Planning Pay Equity Audits
The focus on pay equity is heating up. At the federal level, President Biden’s proclamation on National Equal Pay Day, March 24, provided renewed support for equal pay to advance “America’s values of fairness and equity . . .” His statement reaffirms the policies of the Obama Administration and its Equal Pay Pledge. The Pledge made voluntarily, committed employers “to conducting an annual company-wide gender pay analysis across occupations” and to equal pay efforts in broader enterprise-wide initiatives. By late 2016 over 100 companies had signed onto the pledge. Trump tried to cancel Obama’s policies but he was overruled by a federal judge. Biden restated his support for pay equity in his first speech to Congress.
Support for Pay Equity
Today the federal focus is on the possible passage of the Paycheck Fairness Act, a bill that was first introduced in 1997. It was re-introduced in January, passed by the House this April on a party-line vote, and sent to the Senate. The focus is sex discrimination and requires that employers prove that wage discrepancies are tied to legitimate business qualifications and not gender. It also prohibits employer retaliation for inquiring about, discussing, or disclosing employee pay.
It’s opposed by the US Chamber of Commerce that argues it will “significantly erode employer defenses for legitimate pay disparities . . .”
However, the 80 Business Roundtable member companies have signed on to a new multi-year targeted effort to reform companies’ talent management practices. As an element of that commitment, they have “called on companies to conduct periodic pay equity reviews and regular pay equity analyses, and to implement processes to review and close gaps. . . .”
An important development in January was a Seventh Circuit court decision (Kellogg v Ball State University) that found the Ledbetter Fair Pay Act’s paycheck accrual rule applies to all alleged discrimination. Prior to this, employees had 180 days to file a complaint. The Court ruled that each new pay period begins an employee claim for pay discrimination, making employers liable for decades-old discriminatory pay decisions. Until discrimination is addressed, the costs continue to increase. That is a new reason to consider a pay equity review.
At the state level, in addition to laws prohibiting discrimination in pay, 17 states have passed pay transparency laws that prohibit employers from restricting employees from discussing or disclosing their pay and from discriminating against those that do. In addition, 15 states and several cities now prohibit employers from asking applicants about their salary history. The two most aggressive requirements are California and Colorado. Their laws are discussed later in the report.
Significantly, Canada passed a Pay Equity Act in 2018 that “will come into force on a day or days that the Governor in Council will set”. It’s expected to become effective later this year. A pay equity plan, as required by the Act, must, at minimum, do the following:
- Identify the different job classes made up of positions in their workplace;
- Determine whether each job class is predominantly male, predominantly female, or gender-neutral;
- Determine the value of work [presumably based on a job evaluation method] of each predominantly female or male job class;
- Calculate the compensation of each predominantly female or male job class;
- Compare the compensation between predominantly female and male job classes doing work of equal or comparable value.
A key feature of the law is that merely developing a plan is not enough. Employers must increase pay for identified, underpaid female job classes within three to five years of the Act’s effective date. In addition, they have the obligation to close identified pay gaps disclosed during the updates required by the Act.
Pay equity is a global concern beyond North America. This past fall the United Nations held its first International Equal Pay Day. To mark the day, the Equal Pay International Coalition called all leaders to take necessary steps to ensure pay equity is at the heart of COVID-19 recovery efforts worldwide. The World Economic Forum has reported the Global Gender Gap Index since 2006 that combines data measuring the gap between men and women across health, education, politics and economics. On that index the US ranks 51st out of 149 countries, behind every country in western Europe.
Pay Equity and Comparable Worth
The pay equity issue first became a national issue following the passage of the Equal Pay Act in 1963. Then reformers arguing for ‘comparable worth’ focused on the ratio of women’s to men’s full-time, year-round, median annual earnings. In the 1960 Census, women aged 25 to 34 earned 65 percent of what men in the same age group earned. In 1980, the same women, now aged 45 to 54, were earning only 54 percent as much as men in the same age group. Census data for 1980 disclosed that female professionals were earning less than semiskilled male blue-collar workers, and female college graduates were earning less than male high school graduates who had not attended college.
The ratio was in part attributable to the pay of sex-segregated, “women’s jobs”. The answer was “comparable worth”, raising wages for traditionally female-dominated jobs to the level of those for arguably comparable male-dominated jobs. Further, the decline in the ratio was explained by women dropping out of the workforce to care for their children.
