Opinion
Civil Action No. 19-11437-FDS
08-05-2020
SCOTT BELKNAP, on behalf of himself and all others similarly situated, Plaintiff, v. PARTNERS HEALTHCARE SYSTEM, INC.; THE PENSION MANAGEMENT COMMITTEE; THE RETIREMENT COMMITTEE; and JANE/JOHN DOES 1-5, Defendants.
ORDER ON DEFENDANTS' MOTION TO DISMISS THE AMENDED COMPLAINT SAYLOR, C.J.
This is a putative class action under the Employee Retirement Income Security Act ("ERISA"). Plaintiff Scott Belknap is a former employee of defendant Partners Healthcare System, Inc. He retired early from Partners at age 62 and now receives a type of retirement benefit known as a joint and survivor annuity, which covers both him and his spouse.
Belknap has filed suit on behalf of himself and all others similarly situated, alleging that the way in which Partners calculates the value of his benefit violates ERISA. Specifically, he alleges that Partners uses obsolete actuarial information from 1951 and outdated interest rates, which reduce the value of his annuity and thus violate 29 U.S.C. § 1054(c)(3). In simple terms, § 1054(c)(3) requires that an employee's retirement benefit be valued in terms of the normal benefit that it will yield at normal retirement age, and if it is paid at any other time or in any other form, it must be the "actuarial equivalent" of that normal retirement age benefit.
Originally, the complaint in this action alleged that the way in which Partners calculates Belknap's annuity violated several provisions of ERISA. Partners moved to dismiss the claims in the original complaint on several grounds, including on the ground that § 1054(c)(3) does not require that actuarial equivalence be calculated using reasonable assumptions, leaving employers free to calculate it how they please.
The factual and procedural background of this action are set forth in the Court's memorandum and order on defendant's motion to dismiss Belknap's original complaint. See generally Memorandum and Order on Defendant's Motion to Dismiss (Dkt. No. 33). To the extent that the factual allegations in Belknap's amended complaint differ from his original complaint in ways that matter here, they are set forth below.
The Court granted that motion in part and denied it in part. See generally Memorandum and Order on Defendant's Motion to Dismiss (Dkt. No. 33). As is relevant here, it found no basis in the text of § 1054(c)(3) to require that an "actuarial equivalent" be based on reasonable assumptions. See id. at 11-17. That was because § 1054(c)(3) contains no such reasonableness requirement, while other provisions in ERISA expressly do, which indicates that this omission was deliberate. See Mertens v. Hewitt Assocs., 508 U.S. 248, 254 (1993). Nonetheless, the Court observed that § 1054(c)(3) still requires Belknap's early retirement benefit to be the "actuarial equivalent" of the normal retirement benefit under his plan. And it found that the definition of that term was far from clear. Accordingly, the Court invited the parties to submit supplemental briefing as to the meaning of the term "actuarial equivalent" under § 1054(c)(3) or otherwise propose a framework for resolving that issue.
The parties jointly proposed that Belknap would respond to the issues raised in the Court's memorandum and order by filing an amended complaint, to which Partners could respond. The Court accepted that proposal. On March 3, 2020, Belknap filed an amended complaint that names several additional defendants. (Am. Compl. ¶¶ 18-21). Defendants have now moved to dismiss the amended complaint under Rule 12(b)(6) for failure to state a claim. For the following reasons, the motion will be denied.
According to the amended complaint, Partners calculated the value of part of Belknap's early retirement benefit using a 7.5% discount rate and a mortality table from 1951. (Id. ¶¶ 3, 62-73). The amended complaint alleges that because those inputs are outdated and unreasonable, his benefit is not actuarially equivalent to the benefit he would have received if he retired at his plan's normal retirement age of 65. (Id.). Defendants contend that because the text of § 1054(c)(3) does not require that actuarial equivalence be calculated using reasonable inputs, the amended complaint does not state a claim for a violation of that provision.
According to the amended complaint, another part of Belknap's retirement benefits is calculated using a 7.5% discount rate and the applicable mortality table established by the Internal Revenue Service. (Am. Compl. ¶ 3). The amended complaint alleges that because that discount rate is unreasonable, this portion of Belknap's benefits is also not actuarially equivalent to those he would have received if he retired at age 65. (Id. ¶ 9).
