The Tests You Don’t Want to Fail

February 12, 2014 ( – Good news: your employees are contributing to the 401(k) plan. Bad news: your highly compensated employees are contributing more than others to the 401(k) plan.

Every employer-sponsored 401(k) plan must undergo non-discrimination tests in order to make sure the plan is not too highly weighted in favor of upper-level employees. If the highly compensated employees contribute a higher percentage to the plan or receive a higher percentage of employer contributions than non-highly compensated employees, the Internal Revenue Service (IRS) will change the rules of the plan, potentially risking its tax-exempt status. Plans can be brought into compliance by returning excess contributions to the company’s highly compensated employees, and these contributions must then be reported as taxable income.

The actual deferral percentage (ADP) test compares elective deferrals made by highly compensated employees (on a before-tax basis and excluding the extra catch-up contributions allowed to those age 50 or older) to those made by non-highly compensated employees. The actual contribution percentage (ACP) test compares employer matching and after-tax contributions made on behalf of non-highly compensated employees to those made on behalf of highly compensated employees.

An employer has several options to improve test results, says Derrin Watson, an attorney with the Relius Education division of SunGard Financial. A low participation rate by non-highly compensated employees can put the plan at risk for failing the tests. Try talking to employees to find out why they are not deferring, he suggests. “I know one company that had 65% of participants deferring. They said, ‘We would not accept a 65% success rate in any other area of our business,’ “ Watson tells PLANSPONSOR.

The discussion led to two changes in the design of the plan: auto enrollment and simplifying the investment options, instituting a target-date fund. Auto enrollment brought them to 100% participation, Watson says. (The plan has 200 employees.)

Auto enrollment is being embraced by more small-plan sponsors, Watson says, and historically has increased the number of participants deferring by 20% to 26%. Streamlining the investment options is just common sense, according to Watson. “People pay a price, an internal cost to make a decision,” he says. “and it’s easier to let other people decide.”

When participants are unable to decide how much of their paycheck to defer, Watson points out, or if they should even defer at all, how can they choose from 25 investment decisions? Auto-enrolling participants into TDFs and balanced funds gives the plan sponsor fiduciary protection, according to Watson, and makes it easier for the participant to get into the plan in the first place.

Plan sponsors can consider a matching contribution. “Historically, in an ADP-tested plan, the reason is to make participants defer,” Watson says. Plan sponsors who already have a match in a plan that is not doing well might consider stretching it a little. “If your match is low, say dollar for dollar up to 3%, consider increasing it, to 50 cents on the dollar up to 6%,” he says. “Stretch it out to get more. People are very savvy these days about maximizing their match.”

Watson suggests considering the wide array of design and testing options that can be used to help plan sponsors pass the nondiscrimination test. If a plan uses prior year testing, the highly compensated employee contributions are compared to the non-highly compensated employee contributions for the prior year. Current year testing compares the highly compensated employee contributions to the non-highly compensated employee contributions for the current year. The primary advantage of using prior year testing is the average for the highly compensated employees will be known early in the plan year so a plan sponsor can predict the results and, if they choose, limit their contributions for the year (see "Improving Nondiscrimination Test Results").

“Switch to prior-year testing, and you can tell owners, contribute this much without a problem. You have a target you can aim for,” Watson says.

Another method is to consider using a different definition of compensation if the plan document permits it. “You don’t have to use the same definition to allocate your contributions as you do to test for nondiscrimination,” Watson says. “If the employer makes a profit-sharing contribution, the plan can allocate based on total compensation. But for nondiscrimination testing, if the document allows it, you could test based on compensation net of elective deferrals. It doesn’t hurt the boss’s testing at all, but it helps the staff because it reduces the denominator that you’re testing by. It can be a good trick, and you may not even have to amend the plan.”

Another possibility is to use the “otherwise-excludable employee rule.” According to Watson, new hires are some of the least likely to defer. This rule treats them as if they are in their own plan. Since most highly compensated employees are generally not new hires, separating the two groups keeps the plan in compliance.

Changing the eligibility requirements to participate would require a plan amendment, Watson notes, but could exclude groups of employees, specific workers who have proven to be inactive and unlikely to defer. Another plan amendment would reclassify well-compensated non-owners as non-highly compensated employees. The situation is occasionally found in a medical practice or law firm, Watson says, where a manager who is not an owner makes more than $100,000.

Such employees frequently defer well and can significantly raise the average deferral rate for the highly compensated employees. Using the top 20% rule and taking them out of that group will lower the average deferral rate for highly compensated employees.

Perhaps the simplest thing is to get out of nondiscrimination testing altogether by adopting a safe harbor 401(k) plan, Watson says. But switching to a safe harbor plan carries a price tag, he points out: the plan sponsor will have to make a non-elective contribution of at least 3% of compensation, whether or not employees defer.

“Since many 401(k) plans with testing problems are smaller plans that are top-heavy, the employer is already making a 3% contribution for non-key employees anyway,” Watson explains. “The only extra cost for the safe harbor is that you have to fully vest the money. But there’s nothing else out of pocket.”

For 401(k) plans that are not safe harbor plans, the sponsor can give the 3% contribution and have a six-year vesting schedule. Under safe harbor 401(k) plans, vesting is immediate.

Watson feels this is the reason two-thirds of plan sponsors with fewer than six participants have a safe harbor 401(k) plan. But, giving employees the money allows the plan sponsor to avoid testing altogether, he says.