Monitoring for Compliance in Terminations
OFCCP examines terminations when there are racial, gender or other disparities in the rate of terminations in a job group or title that are statistically significant. In most termination cases, these disparities can be readily resolved with proper analysis.
The first step is to separate voluntary from involuntary terminations. With voluntary terminations, the employee decides on his or her own to leave a job. Some common reasons for voluntary terminations are: getting a more desirable job, going back to school full time, retirement, and relocating out of the area. When examining voluntary terminations, OFCCP asks the employer why the individual left and will attempt to contact the former employee to verify the reason given. If the individual confirms the reason for leaving, OFCCP will usually consider the matter resolved. If the individual cannot be contacted, OFCCP will usually still consider the matter resolved because there is no evidence that the reasons offered by the employer are untrue.
If a disproportionate number of a particular protected class of employees is voluntarily terminating, it is important to look into the reasons even if they would not support an allegation of discrimination from OFCCP. This is especially important if you are trying to correct underutilization for that particular group. Exit interviews may be useful in helping you to understand why employees are leaving. However, you may not always get candid responses especially if the employee's reasons for leaving are related in whole or in part to discriminatory bias. An employee may also be reluctant to provide any negative feedback even if he or she is just feeling disconnected. Without that feedback, it becomes difficult for the employer to remedy the situation and improve retention. One idea you may want to consider is providing a way for employees to report anonymously on why they are leaving and what, if anything, would have increased the likelihood that they would have stayed with the company. Perhaps this could be done through an electronic portal that would mask the identity of the employee. While you may not be able to rely on everything that is input into such a system, it may give you a way to detect earlier rather than later some concerns that could eventually become serious problems. It may also make your good faith efforts have more lasting effect by helping you create an environment where the beneficiaries of your affirmative action efforts want to stay.
Although OFCCP does not find a large number of discrimination violations associated with terminations, those that are found tend to occur in involuntary termination situations. Involuntary terminations can occur for a number of reasons including unacceptable conduct by the employee such as job abandonment, restructuring leading to reductions in force and layoffs.
Conduct-based Involuntary Terminations
To examine conduct related terminations for discrimination, you need to do what I call the "termination two-step." First determine whether the terminated employee actually engaged in the prohibited conduct. For example, you may have leave policies that require an employee to be terminated if they have three unexcused absences. Your policy defines unexcused absence as "no call", "no show." Under the first step, you would check to see if the employee actually had the three unexcused absences. The second step is to examine how the policy was applied. This requires determining whether anyone else violated the policy and whether their comparable misconduct led to their termination. Continuing with our example, Joe has four unexcused absences and Mary has four unexcused absences. Mary is fired but Joe is suspended. The first part of the analysis, determining whether or not Mary actually violated the attendance policy, would be satisfied because she did violate the policy. However, if you stop your analysis at that point, you would miss the fact that Joe also violated the policy and yet retained his job. The second step of the two-step, examining the disciplinary history of employees who have committed the same infraction, picks up this disparity in the treatment of Joe and Mary. Unless the employer can provide evidence of a legitimate non-discriminatory reason for the difference in treatment between Joe and Mary, it will be found to have discriminated against Mary on the basis of gender.
Conduct policies must also be examined for potential disparate impact in the policies themselves. For example, an inflexible leave policy, allowing a set number of leave days with no exceptions, will likely have a disparate impact on employees with disabilities who require leave as a reasonable accommodation. Since employees with disabilities are entitled to a case by case determination of the appropriateness of an accommodation, and accommodation can include a request for leave beyond that permitted under the policy, imposing a rigid policy that does not allow for accommodations would have a disparate impact on individuals with disabilities. Disparate impact in the area of disability discrimination does not require the kind of statistical foundation that is required in race, gender or national origin cases. A single person screened out because of such a policy can allege disparate impact on the basis of disability. It is an alternative approach to cases that might also work under a reasonable accommodation theory. These kinds of violations can be found not only by running analyses of the actual terminations that occurred during the review period but also by a close examination of your leave policies themselves. If such policy is detected during a review, the agency will likely look to see if anyone was terminated because of the policy during a period in which it has jurisdiction to find liability.
Although in most cases the agency's attention is initially drawn to terminations because of statistically significant standard deviations in the job group or job title, quite often the circumstances surrounding the specific terminations tend to eliminate the pattern and practice and turn the focus of the investigation to specific cohorts like Mary and Joe in the prior example. With the right set of facts, it would be possible to establish a pattern and practice of imposing harsher discipline on women in general than is imposed on men. This kind of proof would require a large enough number of terminations for the same infraction to produce reliable statistical patterns. Since this volume of misconduct among the similarly situated is less common, pattern and practice termination violations are more likely to surface in involuntary terminations that are not conduct based.
