I wrote recently about some of the causes of the relative demise of defined benefit (DB) plans. Last Thursday, although it was dated Friday, the IRS released a new proposed nondiscrimination regulation. We had been told that this was coming down the pike. It was going to be all good. It was going to provide relief so that a DB plan that was compliant and was frozen to new entrants (sometimes called soft frozen) would not suddenly become non-compliant due to natural turnover and promotions of active employees in the plan. That the IRS addressed this nearly 23 years after the regulations were finalized was long overdue, but good news.
We didn't know that there were going to be other changes. One of the hallmarks of the suite of nondiscrimination regulations (nondiscrimination, coverage, minimum participation) is that they provided objective tests to determine whether a plan discriminated unfairly in favor of highly compensated employees (HCEs). In fact, the final regulations tell us that satisfying the various tests included in them are the sole method of demonstrating nondiscrimination. While this caused a significant burden for some companies and their plans, it gave plan sponsors comfort in knowing that passing those tests meant that there plans were, in fact, nondiscriminatory. Many design and redesign studies were done because of this knowledge.
Then came last week,
I now digress into some technical mumbo jumbo the likes of which this blog has not seen for a while. If you don't like or care about the technical stuff, please stick with me; I'll return to the more common language soon.
The rules for nondiscrimination in amount under 1.401(a)(4)-2 and -3 contain something known as the general test. Under the general test, there exists a concept known as the rate group. For each HCE, there is a rate group consisting of that HCE and all other nonexcludable employees whose normal and most valuable accrual rates are at least as big as those of that particular HCE (people with similar accrual rates can be grouped to be considered to have the same accrual rate as each other). Once we establish the rate groups, we must demonstrate that each rate group satisfy the head counting portion of the same coverage test under Code Section 410(b) that the plan uses to satisfy coverage. So, if the plan uses the Ratio Percentage Test to satisfy 410(b), then each rate group must have a ratio percentage of at least 70%. On the other hand, if the plan uses the more complex Average Benefit Test to satisfy 410(b), then each rate group must have a ratio percentage at least equal to the lesser of the actual ratio percentage for the testing population or the midpoint of the safe and unsafe harbors (this threshold is far less than 70% and is frequently in the neighborhood of 30%).
This has been the case since 1993. In fact, there are now people performing testing who were not alive when we started doing testing this way.
Under the proposed regulations, however, that lower threshold would not be available to plans that have separate formulas for which there is not a reasonable business criterion. The proposed regulation tells us that naming names (common in designs often known as QSERPs) does not constitute a reasonable criterion. The problem is that when the word reasonable is used, we, the practitioners never know what someone else thinks is reasonable.
So, will we be left doing tests that we think are correct only to learn that someone else with a higher authority thinks that we have misconstrued what is reasonable? Is this a calling to just stop providing retirement benefits? After all, we could design perfectly good retirement plans for our clients today that we learn tomorrow have suddenly failed to still be nondiscriminatory.
From a personal standpoint, the good news is that even most of the aggressive designs that my colleagues work with would seem to satisfy the reasonable business criterion test. There are lots of plans out there, however, that do not, and these plans may be in trouble if and when the proposed regulations are finalized.
Would you like to know if your plans seem safe or potentially in trouble? I'll be happy to help you figure that out.