MEMORANDUM
This is an ERISA case. There are three issues.
The parties agree that the first two are reviewed under the "arbitrary and capricious" standard. (D.I. 37, p.12; D.I. 39, p.15).
The first issue involves calculating how much service Plaintiff has. As I understand it, the Plan does not state how to calculate service for a "partial year." The Plan Administrator uses, and has apparently consistently used for a long period of time, a simplifying assumption, that is, that a year has 360 days, and that each month has 30 days. This is not an accurate assumption. That does not mean, however, that it is irrational. First, its impact is de minimis. For example, Plaintiff states that as a result of this, his monthly pension is reduced by $1.14. Second, even in the age of calculators and spread sheets, simplification has a value. Numbers get entered wrongly, and that sort of error is reduced by limiting the number of decimal points and digits that have to be entered. Third, while Plaintiff says this simplification gives a windfall to DuPont, I do not think that is the case. I believe whether the simplification benefits or harms a retiree depends mostly upon whether one of the months of the partial year is February. If it is, the retiree gains, because he or she gets a credit of 1/12 when the actual credit (in a non-leap year) should be 28/365; if it is not, the retiree either barely gains or loses, because he or she gets a credit of 1/12 when the actual non-leap year credit should be either 30/365 or 31/365. Thus, overall, these gains or losses should average out. I thus think the simplification formula is, overall, neutral. I cannot say that the Administrator abused its discretion in using the 360-days-in-a-year formula.
The second issue involves calculating Plaintiff's "primary social security benefit." Plaintiff sets forth the relevant language from the Plan. (D.I. 41, p.8). It seems to me the interpretation that the Administrator gives to the language is not only not an abuse of discretion, but is the only reasonable reading of the language. The quoted language addresses two issues. One is which social security benefits qualify. The second is how that amount will be determined. On the second issue, a number of alternatives are provided. On the first, however, there is a limitation that only social security benefits for which "service" is recognized count. That limitation applies no matter which alternative is used to make the calculation. Thus, the Administrator did not abuse its discretion in not including Plaintiff's 1966-73 earnings in calculating the "primary social security benefit," because DuPont did not pay FICA taxes on the 1966-73 earnings. Plaintiff raises in his Reply Brief another concern, stating that DuPont did not use his 2013 earnings in making the calculation. (D.I. 41, p.7). He cites to pages in the record that indeed appear not to include 2013 earnings. (D.I. 26-7 at 54-57 [R.429-32]). I cannot tell whether Plaintiff has a point here. I therefore request that the parties meet-and-confer and advise me what the results are using his actual earnings from DuPont including 2013. If there is adispute, the parties should provide calculations that I can follow to see why they come to a different result.
The third involves the Cooper formula. The parties agree that the formula (which is embodied in the Family Court order) is: AB x 50% x Months in Plan during marriage/Months in Plan = QDRO payment to ex-wife. (D.I. 37, Exh. A, p.3). The parties also agree that AB = $9,893.84, and that "Months in Plan during marriage" = 236.1 The parties disagree whether "Months in Plan" should be 487 or 439. Plaintiff says 487.4, which represents the actual number of months he was in the plan (D.I. 37, p.14); Defendants say 439 because for the last 72 months, the plan provided pension benefits that were 1/3 of what they had been, and Defendants say it should be weighted proportionately. Under Plaintiff's calculation, the formula results in a "QDRO" offset of $2,397.28; under Defendants' calculation, the formula results in an offset of $2,659.39. (D.I. 46, p.2). The difference is $262.11.2
There is some dispute about what the standard of review is on this issue. (D.I. 37, p.11; D.I. 39, p.12). In essence, Third Circuit law is that interpretation of the QDRO is reviewed de novo, but construction of the pension plan is reviewed for abuse of discretion. See Hullett v. Towers, Perrin, Forster & Crosby, Inc., 38 F.3d 107, 114-15 (3d Cir. 1994).
I have spent a lot of time trying to figure my way through the mass of calculations in the record, including those proffered by Plaintiff. I think the problem that is identified by Plaintiff isthat having a simple formula to determine what is the fairest division of his pension is not necessarily satisfactory. There are multiple variables that contribute to the formula. Two that are problematic are that (1) DuPont changed the pension calculation in 2007, essentially reducing it by two-thirds, and (2) Plaintiff's salary kept going up, so that his "Average Monthly Pay" or "High Three" also kept going up. The net result of the changes in these variables is that the formula embodied in the QDRO does not work very well.
I have been playing around with numbers trying to understand the problem here. Let's assume the following: P worked for D for 24, 27, or 36 years. After 24 years, DuPont reduced the pension plan by 2/3. He was paid $120,000 per year for 24 years. For the 27 and 36 year careers, his last three years, he was paid either $120,000 or $240,000. He and his wife divorced after the first 12 years. Thus, we can "run the numbers" for what his "average monthly pay" would be under each scenario, and using both his formula and D's formula for computing the QDRO.3
Ex. 1 is 24 years, P formula. Ex. 2 is 24 years, D formula. Ex. 3 is 27 years, last three years, $120,000, P formula. Ex. 4 is 27 years, last three years, $240,000, P formula. Exs. 5 and 6 are Exs. 3 and 4 but with D's formula. Ex. 7 is 36 years, last three years, $120,000, P formula. Ex. 8 is 36 years, last three years, $240,000, P formula. Exs. 9 and 10 are Exs. 7 and 8 but with D's formula. The results:
Years High 3 Formula Pension total QDRO Pension to Plaintiff 1. 24 120K P 3,600 $900 $2700 2. 24 120K D 3,600 $900 $2700 3. 27 120K P 3,750 $833.33 $2916.67
There are two results in this table that are irrational. They are examples 3 and 7, both of which use Plaintiff's formula, and result in the QDRO decreasing as Plaintiff's service continues to increase. The net result of this exercise is that I am convinced that DuPont's method of implementing the QDRO is truer to what...
