Using analytics to close the gender pay gap

Compensation equity has received a lot of attention recently, with a number of states implementing their own pay equity laws.

For example, California, Maryland, Massachusetts, Puerto Rico and New York have enacted or are enacting gender equity laws. In this legal landscape, employers would be well served – for both practical and legal reasons – to take meaningful steps to both measure and reduce gaps in compensation between men and women, and to take steps to ensure that no new problematic compensation gaps arise.

On the legal front, even well-meaning employers who pay their male employees more than similarly-situated female employees– absent a hint of intent – can be held liable under certain state gender equity laws. In some jurisdictions, employers must correct pay gaps that might exist in the labor market, even if they didn’t have any direct hand in their creation.

That is to say, if an employer hires men and women for jobs requiring substantially similar skill, effort and accountability, and pays men and women both 5% more than they were paid at their previous jobs, but the men were paid more than the women in their previous jobs, the current employer’s policy of giving 5% more than new hires’ previous pay could yield a problematic result.

In addition to legal reasons to work to eliminate pay gaps, negative publicity associated with gender pay gaps can negatively impact an employer’s ability to recruit, creating a vicious cycle in which qualified applicants may not even bother applying for a job at an organization with a bad reputation. And, internally, pay imbalances can lead to employee dissatisfaction, which in turn can result in costly turnover.

Unfortunately, absent a voluntary disclosure by a departing employee in an exit interview or other forum, an employer may be unable to discern whether perceived inequity in compensation has contributed to the turnover.

For nearly two decades, working as a data scientist, former law professor and in-house labor and employment law counsel, I’ve helped hundreds of employers to identify and correct gender and race equity issues. Here is some general guidance I’d offer to employers working to navigate this tricky terrain.

Some employers use traditional measures that don’t objectively track skill, effort or accountability to legitimately differentiate compensation. Research has shown how allegedly objective performance evaluations may favor men – which is something that I’ve observed first hand.

Frequently, the performance measures themselves can be biased. Subjectivity can creep in if sufficient care is not taken. In the end, an employer may wind up relying on these measures to account for discrepancies in compensation. They are not necessarily bad, but they may not always be the best available data for the job.

This is why I advise the use of “relational data,” which measure how employees regard their co-workers – e.g., how much they rely on them to get work done, and how critical they are to performance in teams. This is blind-360-degree performance data, instead of top-down single-source information.

When looking for pay gaps, organizations are well-advised to seek outside help instead of relying on internal personnel. Employers sometimes delegate the task of reviewing compensation to someone in payroll or HR. This is not a bad strategy - if it follows after an audit is conducted by an expert or by a set of experts that are sufficiently up to speed on how compensation equity laws are interpreted, as well as experts in the optimal methodologies in statistics and sometimes data science in order to ensure that the evaluation is optimized.

The person making the assessment should also be well-versed in running gender-equity audits. This can lead to identifying risks that don’t exist and/or not noticing areas of real risk. Whoever is making the assessment needs to draw circles around “substantially similar” employees and be capable of evaluating gaps using proper statistical methods.

It’s a good idea to bring in an experienced expert to evaluate an organization’s compensation system. The evaluations should not be based on pre-existing pay bands or pay grades. The experts, instead, can work with employers to identify optimal methods to measure compensation to render meaningful comparisons.

There are many reasons pay gaps can crop up in organizations – including maintaining a policy of offering starting pay at the bottom of a range, while showing willingness to go to the top of that range, particularly when pursuing external candidates. It’s been noted that this creates pay gaps early in employment, because men are more likely than women to ask for the most money possible.

When raises are percentage-based, the problem only gets worse over time. All of this means it’s imperative that organizations regularly analyze data and policies to ensure that they, in fact, are not part of the problem.

For reprint and licensing requests for this article, click here.
Compensation Analytics Data management
MORE FROM EMPLOYEE BENEFIT NEWS