Halline OVERBY and Paulette M. Overby, Appellees v. NATIONAL ASSOCIATION OF LETTER CARRIERS, et al., Appellants, 389 U.S. App. D.C. 234

Opinion for the Court filed by Chief Judge SENTELLE. SENTELLE, Chief Judge: Halline Overby, an annuitant in a retirement trust fund operated by National Association of Letter Carriers (NALC), and his wife Paulette Overby brought suit in district court seeking a declaration that a purported amendment to the trust plan which would have rendered Paulette Over-by ineligible to receive benefits under the plan as a surviving spouse was not properly adopted and is therefore inoperative.

The district court found that the trustees of the plan had not submitted the amendment to the fund’s actuaries for an evaluation and estimate of its cost, as required by the governing provisions of the plan, and therefore held that the amendment was not properly adopted. NALC appeals, arguing that the district court erred both in its findings of fact and in its conclusions of law. Upon review, for the reasons more fully set forth below, we hold that the district court committed no reversible error in either its factual determinations or in its conclusions of law.

We therefore affirm the judgment of the district court. I. BACKGROUND The National Association of Letter Carriers (NALC) is a national labor union which sponsors a retirement plan, the National Association of Letter Carriers Annuity Trust Fund (ATF or plan). The ATF is a “qualifying plan” under the Employee Retirement Income Security Act of 1974, 29 U.S.C. § 1001 et seq. (ERISA). The plan provides retirement annuities for NALC’s national officers, national business agents, certain branch officers, headquarters employees, and employees of NALC’s health plan, though not postal carriers themselves.

The president of NALC serves as the plan administrator for the ATF, and the board of trustees of NALC has oversight responsibilities for the plan. Appellee Halline Overby became a letter carrier in 1960. He was elected president of his local union in 1969. Halline Overby then joined NALC’s Board of Trustees in 1978. When he joined the Board of Trustees, he also became a participant in the ATF and has remained so ever since. Over his career of working for NALC, he has been a trustee of the ATF, a member of the Executive Council, and the Assistant Secretary Treasurer.

Halline Overby retired from NALC on October 22, 1990, and began receiving his annuity payments on February 1, 1991. In May of 1991, he married Paulette Overby, his co-plaintiff in the district court and co-appellee before this court. One provision of the plan provides a benefit to a surviving spouse of a deceased annuitant, calculated at 60% of the benefits of the deceased. When Mr. Overby experieneed serious health difficulties in the late 1990s, Mrs. Overby, concerned about her own financial stability, inquired into life insurance and survivorship benefits.

NALC’s accounting office informed her that she would not be eligible to receive surviving spouse benefits because a purported amendment to the plan made in 1985 changed the definition of “surviving spouse.” The parties agree that prior to the purported amendment, the plan followed a “one-year-at-death” rule, under which the surviving spouse was “one to whom the Annuitant was married for at least one year immediately preceding the Annuitant’s death, or is the parent of issue by such marriage.” Overby v. Nat’l Ass’n of Letter Carriers, 601 F.Supp.2d 101, 103 (D.D.C.2009).

In order to qualify as a surviving spouse, under the purported amendment, a claimant must be “one to whom the Annuitant was married for at least the year immediately preceding and ending on the Annuitant’s annuity commencement date.” Id. However, the parties dispute whether the plan was effectively amended to replace the “one-year-at-death” rule with the so-called “marriage-at-commencement” methodology for determining who qualifies as a smwiving spouse.

After receiving the bad news, the Overbys brought the instant action, seeking a declaration that, inter alia, “the alleged 1985 amendment to the survivor annuity rule was not adopted in accordance with the Plan’s amendment procedures in

Article IX,

Section I, as amended, and is therefore invalid.” Pis. Compl. 16. The district court received evidence on the require-merits of the plan and the purported adoption of the surviving spouse amendment. Considering the evidence in the light of the requirements of ERISA, the court held that the amendment had not been properly adopted, that the amendment was therefore ineffective, and that Paulette Overby would qualify as a surviving spouse under the plan in its unamended form. Overby v. Nat’l Ass’n of Letter Carriers, 601 F.Supp.2d 101 (D.D.C.2009). NALC appeals. II. ANALYSIS A.

