Splitting Up DB Plans Can Reduce PBGC Premiums

Splitting up a defined benefit (DB) plan into a plan that covers participants subject to the variable-rate premium headcount cap and a plan that covers all other participants may reduce the Pension Benefit Guaranty Corporation (PBGC) variable-rate premiums in two ways, according to an article penned by Brian Donohue, partner at October Three Consulting, based in Chicago.

PBGC premiums consist of two parts. The flat-rate premium for single-employer plans in 2018 is $74 times the number of plan participants. Donohue explains that all plans pay this, as it is a function of headcount.

The variable-rate premium (VRP) for single-employer plans in 2018 is $38 per $1,000 of unfunded vested benefits (UVBs). The VRP is capped at $523 times the number of participants—what Donohue calls the head-count cap (HCC).

Donohue explains that the VRP and the HCC trade off against each other. For example, in a one-participant plan, if the funding shortfall for that participant is greater than $13,763 ($523 divided by 3.8%), the HCC will apply. If the shortfall is $13,763 or less, it won’t apply. The VRP for DB plans is the lesser of the VRP and the HCC.

“Assets are allocated to participants based on ordering rules applicable to plan terminations. (Note, simplified rules apply in the case of de minimis spin-offs of <3% of plan liabilities.) These calculations require a separate actuarial valuation using assumptions dictated by PBGC, and the ordering rules generally favor current retirees. Once assets are assigned to participants, the calculation of each participant’s unfunded liability is straightforward,” Donohue’s article explains.

The idea is to put participants with lesser unfunded vested benefits (UVBs) in one plan, and those with greater UVBs in another.

Legality and nondiscrimination concerns

Plan sponsors are allowed to have more than one DB plan.

One might think that since asset allocation generally favors retirees, one plan may contain more older participants and the other may contain more younger participants. So, would this cause nondiscrimination testing problems?

Donohue says the thinks there are no nondiscrimination testing issues with splitting plans. If the plan is frozen and no one is accruing benefits, testing will not be a problem. If the plan is not frozen and is providing future benefit accruals, permissive aggregation allows the two plans created by the split to be combined for testing purposes, eliminating nondiscrimination issues.

How splitting the plan helps

Splitting the plans can be used to maximize the effect of the headcount cap, and it may allow the DB plan sponsor to make contributions that reduce variable-rate premiums where the headcount cap would otherwise prevent that result.

Donohue considers a two-person plan. One person has a UVB of $5,000 and the other has a UVB of $45,000. If the first person is put into a separate plan, the VRP is going to be the lesser of 3.8% of $5,000, which is $190, or $523 times one participant, or $523—so the VRP is $190. For the participant in the second plan, the VRP is the lesser of 3.8% of $45,000, which is $1,710, or one participant times $523, or $523—so the VRP is $523. So the total VRP the plan sponsor will pay for both plans is $713.

If the plan sponsor kept both participants in one plan, the average of the underfunding of the two participants’ benefits is used for each participant, which would be $50,000 divided by two or $25,000, so the UVB for the plan is $50,000. The VRP is the lesser of 3.8% of $50,000, which is $1,900, or $523 times two participants, which is $1,046—so the VRP is $1,046.

This simple example shows how DB plan sponsors can save on PBGC VRPs by splitting plans.

But, a second advantage, according to Donohue, is that if the plan sponsor makes a voluntary contribution to the first plan of $5,000, it would reduce the UVB to $0 and wipe out the premium for that plan altogether. “What we see in practice is plan sponsors that are limited to the HCC don’t have same incentive to contribute that other plan sponsors do,” he says.

There are some complications to implementing this strategy that Donohue mentions. For one, the VRP is set to go up to 4.2% or 4.3% next year, and the “threshold” for identifying “HCC participants” will decrease to about $12,500 in 2019 and stabilize around that number thereafter. He says actuaries may want to model the scenario out a few years.

“In general, the greater the number of participants moved into the ‘non-HCC plan’, the more significant the impact on PBGC premiums. But also, the greater the appetite the sponsor has for making voluntary contributions, the greater the number of participants that can be moved to the ‘non-HCC plan’. So, in practice, the process of determining the appropriate split can be iterative. And, the asset allocation calculation is costly.

According to Donohue, plans with about $50 million in assets would save on PBGC premiums an amount that would be greater than the cost to implement the strategy, but smaller plans may have the benefit eaten up by the costs.

“Splitting plans is within the law and a good policy outcome, but the bottom line is, if we can get DB plan sponsors to a situation where they can fund the plan better than they do today, that will be a better outcome for everyone,” Donohue concludes.