As the Court previously found, § 1054(c)(3) does not—at least on its face—require an "actuarial equivalent" to be calculated using "reasonable" assumptions. And the Court must assume that the statutory omission was deliberate. See Mertens, 508 U.S. at 254. Nevertheless, § 1054(c)(3) does require that the two benefit forms be "actuarial equivalent[s]." That term must mean something. But despite using it, "ERISA does not further define actuarial equivalence." Stephens v. U.S. Airways Grp., Inc., 644 F.3d 437, 440 (D.C. Cir. 2011). Presumably, then, "Congress intended that term of art to have its established meaning." Id. (citing McDermott Int'l, Inc. v. Wilander, 498 U.S. 337, 342 (1991)). Thus, the question becomes ascertaining the "established meaning" of "actuarial equivalence." See id.
One possibility is that "actuarial equivalence" simply requires that two benefit forms be equal when compared using a common set of assumptions—that is, one discount rate and one mortality table. The D.C. Circuit embraced this view in Stephens, finding that "[t]wo modes of payment are actuarially equivalent when their present values are equal under a given set of actuarial assumptions." Id. (citing JEFF L. SCHWARTZMANN & RALPH GARFIELD, EDUCATION & EXAMINATION COMM. OF THE SOC'Y OF ACTUARIES, ACTUARIALLY EQUIVALENT BENEFITS 1, EA1-24-91 (1991), archived at https://perma.cc/A7U2-G3M8). But the Stephens court identified this "established meaning" of "actuarial equivalence" as a term of art based on only one source: a set of "study notes to persons preparing for the examinations of the Society of Actuaries" that "do not . . . represent any official opinion, interpretations, or endorsement of the Society of Actuaries." Id.; SCHWARTZMANN & GARFIELD, supra, at 1.
It is true, of course, that statutory interpretation is a matter of law for the court to decide. And it may well be that those study notes reflect the accepted meaning in the industry of the term "actuarial equivalence." Nonetheless, prudence would suggest that the Court should attempt to explore the issue in greater depth before resting its decision on such an apparently doubtful source. Indeed, the amended complaint cites several other sources that appear to describe actuarial equivalence differently. (See Am. Compl. ¶¶ 23-29). Furthermore, because the definition of actuarial equivalence in the study notes is apparently unmoored from any concept of reasonableness, it could conceivably produce absurdly unfair results. As Belknap argues, under such a definition, "[a]n employer could calculate early retirement benefits using the mortality table that Sir Edmond Halley developed in 1693, coupled with a corporate bond rate from the early 1980s at around 15 per cent." (Pl. Opp. at 11). Surely, Congress intended the "actuarial equivalence" requirement of § 1054(c)(3) to provide some degree of protection to beneficiaries, and not to permit employers to use any assumptions they chose, no matter how outmoded or inapt.
Belknap proposes a different definition, but that, too, raises potential problems. The amended complaint, citing Stephens, alleges that "'[a]ctuarial equivalence' means that two forms of payment have an equal present value under a given set of assumptions." (Am. Compl. ¶ 24). But it further alleges that "[a]ctuarial equivalence requires mortality and interest rate assumptions that are current and reasonable when a participant starts receiving benefits." (Id. ¶ 25). It cites several sources in support of that interpretation, including actuarial standards of practice issued by the American Academy of Actuaries. (Id. ¶¶ 23-29). One federal district court has recently taken a similar view. See Smith v. Rockwell Automation, Inc., 438 F. Supp. 3d 912, 921 (E.D. Wis. 2020) (concluding that § 1054(c)(3)'s actuarial equivalence requirement "means that plans must use the kind of actuarial assumptions that a reasonable actuary would use at the time of the benefit determination").
Notably, it does not appear that the parties in that case disputed whether actuarial equivalence must be calculating using reasonable actuarial inputs. Rather, they disagreed over whether those inputs must be reasonable at the time that the value of a retiree's benefit is determined or need only be "reasonable at the time they were written into the plan." Smith, 438 F. Supp. 3d at 915-16.