Amendment of the Plan As the district court held, ERISA requires in 29 U.S.C. § 1102(b)(3) “that every employee benefit plan ‘provide a procedure for amending such plan, and for identifying the persons who have authority to amend the plan.’ ” 601 F.Supp.2d at 108 (quoting 29 U.S.C. § 1102(b)(3)). The statute is silent as to the level of detail and as to the nature of procedural requirements, but “[t]he provision requires ... that there be an amendment procedure.” Curtiss-Wright Corp. v. Schoonejongen, 514 U.S. 73, 80, 115 S.Ct. 1223, 131 L.Ed.2d 94 (1995) (emphasis in original).

The statutory scheme further “follows standard trust law principles in dictating only that whatever level of specificity [an employer] ultimately chooses, in an amendment procedure or elsewhere, it is bound to that level.” Id. at 85, 115 S.Ct. 1223. Therefore, as the district court correctly held, “a proposed amendment not done in accordance with a plan’s amendment procedure is ineffective and does not amend a plan.” 601 F.Supp.2d at 108 (citing, inter alia, Curtiss-Wright Corp., 514 U.S. at 78, 115 S.Ct. 1223). As the statute requires, the ATF’s plan includes a formal amendment procedure.

That procedure includes the following three requirements:

(1) The trustees must first submit the proposed amendment to the Fund’s actuaries “for an evaluation and estimate of its cost;”

(2) The trustees must then “adopt” the proposed amendment; and

(3) NALC’s Executive Council must then “approve” the proposed amendment. Id. at 108. The district court held that on the evidence before it the trustees had not met the first of those requirements; that is, they had not submitted the proposed amendment to the fund’s actuaries for an evaluation and estimate of its cost. The district court therefore held that the amendment had not been validly adopted, and that the effective definition of surviving spouse would entitle Paulette Overby to that status should she survive Halline while still married to him.

Appellants do not challenge the district court’s ruling as to the nature of the amendment process required under the plan, but challenge its conclusions on two principal bases. First, they contend that the district court did not have sufficient evidence to have reached the finding that the amendment was not submitted to the actuaries, as required by the amendment process. Second, they contend that even if the court had sufficient evidence for the finding, it nonetheless erred in its conclusion that this invalidated the amendment.

Before considering appellants’ other assignments of error, we will dispose of these two principal arguments in turn. 1. The Sufficiency of the Evidence We note at the outset that an appellant seeking reversal of a trial court’s findings of fact in a bench trial faces a daunting task. Such findings, “whether based on oral or other evidence, must not be set aside unless clearly erroneous, and the reviewing court must give clue regard to the trial court’s opportunity to judge the witnesses’ credibility.” Fed.R.CivP. 52(a)(6).

We further note that “[tjhis standard applies to the inferences drawn from findings of fact as well as to the findings themselves.” Halberstam v. Welch, 705 F.2d 472, 486 (D.C.Cir.1983). Cases addressing the question of the sufficiency of evidence to support a trial court’s findings are similar. We will determine that the evidence is sufficient if a reasonable fact finder could have reached the conclusion adopted by the trial court. Anderson v.

City of Bessemer City, 470 U.S. 564, 573-74, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985) (“If the district court’s account of the evidence is plausible in light of the record viewed in its entirety, the court of appeals may not reverse it.... Where there are two permissible views of the evidence, the factfinder’s choice between them cannot be clearly erroneous.”). Appellants have not met this standard.

Appellants argue that “the District Court based its finding that the ‘married at commencement’ amendment was never submitted for an actuarial cost review on the fact that the review is not explicitly mentioned in either the April 19, 1985 trustee meeting minutes or the contemporaneous notes of NALC’s outside counsel, Jani Rachelson.” Appellants’ Br. 15.

It is true that one of the evidentiary bases relied upon by the trial court was the undeniable fact that the submission of the proposed amendment for the actuarial cost review was never mentioned at all in the minutes of the meeting at which it was purportedly adopted, or in the contemporaneous notes of the outside counsel. However, it is not true that the absence of an “explicit” reference is not evidence that it was never done. It is, of course, not direct evidence, but that does not mean it is not evidence. It is inherently difficult to prove a negative. See Elkins v.

