District Court Affirms Withdrawal Liability Calculations, But Appears to Leave an Opening

Working right before the holiday, New Jersey District Court Judge Kevin McNulty issued a lengthy opinion late on July 3 in Manhattan Ford Lincoln, Inc. v UAW Local 259 Pension Fund. The case concerns the withdrawal liability assessed against Manhattan Ford, a withdrawing employer from a multiemployer pension plan and the actuarial assumptions used for the calculations. While this is far from the first case on this topic, it may be instructive to contributing employers who are considering withdrawing from multiemployer plans and for consultants and actuaries who work with these plans.

Before diving too deeply into the case, we need some background for readers who do not work in the multiemployer plan world on a daily basis. As the background is geared toward the more casual reader than to the multiemployer expert, we've intentionally omitted some details.

In 1980, Congress passed and President Carter signed into law the Multiemployer Pension Plans Amendments Act (MPPAA). In part, MPPAA established the concept of withdrawal liability as a means of ensuring that employers who choose to leave those plans pay their fair share of any unfunded liabilities.

For years, one of the frustrations of many who are involved with multiemployer plans has been the lack of guidance on actuarial assumptions particularly the discount rate to be used in determining the vested benefit liability (VBL) under a plan. That said, the statutory language that provides much of the guidance that we have in this arena and the language that the fund's actuary looks to in determining the discount rate to be used in annual actuarial valuations are somewhat instructive.

Quoting from a footnote in Judge McNulty's opinion, "The main upshot, for our purposes, is that under current law, 'each' actuarial assumption must be reasonable for the purpose of minimum funding, whereas they must be reasonable 'in the aggregate' for purposes of withdrawal liability." The language here is strikingly similar yet as we will discuss later, the calculations are often very far apart.

In this particular case, the plan's Enrolled Actuary (EA) used the Segal Blend (a method developed by the Segal Company in the early days of MPPAA) as a means of discounting in order to determine the fund's VBL and therefore the unfunded vested benefits (UVB) as well. To understand this case and the distinctions we will make later, we'll have to get hypertechnical (yet still oversimplified) for a moment to explain the Segal Blend. The Segal Blend essentially does two calculations and blends them. It considers that liabilities that can be settled by plan assets are assumed to have their risk transferred and therefore use PBGC rates (the rates inherent in insurance company annuity settlements) to discount the liabilities. But, for the portion of the liabilities not covered by plan assets, it acknowledges a risk premium and uses the funding interest rate (the EA's best estimate of future investment returns).

In Manhattan Ford, the fund's calculation performed by the EA showed that the withdrawing employer owed a roughly $2.5 million withdrawal liability based on the Segal Blend. The employer challenged the calculation and an arbitrator found in favor of the fund. Manhattan Ford appealed to the District Court.

Judge McNulty found that two essential questions were raised [quoting]:

1. As a matter of ERISA law, must a pension plan's actuary use identical actuarial assumptions to calculate the plan's satisfaction of minimum funding requirements and its unfunded vested benefits ("UVB") for withdrawal liability?
2. Assuming the answer to question 1 is "no," did the Arbitrator err in this case when he found that the discount rate applied by the Pension Fund's actuary to determine Manhattan Ford's withdrawal liability, the Segal Blend, did not render the actuarial assumptions "in the aggregate, unreasonable (taking into account the experience of the plan and reasonable expectations)"?

The Judge found that answer to both questions to be "no" and granted summary judgment to the Pension Fund. Summary judgment is granted only when the Court finds that the movant shows that there is no genuine dispute as to any material fact and the movant is entitled to judgment as a matter of law. Stated differently and from an extremely non-legal perspective, the Court found that even if all of the facts claimed by Manhattan Ford were true, it still had no valid case.

Thinking about this as a consulting actuary or even putting us in the lens of a withdrawing employer, this would appear to leave little room. And, this would appear that based on the fact pattern in this case that at least in the District of New Jersey that the Segal Blend produces results that are not unreasonable (note that the burden of proof here by statute is on the withdrawing employer).

What this opinion specifically does not say, however, is that the EA's judgment is infallible or indisputable. While the opinion did not address this, that the statute gives withdrawing employers the ability to challenge withdrawal liability calculations suggests that ERISA contemplates that there are, in fact, valid challenges.

What might they be?

We've seen a number of calculations of withdrawal liability where all of the discounting is done using PBGC rates. To understand the significance of this, note that in recent years, PBGC rates for this purpose have often been less than 3% while funding interest rates are often in the vicinity of 8%. Conservatively, changing the discount rate from 8% to 3% could increase liabilities by anywhere from 40% to more than 100% depending on the demographics of plan participants. For purposes of an example, let's use 60% and further, for extreme simplicity, let's assume that all liabilities of our hypothetical multiemployer plan are vested.

Suppose our plan has assets of $1 billion and liabilities discounted at 8% of $1 billion. Then, the unfunded liability at 8% is $0. However, decreasing our discount rate to 3% increases our liabilities to $1.6 billion and increases our unfunded liability from $0 to $600 million. That helps us to illustrate the extreme leverage inherent in many of these calculations.

So, to return to some of the questions at hand, let's suppose that the 8% discount rate applied in the annual valuation of the plan is the EA's actuarial assumption. In fact, when he has put it on the Schedule MB to Form 5500, he has asserted under penalty of perjury that it is reasonable. And, he has similarly asserted under penalty of perjury that each of his other assumptions is reasonable. Finally, he has asserted that, in combination, these assumptions offer his best estimate of expected future experience under the plan.

Now, the EA is asked to perform a withdrawal liability calculation. In performing that calculation, he keeps all assumptions except for the discount rate the same as they were for minimum funding. But, he changes the discount rate from 8% to 3%.

If each assumption is individually reasonable, one can extrapolate that the set of assumptions, in the aggregate, is reasonable. 

Given that the unfunded liability has increased from $0 to $600 million and even conceding that the purpose of the measurement is different, this begs the question as to how the revised assumptions (only one is revised), in the aggregate, can also be reasonable.

Withdrawing and potentially withdrawing employers should weigh these issues carefully when withdrawing. In combination with counsel and with other experts including actuaries, they should also weigh these issues when appealing a determination of withdrawal liability.