Our company employees union employees from numerous trade unions. Each trade offer its own retirement benefits to its members; many of which consist of 401(k) plans administered by the various union trusts. As an incentive to work for our company, we have offered an additional 401(k) plan to these union employees with a 3% match. Historically, for nondiscrimination testing, we have only counted employees who participated in our plan. We did not count eligible employees who did not participate in our plan because they had 401(k) plans offered by their individual unions that they were participating in. In other words, rather than taking deferrals from pay and adding to their company 401(k) accounts, we were making contributions to their union plans out of our payroll.
This year, we were informed by our TPA that we should have been counting all "eligible" union employees, not just the employees who participated in the plan. When these additional "eligible" employees were added to the pool of employees tested, we failed the ADP test due primarily to all of the non-participants.
We fixed the problem this year and issued refunds to the impacted highly compensated employees. However, because we performed the nondiscrimination testing using incomplete data for many years, we are facing the rather daunting task of remedying past failures. We have been advised that the remedy involves making refunds to highly compensated employees for past failures and making one-to-one contributions to the non highly compensated pool of plan participants. Besides the financial burden this poses, it leads to several other problems such as deceased participants and participants who retired and pulled their money from the plan.
Has anyone had a similar experience and what recommendations do you have for dealing with the issue? Is anyone aware of a rule or exception that would justify the way we were counting employees for discrimination testing?
Your comments and suggestions will be much appreciated.