United States, 364 U.S. 206, 218, 80 S.Ct. 1437, 4 L.Ed.2d 1669 (1960) (“[A]s a practical matter it is never easy to prove a negative.... ”). It was plaintiffs’ task below to prove that something had not been done. If that proof requires plaintiffs to produce direct evidence, then plaintiffs’ burden can virtually never be carried. Rare will be the case in which a plaintiff can offer witnesses who will testify “I saw him not do it.” Thus, if direct evidence of negatives were requii'ed, there would be little point in the law requiring any person to do anything as the failure to do it could rarely be proved.

Therefore, we must expect negatives to be proved by circumstantial evidence. Where, as hei-e, the law requires

an act to be done, and the person whom the law requires to act has maintained minutes of the time during which the act should have been done, it is certainly probative that the written record does not include a recording of the required act.

That evidence is even more probative in this case, where the district court found that “[t]he omission of [the] actuarial evaluation and estimation is telling when juxtaposed against the next subparagraph of those same minutes, in which it is recorded that before the trustees adopted” another amendment at the same meeting, the opinion of the actuaries had been obtained and the cost estimate produced by the actuaries is set forth. Overby, 601 F.Supp.2d at 109.

The district court further found it “telling” that the minutes of another meeting in December of 1985 specifically records that the fund’s actuaries undertook an actuarial evaluation of another subject matter. Id. Thus, the district court’s opinion, id. at 106-07, sets forth significantly more evidence than the appellants admit to exist. Though appellants contest the district court’s reliance on Ms.

Rachelson’s testimony, it is only after discussing the lack of record evidence before the court for the submission of the amendment to the actuaries and the conspicuous omission of such an actuarial evaluation in the face of the fund’s apparent practice with respect to other amendments that the district court expressed additional reliance on the testimony of outside counsel.

As the district court put it, “[t]hat no actuarial evaluation or estimation ever occurred with respect to the ‘marriage-at-commencement’ rule prior to its asserted ‘adoption’ is further established by the testimony [of the outside counsel].” Id. at 109 (emphasis added). Therefore, the issue before us has never been whether the testimony of the outside counsel would be sufficient standing alone, but whether the record evidence including corroboration by her lack of memory is sufficient to support the district court’s factual finding.

While we are not implying that the court’s findings would have been questionable in the absence of the deferential standard, given the deferential standard, the record is more than ample to support what the district court found and the inferences it drew from the evidence before it. The trial court’s findings were certainly not clearly erroneous. 2. The Effect of the Omission Appellants next argue that a “procedural irregularity” in the adoption of an amendment “cannot support a holding invalidating it.” Appellants’ Br. 19.

In appellants’ view, “[c]ourts should not invalidate amendments to ERISA plans that are adopted without strict adherence to plan amendment procedures unless there is evidence of bad faith regarding the amendment procedure, active concealment of the amendment itself, or plaintiffs detrimental reliance on the plan procedures.” Id. Unfortunately for appellants, that is not the law.

The Supreme Court has told us that “ERISA ... follows standard trust law principles in dictating only that whatever level of specificity a company ultimately chooses, in an amendment procedui’e or elsewhei’e, it is bound to that level.” Curtiss-Wright Corp. v. Schoonejongen, 514 U.S. 73, 85, 115 S.Ct. 1223, 131 L.Ed.2d 94 (1995). While arguably the Curtiss-Wnght statement qxxalifies as dicta, “carefully considei'ed language of the Supreme Court, even if technically dictum, generally must be treated as authoritative.” U.S. v. Dorcely, 454 F.3d 366, 375 (D.C.Cir.2006) (quotation omitted).

This is especially so hei’e as the Supreme Court has reitei'ated the same teaching in Inter-Modal Rail Employees Ass’n v. Atchison, Topeka and Santa Fe Railway Co., 520 U.S. 510, 515-16, 117 S.Ct. 1513, 137 L.Ed.2d 763 (1997) (“An employer may, of course, retain the unfettered right to alter its promises, but to do so it must follow the formal proeedxu-es set forth in the plan.”) (discussing ERISA plans). The clear implication of the Supi’eme Court’s language is that there must be amendment procedures in a plan, and those amendment proeedui*es must be followed for the valid adoption of an amendment.