Non-conduct based Involuntary Terminations
Layoffs and Reductions in Force (RIFs)
Mass reductions in force or mass lay-offs are more susceptible to a pattern and practice analysis. In these cases, the employer has often identified a group of people it views as similar with respect to their vulnerability for termination. For example, an employer may decide it needs to lay off 50 people who work in the same division in the same or a similar job. In these situations, the analysis of potential discrimination begins by determining what criteria the employer has established for determining who stays and who goes.
For example, an employer may establish a point system with 10 points for the dollar value of business the employee has brought in over the last year, 7 points for performance, and 5 points for tenure. The terminations in the reduction in force or layoff would come to the attention of the agency because they appear at least initially to be statistically significant. The agency would identify the pool of people who potentially could have been terminated. In this example, it would be the employees in the division who worked in the jobs targeted for the reduction in force. The first step in the analysis would be to verify that the employees who would be expected to be terminated under the terminations policy were actually the ones terminated. In this example, suppose those with higher point totals were supposed to have a greater chance of staying than those with lower point totals. Lee, who is Asian, has 10 points for dollar value of business, 7 for performance and 2 for tenure. Brad, who is white, has 5 points for business, 6 points for performance and 5 points for tenure. Under the policy, Brad would be expected to be terminated before Lee. Lee has 19 points and Brad only has 16 points. If instead, Lee were terminated, OFCCP would be concerned that the reason for his lay off was his national origin, Asian. If OFCCP found that other Asian employees were being terminated who had higher scores than other favored group members, it may find a pattern and practice of discrimination in how the reduction in force or lay-off was carried out. Even if no pattern was found, the agency would look to remedy the apparent discrimination against Lee.
Sometimes involuntary terminations are the result of poor performance. The employee is simply not delivering the performance up to your standards. The examination of performance based terminations is similar to the evaluation of conduct based terminations. First, determine what the performance standards are and what process, if any, the employee is due under your performance system. Then determine if the policy was followed and whether or not the employee met the standard.
Performance objectives established in advance will usually provide the best evidence of what the performance standards actually were. It is helpful if these standards are sufficiently descriptive so that a third party, here an outside agency, can fairly readily determine whether or not the standard was met and whether it was applied in a non-discriminatory fashion. Factors that may be addressed by the standard are timeliness, quality, output and whatever other factors are relevant to determining the level of performance.
Second, determine if other employees with similar performance received similar discipline and if not, why not. If you have a performance improvement system or some other intervention measure designed to provide the underperforming employee an opportunity to improve, the agency will want to see that all employees entitled to this procedure are, in fact, being treated in a non-discriminatory manner. For example, if a Black employee is fired for poor performance and a White employee with comparable performance problems is transferred to a job the employee might be better suited for, this raises an inference of race based disparate treatment. Unless the employer can provide a persuasive, legitimate, non-discriminatory reason for this difference in treatment, it will be found to have discriminated against the Black employee.
It is important to review the fairness of the performance system itself and also to review whether rating officials are, in fact, implementing the system as designed. A failure of either of these can lead to differences in treatment that are hard to explain. The absence of a reasonable, non-discriminatory explanation is enough to allow the inference of discrimination to stand.
If you are considering terminating an employee with a disability for poor performance and the employee suggests that the poor performance is related to the disability, you will need to be able to show that you have provided any required reasonable accommodations, that you have engaged in the interactive process with the employee if the necessary accommodation is not readily apparent to either of you and that you have examined whether there are any comparable or lesser positions that the employee could be reassigned to as a reasonable accommodation. Only after these avenues have been exhausted and the individual is still unable to perform the essential functions of the position held or desired would termination for poor performance be acceptable.
Employers should be especially conscious of termination actions that are taken against employees who have exercised their rights under the civil rights laws and statutes. For example, Tom has been a marginal employee for a number of years. You really should have gotten rid of him long ago but you did not. You find out that during the compliance review, Tom was letting the company have it with both barrels in his conversations with the OFCCP compliance officer. He also started filing complaints against the company with the EEOC. Suddenly, your tolerance for Tom reaches its limit. You have lots of documentation about how badly he has performed over time so you fire Tom.
The question the agency will have, or should have, is why now? If Tom's performance has been tolerated for an extended period of time, why is it viewed as requiring action only after Tom has begun to exercise his rights under the civil rights laws? The nearness in time between the discipline and engaging in protected equal employment opportunity activity is likely to raise an inference of retaliation. Even if the employer is found not liable on any of the other allegations made by Tom, it could still be found to have retaliated in taking discipline that would not have been taken but for the protected activity. It is extremely easy to persuade yourself that you have just grounds for this action, especially since you probably had ample grounds long before he engaged in protected activity. However, because of the timing and your apparent tolerance of this performance before, firing Tom at this point is likely to prove problematic. This is one reason why it is better to take prompt action when the performance problem arises rather than allowing a problem to go unaddressed until the employee starts exercising his civil rights.
OFCCP discrimination findings in terminations are fairly rare, but the obligation to monitor compliance in this area is ongoing. Hopefully, this information will assist you in making sure that when terminations happen, they are handled in a fair and non-discriminatory manner that reflects well on the company and that ensures fairness for the individuals affected.