Comparable worth should not be confused with the focus of the Equal Pay Act, “equal pay for equal work”. Rather, comparable worth policies promote equal pay for comparable work. Advocates of comparable worth argue that sex discrimination in employment and pay setting reflects centuries of discrimination and the devaluation of the work traditionally performed by females. According to supporters, the pay of traditionally women’s jobs should be reset to better reflect the true value of the work.
Through the 1980s lawsuits brought the issue of comparable worth to political prominence. In the state of Washington, the AFSCME union won a case that provided raises and compensatory back pay to female state employees, who were found to be earning 20 percent less than their male coworkers. Although the decision was overturned on appeal, the state of Washington agreed to make women’s wages equal to those paid to men. Through the decade pay equity laws were adopted in Canada, Australia, and in a number of states and cities. Since then, women have made significant strides in moving into formally male-dominated occupations. Support for comparable worth waned after the 1990 recession, the decline in office support jobs traditionally held by women, and the ongoing changes in the work paradigm.
The pay ratio in 1990 was 71.6%. In 2019 it was 82.3%. With the COVID-19 crisis, the ratio would be different today because many of the jobs lost were in retail, travel, and service fields still dominated by women. But a key issue is that equal pay laws focus narrowly on occupational pay data, comparing “like with like”, where the gaps are significantly smaller.
A key for employers is the understanding of relevant labor markets, local area differentials, specific skill requirements, and other factors (e.g., volunteering to work overtime) that explain w-2 pay differentials.
Emerging Federal and State Reporting Requirements
The EEOC is responsible for enforcing the Equal Pay Act, Title VII of the Civil Rights Act, the Age Discrimination in Employment Act, and Title I of the Americans with Disabilities Act. Most employers with at least 15 employees are covered by EEOC laws (20 employees in age discrimination cases). The data that employers are required to report can be used under any of the laws.
More specifically, “The Equal Pay Act requires that men and women in the same workplace be given equal pay for equal work. The jobs need not be identical, but they must be substantially equal. Job content (not job titles) determines whether jobs are substantially equal. All forms of pay are covered by this law, including salary, overtime pay, bonuses, stock options, profit-sharing, and bonus plans, life insurance, vacation and holiday pay, cleaning or gasoline allowances, hotel accommodations, reimbursement for travel expenses, and benefits.” That’s clearly ‘total compensation’.
The 2019 & 2020 EEO-1 Component 1 data collection opened on April 26. The data collection deadline is July 19. For a time, the EEOC required submission of a Component 2 form reporting the number of male and female employees and the hours worked for 10 job categories, from executives to laborers and service workers. The data are reported for the usual race and ethnic groups. It’s very possible the form will be required again in the near future.
At the state level, the recently enacted laws and associated information required by California and Colorado could be the models for future compliance practices in other states.
Last September California adopted a reporting requirement essentially the same as the discontinued EEO – Component 2 form, showing the number and hours worked by male and female employees in the same 10 job categories. The data had to be reported by March. Over the last few years, the state also broadened the coverage of its equal pay legislation to include race and ethnicity. They also eliminated the requirement that jobs must be located at the same location; now employees need to be paid the same, regardless of location across the state. That ignores the significant differentials for larger, urban areas.
Colorado banned requests for a salary history in its 2019 Equal Pay for Equal Work Act (effective 1/1/2021). The law applies to all employers and aims to “help close the pay gap in Colorado and ensure that employees with similar job duties [regardless of job title] are paid the same wage rate regardless of sex, or sex plus another protected status.” The law recognizes “bona fide factors” that can justify pay differentials including:
1. A seniority system;
2. A merit system;
3. A system that measures earnings by quantity or quality of production;
4. The geographic location where the work is performed;
5. Education, training, or experience to the extent that they are reasonably related to the work in question; or
6. Travel, if the travel is a regular and necessary condition of the work performed.
Colorado now requires employers to notify employees of job openings and promotional opportunities, including the pay rate or range, a general description of additional cash payments (e.g., bonuses) and all benefits offered to hired applicants. That is a ‘red flag’ change that makes it easy for employees to assess their pay.
Significantly the state provided a safe harbor for employers that conduct proactive pay equity audits. Employers may avoid damages if they can demonstrate with reason that they believed they were not in violation of the law.
