occurring very often. But in almost all cases there is a sub-set of the population that appear to have pay out of sync with their skills, experience, and performance and remediation actions should be undertaken.
Strong processes and salary administration systems help to combat systemic bias and include these characteristics:
- Detailed, accurate HRIS databases;
- Clarity around job structures and hierarchy;
- Clearly defined pay ranges that are aligned with an organization’s array of jobs;
- Ensuring formal pay-setting policies (e.g., new hire, promotion); and
- Rigor in merit increase planning.
What’s likely to happen going forward?
The attention on pay equity and discrimination issues is not going to ease and has in fact sharpened in recent months. More specifically, the media will continue to conflate pay equity and the pay gap—two vastly different problems with very different solutions. For public companies, proxy advisors are jumping in to make assessments and shareholders will continue to put forth resolutions related to gender, race, and pay, taking their cue from Arjuna Capital which is targeting high-profile industry leaders. More corporate giants will join in with Bank of America and Google to voluntarily disclose results of pay equity and/or pay gap assessments and remediation plans. And the wave of legislation will continue, as evidenced by active or proposed board diversity laws, the pay gap disclosure mandate in the UK, and expanded EEOC reporting requirements.
An accurate assessment of your company’s pay equity position will not absolve it of broader responsibilities or alleviate scrutiny, but it is a critical first step best undertaken proactively and thoroughly.
 The Equal Pay Act of 1963; title VII of the Civil Rights Act of 1964; the Lilly Ledbetter Fair Pay Act of 2009