Oxxr sibling circxiits follow this view of the law with near unanimity. In Coffin v. Bowater Inc., 501 F.3d 80, 91-92 (1st Cir.2007), the First Circuit stated that “an ERISA plan amendment must be in writing; it must be executed by a party authorized to amend the plan; the language of the amendment must clearly alert the parties that the plan is being amended; and the amendment must meet any other l-equix-ements laid out for such amendments in the plan’s governing documents.” Similarly, in Halliburton Co. Benefits Committee v.

Graves, 463 F.3d 360, 371-72 (5th Cir.2006), the Fifth Circuit opined that “[i]n order to amend a welfare benefit plan governed by ERISA, the employer must provide a procedure for amending such plan, and for identifying the persons who have authority to amend the plan.... [0]nly an amendment executed in accordance with the plan’s procedures is effective.” (internal citations and quotations omitted). The Third Circuit in Depenbrock v. Cigna Corp., 389 F.3d 78, 82-83 (3d Cir.2004), held that “an amendment is ineffective if it is inconsistent with the governing instruments ....

ERISA specifies that a valid amendment can only be made in the manner specified in the plan document.” Other circuits have acted consistently. See Winterrowd v. Am. Gen. Annuity Ins. Co., 321 F.3d 933, 937 (9th Cir.2003) (“Section 402 of ERISA requires employee benefit plans to specify both an amendment procedure and a procedure for identifying persons with authority to amend. These amendment procedures, once set forth in a benefit plan, constrain the employer from amending the plan by other means.”) (citation omitted); Aramony v.

United Way Replacement Benefit Plan, 191 F.3d 140, 148—49 (2d Cir.1999) (looking to corporate formalities to see whether a pension plan was amended); Miller v. Coastal Corp., 978 F.2d 622, 624 (10th Cir.1992) (“ERISA requires all modifications to an employee benefit plan ... to conform to the formal amendment procedures.... ”); Coleman v. Nationwide Life Ins. Co., 969 F.2d 54, 58-59 (4th Cir.1992) (“[A]ny modification to a plan must be implemented in conformity with the formal amendment procedures and must be in writing.”); Confer v.

Custom Eng’g Co., 952 F.2d 41, 43 (3d Cir.1991) (“Only a formal written amendment, executed in accordance with the Plan’s own procedure for amendment, could change the [ERISA medical] Plan.”); Albedyll v. Wi. Porcelain Co. Revised Ret. Plan, 947 F.2d 246, 254-55 (7th Cir.1991) (holding that an ERISA pension plan amendment was invalid because it did not comply with the plan’s procedures for amendment).

In the face of this otherwise unanimous array, appellants ask us to follow what they call “the leading case on the proper standard for determining whether an ERISA plan amendment should be invalidated.” Appellants’ Br. 19. The case appellants cite is Loskill v. Barnett Banks, Inc. Severance Pay Plan, 289 F.3d 734 (11th Cir.2002). Concededly, that ease appears to have reasoned that bad faith was necessary for an invalidation of an amendment adopted in disregard of the procedures required by the plan. We cannot, however, see what makes it “the leading case” when no other court has ever followed it. Nor will we.

Indeed, even the Eleventh Circuit seems to have ignored the holding of Loskill. See Shaw v. Conn. Gen. Life Ins. Co., 353 F.3d 1276, 1283 (11th Cir.2003) (invalidating a purported amendment to an ERISA-covered disability plan because of lack of compliance with plan amendment procedures without any discussion of bad faith, active concealment, or detrimental reliance). The Supreme Court’s guidance on the necessity of amendment procedures drives us toward a conclusion that those procedures should not be ignored.