The Argument for Supporting Pay Equity
The most obvious reason is compliance with the law. All the evidence suggests the state laws will become more restrictive and noncompliance more costly. Only Mississippi has not enacted an equal pay act. The transparency laws were all enacted in the past three years. It’s very likely Congress will at some point pass the Paycheck Fairness Act. Plus, there will be increasing pressure from investors as well as public and media attention. There is no evidence support will fade.
To emphasize a key point, the focus of the research has been the overall US workforce and pay means and medians – but the laws are relevant to women and minorities at all levels. Analyses by McKinsey in 2019 found that in the prior five years “representation of women at the C-suite level” increased by 24%. At lower levels of management, women increased by over 25%. Their study found 87% of companies were committed to gender diversity, compared with 74% in 2015. President Biden demonstrated his commitment by appointing a number of women to leadership roles in government.
It should not be forgotten that the pay package for executives, managers and many professionals is not limited to salary. Comparative analyses need to take into account cash incentives, income from stock-based plans and eligibility for perquisites. Salaries are often a small percentage of compensation at senior levels.
An added factor is the labor market, the talent shortages in a number of fields, and the high demand for professionals in STEM and medical fields. Technology is expected to replace many low skill production jobs but fields where women predominate – education, business services and healthcare – will continue in large numbers. Prior to the pandemic, for the first time women outnumbered men in the workforce (excluding agriculture). In knowledge fields, the shifting occupational supply and demand will push up salaries for the better qualified applicants, regardless of gender. The pay gap with lower level jobs will increase over time, making adjustments more costly. That tilts recruiting to focus on women.
Perhaps most important is building a diverse, trusting and committed workforce. That’s important as organizations emerge from the pandemic and begin rebuilding their workforces. The next year will also see efforts to adjust to the “new normal” with redefined supervisor/employee relationships, possible reorganizations, and renewed interest in mergers and acquisitions. When the alternatives are assuming a proactive strategy to address equity problems or waiting until problems emerge, employees want to work for forward thinking employers with clear values that enable employees to work for a purpose, and balance work and life.
The work environment relevant to compensation management has changed significantly in the past decade or so. More important than the legislation is the availability of pay information on websites like Glassdoor. Those sources are far from valid, the data are fragmented and unreliable – but the information influences how employees view their pay and their employer. Employers need to consider their policies and their communications in light of what employees learn from those sources.
Now we understand the old ‘keep them in the dark’ philosophy – and the repercussions when employees were caught discussing pay — never benefited employers or their employees. In those years employees had no reason to believe wage and salary management was fair.
But pay transparency is a double-edge sword — it can be beneficial only when an employer is committed to maintaining ‘fair pay’. An important question is – How would employees react if salaries were posted on a webpage available to everyone? How would women and people of color react? Leading edge practice is moving in that direction.
We now have a better understanding of how the employee experience influences their commitment – how pay that is seen as unfair affects their commitment. Their understanding of the pay program and how its managed both internally and relative to market practices is central to its effectiveness. Pay equity studies confirm an employer’s commitment to a fair workplace. This is not an issue that employers can afford to ignore.
The bottom line is that the pattern as confirmed in research across the private sector is broadly consistent – employers still pay women and people of color less than white men for the same work. That’s true at all career stages which translates into differences in lifetime earnings – as much or more than $1 million — and influences family planning decisions that will affect future generations. Organizational leaders need to decide if that is consistent with their management philosophy and values.
The bottom line is that employers still pay women and people of color less than white men for the same work.
Equal Pay Language – the Starting Point
Federal and state equity laws focus internally on the management of pay for ‘substantially equal’ jobs. That is a different focus than the advocates of comparable worth. The legal requirement is more consistent with the reality of labor markets and the dynamics that drive up salaries in high-demand fields.
In the 1980s, when pay equity first became a national issue, the early methodologies were a response to the comparable worth argument and followed the logic of job evaluation, covering all exempt jobs in an organization. These systems were the first ‘comp’ applications using multiple regression to analyze data. A core problem then and now is that few people understand the math; it goes well beyond the basic stat course common in college curriculums. It also goes beyond the requirement in federal and state statutes.
Federal law has been consistent since the passage of the Equal Pay Act in 1963, which requires: “men and women be given equal pay for equal work in the same establishment. The jobs need not be identical, but they must be substantially equal.”
