ANALYTIC ERROR IN THE EVALUATION OF "OTHER" FORMS OF COMPENSATION

Last year’s revisions to OFCCP’s scheduling letter included an addition of detailed compensation data (e.g., bonuses, incentive pay, commissions, merit increases, etc.) to be submitted as part of the itemized listing requirements.  Predictably, we are beginning to see requests for these types of compensation data during OFCCP audits (even those submitted before the change to the itemized listing). OFCCP’s strategy for analyzing these data remains unclear, but provisions in OFCCP’s proposed Equal Pay Report (EPR) may provide some insight.

The EPR proposal stated that Federal contractors would be obligated to submit yearly W-2 earnings for each employee. OFCCP could then compute statistical benchmarks of salary differences between protected class subgroups, although the appropriateness and meaning of such an analysis is dubious. The OFCCP Institute submitted a public comment through the rulemaking process to express reservations about the EPR plan. The OFCCP Institute illustrated, through simple data simulations (Appendix C), the gross analytic problems and interpretation difficulty posed by using W-2 earnings as the variable of interest.

If, during audits, the OFCCP plans to aggregate all compensation during the year for each individual and conduct protected class subgroup comparisons, they will run into many of the same problems outlined in the OFCCP Institute’s EPR public comment. To illustrate, we provide four simplistic simulations below:

Issue 1: Merit Increase and Part-time vs Full-Time status

Problem Statement: When part-time and full-time jobs are evaluated together, evaluation of earnings, rather than base pay, may show disparities simply due to full-time versus part-time hours worked.

Example: In this example, there are 3 men and 3 women in a job title. One man and two women work part-time at 20 hours a week, per employees’ hour requests. The hourly rate for this job title is the same for all employees. All employees received a bonus that was equivalent to their hourly rate multiplied by annual hours worked.

Conclusion: If bonus compensation is added to annualized base pay and average total compensation for men is compared to that for women, without regard for differences in full-time/part-time status, it appears that men make $250 a year more than women on average. If average bonus by sex is compared, without regard for differences in full-time/part-time status, it appears that the average bonus for women is 80% that of the average bonus for men ($1,250/$1,000).

Issue 2: Promotions/Demotions Differences

Problem Statement: When total salary earnings are evaluated, individuals who were promoted or demoted during the year will influence the analysis, despite similar pay at the data snapshot.

Example: In this example, there are 3 men and 3 women listed as cooks with an average salary of $30,000 per year, and the snapshot date is September 1^st, 2015. Of the six individuals who were cooks as of the snapshot, two (2 women) were promoted into the cook role from a $20,000 a year role on March 1^st, 2015. In addition, one sous chef (1 man; $70,000 a year) was demoted on March 1^st, 2015.

Conclusion: If salary earnings or hourly rates during the AAP time period are compared, women cooks would appear to be making approximately 74% of what the men cooks make ($27,490.67/$37,389.67 annually or $13.22/$17.98 hourly). Clearly, any apparent discrepancies are due to promotion/demotion activity, but there is a possibility for misconception if OFCCP were to analyze these types of data.

Issue 3: Employees with less than one year tenure

Problem Statement: Analyzing time period earnings for a group of employees, in which some employees were hired during the AAP time period, may result in the appearance of pay disparities where there are none.

Example: In this example, there are 3 men and 3 women distributed in one job with approximately the same exempt, full-time salary. Two women in this example were not hired until the end of February 2015. If the AAP snapshot date is September 1^st, 2015, the time period salary earnings for the two women reflect six months of base pay.

Conclusion: Comparing the actual earnings of women to that of men would indicate that women, on average, are paid approximately 89% of what men are paid ($80,012/$90,010).

Issue 4: Earnings other than base pay

Problem Statement: Analyzing all compensation for a group of employees in which some have received other forms of compensation (i.e., relocation funds, unreturned per diem payments, in-town evening meal reimbursements) may result in a false positive determination of potential discrimination.

Example: In this example, the 3 men and 3women have almost the same base pay values. Notice that all employees received some additional compensation in the form of unreturned per diem and/or in-town evening meal payments.  However, two women also received moving expense reimbursements.

Conclusion: Comparing the average total compensation of men and women would indicate that men earn $0.95 for every dollar earned by women, despite average base pay being nearly identical.

Conclusion

The examples provided are overly simplistic, compared to the scope and complexity of organizational compensation data and policies, yet the complications inherent in evaluating non-base pay compensation are clear even with such basic examples. In 2015, we expect OFCCP to continue requests of “other forms” of compensation, and we encourage contractors to take a considered approach to such requests, and seek professional assistance in formulating responses.

By Kristen Pryor, M.S., Associate Consultant and Kayo Sady, Ph.D, Senior Consultant, DCI Consulting Group