Likewise, the near unanimous conclusions of our fellow circuits weighs heavily against the novel construction sought by appellants. Appellants give us no reason why we should treat the written procedures of the plan so lightly, nor can we think of any. An amendment procedure is there to be followed. It is there to give fair notice to the beneficiary under the plan. We have already upheld the district court’s finding that it was not followed in this case.

In short, we adopt the near unanimous view of the other circuits that a failure to follow the amendment procedures of a plan invalidates an amendment without regard to a showing of bad faith. We therefore uphold the conclusion of the district court that the effect of the failure in this case renders the amendment invalid. 3. Appellants’ Other Arguments Appellants offer three other arguments, none of which warrant reversal. Appellants first argue that “[l]aek of an actuarial cost review was not a central theory of the Overbys’ case.” Appellants’ Br. 18.

They assert that “the plaintiffs did not rely heavily on the theory that a lack of actuarial cost review made the amendment invalid.” Id. (emphasis added). They argue in a single paragraph that the Overbys’ “main theory” was that no vote on the amendment took place at all. While this is at least arguably true, it is without legal effect.

While it is no doubt true that the complaint and the evidence of a plaintiff must be sufficient to put defendants on notice of any theory of recovery upon which the plaintiff is relying, this does not preclude the possibility of plaintiffs arguing alternative theories, nor the possibility of the court’s relying upon a “lesser” alternative in its decision. Appellants offer no authority for a contrary position, nor have we independently found any.

The complaint, the discovery, the evidence at trial, and the complete record established in the district court taken together — indeed taken separately— provided defendants with ample notice' that plaintiffs sought to have the amendment invalidated because of the trustees’ failure to follow the plan’s amendment procedure in the adoption of the disputed amendment. More specifically, the Overbys asked for discovery on the actuarial review. Appellants’ counsel admitted on appeal that these requests for admissions were in the record and that the actuarial review came up in testimony as well. See Trial Tr. at 130.

It is abundantly clear that appellants were on notice that this argument was before the district court and could have presented evidence at trial that NALC complied with the amendment procedures if any such evidence existed. We have no precedent for requiring any more. Appellants further offer two brief arguments. They first contend that the amendment to the plan was necessary to comply with a change in the law required by an amendment to the governing statute effected in the Retirement Equity Act of 1984.

That amendment required plans qualifying under ERISA to include a “married at commencement” eligibility requirement. Again, this is inarguably true. See Retirement Equity Act of 1984, Pub.L. No. 98-397, 98 Stat. 1426. This amendment did not, however, require the deletion of pre-existing “one year at death” language, nor free the trustees from complying with the amendment procedures required by the plan. Whether the resulting invalidity of the amendment means that the two methods of qualification must coexist or that the amended statute mandates further amendment to the plan is not before us.

Appellants are no doubt correct that such coexistence will increase the financial burden on the plan. That well illustrates a problem that could have been addressed by the submission of the proposed amendment for review by the actuaries as required by the amendment procedures of the plan. Last, appellants argue that the district court should have narrowly tailored the relief ordered in the judgment.

That is, they contend that the court should have limited the effect of its findings and holdings to the claims of the Overbys, leaving the questions of the validity of the amendment to other annuitants for later cases. We cannot conclude that the court erred in granting the relief it provided. The district court performed its

Article III function. The judge decided the case before her. The precise breadth and strength of the preclusive effect of that judgment on later litigants can await cases in which that preclusive effect is at issue. CONCLUSION We conclude that the district court’s factual finding that the proposed amendment was not submitted to the actuaries for a cost review is not clearly erroneous. Because an ERISA plan is held to whatever level of specificity it has adopted, the failure of NALC ATF to follow its own amendment procedures means that the amendment was not adopted.

For the reasons set forth above, the judgment of the district court is therefore affirmed. So ordered. . Unless the context indicates otherwise, "NAI.C” refers collectively to the four defendant-appellants in this case: the National Association of Letter Carriers, the National Association of Letter Carriers Annuity Trust Fund, William H. Young, and the Board of Trustees of the National Association of Letter Carriers Annuity Trust Fund. .

While there is a discrete exception qualifying spouses who were married to the annuitant within the year before the commencement of the annuity and for at least one year ending on or before the annuitant’s death, that exception has no application to the facts before us.