In planning a pay equity analysis, the starting point is the language on the EEOC website:
Skill: “Measured by factors such as the experience, ability, education, and training required to perform the job. The issue is what skills are required for the job, not what skills the individual employees may have.”
Effort: “The amount of physical or mental exertion needed to perform the job.” The focus is on extra effort. An example from the EEOC website focuses on a job requiring heavy lifting.
Responsibility: “The degree of accountability required in performing the job.”
Working Conditions: “This encompasses two factors: (1) physical surroundings like temperature, fumes, and ventilation; and (2) hazards.”
Establishment: “An establishment is a distinct physical place of business rather than an entire business or enterprise consisting of several places of business.”
“Pay differentials are permitted when they are based on seniority, merit, quantity or quality of production, or a factor other than sex. These are known as “affirmative defenses” and it is the employer’s burden to prove that they apply.” (The key words are in italics.)
That language can be used to explain and justify pay differentials. A clear example is the current focus on cybersecurity specialists and the high demand driving up their salaries. The pay equity argument supports comparing the pay of men and women in this new specialty but combining those specialists with others in the IT field makes no sense since the skill sets, especially the non-technical skills, are not the same.
An important ‘however’ is that under Title VII, ADEA, and ADA, discrimination claims are not limited to jobs that are substantially equal or that employees work in the same establishment. The EEOC website uses the example of a disabled employee paid less and the employer’s explanation does not satisfactorily explain the differential.
Data Concerns Need to Be Addressed
Regardless of the jobs in question, it will be important to confirm relevant, current data are available. Its often the case that multiple data platforms are involved, or the data are incomplete or not current.
For each employee included in the analysis, the data file should include the basics – salary, grade/band, age, gender, race/ethnic group, date of hire, date of last promotion, last performance rating, etc. For jobs where overtime, shift differentials or bonus payments are involved, w-2 earnings is probably the best measure (since differences in male /female schedules and work hours are relevant). Differences in the scope of financial or managerial responsibility should also be considered.
A common problem is that job titles and job descriptions are not current. Or that the same title is used with jobs requiring different skill sets. Similarly, analyses that compare jobs at the same salary grade (or band) should start by confirming jobs are graded accurately. It may be that a survey to confirm job duties and skills is the right first step. A ‘like with like’ analysis has to start with reliable employee and job information.
Simple descriptive statistics – means, medians and percentiles – are understood by virtually everyone. But most pay equity studies are based on regression analysis – which at its simplest can be illustrated on a graph where the X or bottom axis is the independent variable and pay is the Y or dependent axis.
Actual pay decisions are based on several “independent” variables (simplistically captured in the law by skill, effort, responsibility and working conditions). That means pay equity analyses require multiple regression methods – and multiple ‘Y’ variables. Its not a subject a lot of people have studied or used.
One reality of regression analysis is that the math always produces an equation, regardless of the data. That would be true even if a janitor’s pay was compared with the CEOs pay. A related issue is that the equation can be highly influenced by extreme or outlier values – like the janitor/CEO example. That makes it important to evaluate the regression results. It can be GIGO – garbage in, garbage out – if the assumptions in planning the analysis do not make sense.
An added issue is that simple regression assumes a straight line relationship (to illustrate, that would mean each year of experience adds a constant dollar amount to an employee’s pay). That’s overly simplistic. When salaries are graphed, the data pattern is typically not linear. Pay increases are generally expressed as a percentage, which means the dollar amounts get larger at high salaries. Further, in organizations that relay on a merit matrix based policy, the percentage increases are larger early in a new job but then get smaller as tenure and salaries increase. To accommodate the percentage issue in the analysis, salaries are often switched to logarithmic values. (The step increases in the General Schedule reflect this pattern.) The resulting estimates then have to be switched back to dollars for discussion.
In a multiple regression equation, the coefficient linked to each X variable shows the change in Y for an incremental change in the Y variable. Gender (or race) is normally assigned a 1/0 value and the coefficient is a measure of the overall difference between men and women.
Another technical problem is multicollinearity which affects the validity or accuracy of the coefficients. That occurs when there is a high intercorrelation between two or more X variables. When the problem exists, it distorts the coefficients as well as the pay estimates. Its common with the factors that govern pay since older, longer service, better educated employees are generally higher paid. A series of preliminary analyses is useful to assess the relevance of the issue.