That interpretation has the benefit of requiring that the actuarial assumptions be both current and reasonable, which seems perfectly sensible on its face. Again, however, that interpretation reintroduces the concept of "reasonableness" into the statute, which runs counter to basic principles of statutory interpretation. If "actuarial equivalence" necessarily requires the use of reasonable inputs, then it seems unnecessary to require employers to use "reasonable" "actuarial assumptions and methods," as Congress did elsewhere in ERISA. See, e.g., 29 U.S.C. §§ 1085a(c)(3)(A), 1393(a)(1). There may be a way to harmonize the different provisions of the statute, but at a minimum, that may require further development of the record.
For example, it may be that Congress thought it unnecessary to further clarify what "actuarial equivalent[s]" meant as a term of art among actuaries, but considered it important to state that "actuarial assumptions"—a term that perhaps does not have an established meaning—had to be "reasonable." Compare 29 U.S.C. § 1054(c)(3) with 29 U.S.C. §§ 1085a(c)(3)(A), 1393(a)(1).
Defendants further contend that even if calculating "actuarial equivalence" requires using reasonable inputs, Belknap's "own factual allegations here demonstrate that [he] is receiving more than he would have received from what [he] claims are 'reasonable' actuarial assumptions." (Def. Mem. at 9). Specifically, they contend that in his original complaint and prior briefing, he admitted that calculating the value of his early retirement benefits using reasonable actuarial assumptions would still have reduced them by 28.7%. And they contend that because the amended complaint now indicates that the way in which Partners calculates his benefits reduces them by less than 28.7%, those calculations must be reasonable and it does not allege a violation of § 1054(c)(3).
The amended complaint will not be dismissed on that basis. Defendants essentially ask for Belknap to be held to the position that any inputs that reduced his early retirement benefit by 28.7% or less would have been reasonable per se. They do not specify whether they ask for that result under the doctrine of judicial estoppel or as a judicial admission. (See Def. Mem. at 9). At a minimum, both doctrines require some degree of clarity. Perry v. Blum, 629 F.3d 1, 9 (1st Cir. 2010) (explaining that judicial estoppel is only appropriate where, among other things, a party takes "clearly inconsistent" positions"); Harrington v. City of Nashua, 610 F.3d 24, 31 (1st Cir. 2010) ("To be binding, a judicial admission must be 'clear.'"). And it is not clear that Belknap actually took that position. He asserted that the use of one example set of inputs, which resulted in a 28.7% reduction, would have been reasonable, not that any reduction of less than 28.7% was reasonable no matter the inputs. (See Pl. Sur-Reply (Dkt. No. 27) at 3). Moreover, the allegations in his original and amended complaints are not obviously inconsistent. Both allege that outdated inputs used by Partners to value his retirement benefits were unreasonable. To be sure, the allegations about the precise numeric consequences of those inputs have changed. But that appears to be because Belknap has obtained new information from Partners about how his benefits are valued. (Pl. Opp. at 15-16). And, in any event, his position ultimately remains the same: the actuarial assumptions used by Partners were unreasonable.
In summary, there is no clear answer, at least at this stage of the proceeding, as to what is necessary for two retirement benefit forms to be "actuarial equivalent[s]" as required by 29 U.S.C. § 1054(c)(3). ERISA does not define that term, and courts have yet to agree on a definition. Compare Stephens, 644 F.3d at 440 (defining actuarial equivalence as having present values that are "equal under a given set of actuarial assumptions") with Smith, 438 F. Supp. 3d at 921 (interpreting it to "mean[] that plans must use the kind of actuarial assumptions that a reasonable actuary would use at the time of the benefit determination") (emphasis added). Nor is it clear, at least on this record, whether "actuarial equivalence" is in fact a term of art, or how actuaries themselves would interpret the term. It may be that expert testimony on the topic is required to resolve the issue. In any event, the Court cannot resolve the issue, on this record, on a motion to dismiss.
Accordingly, and for the foregoing reasons, the motion to dismiss is DENIED.
So Ordered.
/s/ F. Dennis Saylor IV
F. Dennis Saylor IV
Chief Judge, United States District Court Dated: August 5, 2020