A practical issue is that all regression results are best understood and pictured as that X-Y graph with data points plotted above and below the line. In concept that’s true with a multivariate analysis. Pay adjustments close the gaps, effectively raising below-the-line salaries – relative to the highest salaries.
The concerns make it important to invest adequate upfront time to plan the analyses. To reiterate, regression analyses always produce an equation. That’s true even if its a CEO/janitor comparison. Once the data crunching ends, the results have to be credible to everyone. The credibility is likely to ride on the planning and ongoing communication.
Caveat: Pay analyses can uncover evidence of discrimination and place employers at significant legal risk. Employers should consider if it’s important to preserve attorney-client privilege.
Planning the Job to Job Comparisons
An employer’s commitment to pay equity is one of several issues that need to be considered in developing an overall compensation strategy. The core questions are: Is total compensation planned and managed to support the talent management strategy? Is the organization able to hire and retain talent with essential knowledge and skills? If not, is compensation an issue? Does the management of financial rewards engage employees and support planned performance levels? And of course – Is there evidence of bias or discrimination in the management of compensation?
If the answer to the final question is affirmative or uncertain, it warrants the investment to develop and assess the evidence. Over and above what is essentially a ‘CYA’ reason, conducting a pay equity analysis sends an important message. When the analysis is initiated to avoid litigation, narrowly defined job-to-job comparisons are warranted. But if leaders are making a commitment to employees, the message is important and the goal argues for a somewhat broader grouping of jobs.
The necessary analyses can be done behind closed doors but that is taking a significant risk. The initiative will have more value if employees are updated as the project unfolds. Even better is involving them in the planning and analyses and in developing recommendations. That is a common strategy when new pay programs are developed in higher education and healthcare. Experience with employee teams shows they take their responsibility and their planning decisions very seriously.
To make certain any adjustments address the right issues, an early step employers should consider is soliciting feedback from managers and employees. Bias as well as discrimination is generally not flagrant or simple so focus groups are likely to be the most informative sources. But the old saying is relevant — Don’t ask if you don’t want to know the answer.
Every employer should plan to consider what pay equity means within their talent management strategy, work environment, and recent or planned organizational changes. If a reorganization or merger is in future plans, an equity audit should be included in the planning.
The strategy for developing an understanding and the steps to address problems need to be consistent with the values of leaders and their approach to addressing problems. In more cautious environments that could translate into limiting the analysis to specific jobs with the same or similar titles. It could also mean limiting the analysis to separate work locations (except in California).
The planning focuses on three core issues: the jobs and career ladders to be compared, the factors are seen as relevant to pay management – grades, skills, education, experience, performance, etc – and the database. The common data element of course is gender (or race/ethnic group).
Two additional issues need to be confirmed before starting the analysis – top management’s commitment and the plan to address needed adjustments. A communication strategy should also be confirmed.
Setting the Stage for Talent Management
Three decades ago, the workforce was managed as a cost. HR was an administrative function. Then Gallup concluded that engaged workers are far more productive – as much as 30 to 40% more. About the same time the Great Places to Work Institute was created and initiated the lists that are now found in numerous websites and publications here and in other countries. The pandemic gave every employer a reason to value committed workers. Now people management is a key to success.
A pay equity analysis will help to understand the past. It provides insight into what factors have affected pay and careers. Differences will become apparent across large organizations. That’s true at each level of the organization. The analysis could provide insight relevant to each employee. Regression equations provide estimates of gender neutral, “fair” pay for each employee. Employers need to use the evidence to address problems.
Going forward, dashboards can be used to highlight departmental or job family issues and to keep managers and possibly employees informed. The dashboards set the stage for future equity audits.
With broader planning, the data collection can generate the information needed to assess additional talent management concerns –
- To confirm salary grade/band levels;
- To confirm or define career ladder skill profiles;
- To set the stage for market analyses;
- For managers and supervisors, to compare their budget and talent management responsibilities.
Compensation is typically not the highest priority in attracting and retaining talent but it’s always a consideration. The management of pay as careers unfold sends important messages. With the new emphasis on transparency, employees expect to be kept informed. They want to know where they stand. They will compare notes with co-workers. A commitment from leaders will be important.