DAVIS v. United States

Opinion

In the United States Court of Federal Claims No. 24-364 (Filed: May 20, 2026)

************************* RODNEY L. DAVIS, et al.,

Plaintiffs,

v.

THE UNITED STATES,

Defendant.

*************************

Kenneth T. Cuccinelli, II, Spotsylvania, VA, for plaintiffs. Earl N. Mayfield, III, Fairfax, VA, of counsel.

Galina I. Fomenkova, Senior Trial Attorney, United States Department of Justice, Commercial Litigation Branch, Civil Division, Washington, DC, with whom were Stanley E. Woodward Jr., Associate Attorney General, Brett A. Shumate, Assistant Attorney General, Patricia M. McCarthy, Director, and Albert S. Iarossi, Assistant Director, for defendant.

OPINION

BRUGGINK, Senior Judge.

Pending are plaintiffs’ motion for partial summary judgment and the government’s cross-motion for summary judgment. The opposing motions raise many issues, most of which are of first impression. The crux of this dispute hinges on the interplay between the “Ascertainment Clause” found in Article I, Section 6, Clause 1 of the United States Constitution, the Twenty-Seventh Amendment to the United States Constitution, the automatic cost-of-living adjustment (“COLA”) provisions of the Ethics Reform Act of 1989 (“ERA”), and Congress’s repeated rejection of COLAs that otherwise would have automatically taken effect under the ERA.

1 Plaintiffs assert that each time Congress enacts legislation that blocks an otherwise automatic COLA under the ERA from taking effect (“COLA- blocking legislation” or “COLA-blocking law(s)”), it has passed a law varying congressional compensation that is subject to the requirements of the Twenty-Seventh Amendment. In plaintiffs’ view, Congress has enacted such COLA-blocking legislation in a manner inconsistent with the Twenty- Seventh Amendment twenty times since the Amendment was ratified in 1992, including every year since 2009. Thus, plaintiffs assert that they are entitled to backpay resulting from those missed COLAs starting March 7, 2018. 1

Among other things, the pending motions call on this court to address whether it is appropriate to allow these plaintiffs to recover their asserted damages and, if so, whether COLA-blocking legislation is subject to the requirements of the Twenty-Seventh Amendment. To that end, and recognizing that it would be unwise to address in one go the myriad of potential issues raised by the briefing, we directed the parties to focus oral argument on the following four “comprehensive defenses” the government raised in its briefing: (1) whether plaintiffs are estopped from recovering their asserted damages; (2) whether the Twenty-Seventh Amendment applies to laws decreasing Congressional compensation; (3) whether COLA-blocking legislation “varies” Congressional compensation; and (4) whether the Twenty-Seventh Amendment renders noncompliant legislation void ab initio or operates merely as a timing provision, in which case the question becomes whether and to what extent COLA-blocking legislation can be retroactively applied.

Consistent with our approach to oral argument, we limit this opinion to those four issues. For the reasons set out below, we hold that (1) plaintiffs are not estopped from recovering their asserted damages, (2) that the Twenty- Seventh Amendment by its terms applies to laws decreasing Congressional

1 We previously rejected plaintiffs’ argument that the continuing claims doctrine applied to this case and held that “[a]ny claim originating prior to March 7, 2018[,] is barred and COLAs nullified more than six years earlier are not resurrected for the 2018–2024 period.” Davis v. United States, 174 Fed. Cl. 563, 579 (2024). Although we do not address that issue in this opinion, as we explained at oral argument and discuss further below, we treat the issue as unresolved. Therefore, that question is hereby reopened, and we expect it to be the subject of further briefing and oral argument.

2 compensation, (3) that COLA-blocking legislation “varies” Congressional compensation, and (4) that the Twenty-Seventh Amendment renders non- compliant legislation ineffective to the extent such laws seek to effectuate a change in congressional compensation before an election intervenes. We defer ruling on all other issues raised by the parties’ motions pending further briefing.

BACKGOUND

I. The Mechanics of Congressional Pay

We begin with the salient constitutional and statutory provisions. We will then discuss the relevant (albeit scarce) case law.

A. The Constitution

The Constitution states that “[t]he Senators and Representatives shall receive a Compensation for their Services, to be ascertained by Law, and paid out of the Treasury of the United States.” U.S. Const. art. I, § 6, cl. 1. 2 Thus, consistent with Congress’s power of the purse, the responsibility of determining Congress’s pay is left in the hands of Congress. For most of our nation’s history, this “Ascertainment Clause” was the only constitutional provision that concerned congressional pay.

That changed, however, on May 7, 1992, when “Michigan’s approval of the proposed Twenty-Seventh Amendment supplied the necessary 38th state to bring about ratification, officially enshrining James Madison’s compensation Amendment two hundred years after he had proposed it.” Davis v. United States, 174 Fed. Cl. 563, 567 (2024). The Twenty-Seventh Amendment states that “[n]o law, varying the compensation for the services of the Senators and Representatives, shall take effect, until an election of Representatives shall have intervened.” U.S. Const. amend. XXVII. Therefore, while it is within Congress’s power to set its own compensation, it now must do so in a manner consistent with the Twenty-Seventh Amendment.

2 This provision is often referred to as the “Ascertainment Clause.” As will become relevant later, Article I, Section 6, Clause 1 of the Constitution also contains the “Speech or Debate Clause.”

3 B. The ERA

The relevant statutory scheme for Congressional pay arises from the 1989 enactment of the ERA, which “overhauled compensation and ethics rules for all three branches of government.” Beer v. United States, 696 F.3d 1174, 1177 (Fed. Cir. 2012) (en banc). 3 In exchange for rather drastically limiting outside income, see ERA, Pub. L. No. 101-194, § 601, 103 Stat. 1716, 1760–61 (1989) (now codified at 5 U.S.C. § 13143), the ERA granted “an immediate one-time salary increase and in subsequent years an annual COLA.” Boehner v. Anderson, 30 F.3d 156, 158 (D.C. Cir. 1994) (citing ERA, Pub. L. No. 101-194, 103 Stat. 1716 (1989) (codified at 2 U.S.C. § 4501 and 5 U.S.C. § 5318 note)). 4

The COLA provisions of the ERA are codified at 2 U.S.C. § 4501, which states:

The annual rate of pay for [members of Congress]–

...

shall be the rate determined for such positions under chapter 11 of this title, as adjusted by paragraph (2) of this section.

(2)(A) Subject to subparagraph (B), . . . each annual rate referred to in paragraph (1) shall be adjusted by an amount, rounded to the nearest multiple of $100 (or if midway between multiples of $100, to the next higher multiple of $100), equal to the percentage of such annual rate which corresponds to the most recent percentage change in the ECI (relative to the date described in the next sentence), as determined under section 704(a)(1) of the Ethics Reform Act of 1989. The appropriate date under this sentence is the first day of the fiscal year in which such adjustment in the rates of pay under the General

3 Although originally only applicable to Representatives, in 1991, Congress extended the ERA to include Senators. See Legislative Branch Appropriations Act, 1992, Pub. L. No. 102-90, § 6(a), 105 Stat. 447, 450 (1991) (codified at 5 U.S.C. § 5318 note). 4 2 U.S.C. § 4501 was previously codified at 2 U.S.C. § 31. For simplicity’s sake, we treat that section as though it was codified at 2 U.S.C. § 4501 at all relevant times.

4 Schedule takes effect.

(B) In no event shall the percentage adjustment taking effect under subparagraph (A) in any calendar year (before rounding), in any rate of pay, exceed the percentage adjustment taking effect in such calendar year under [5 U.S.C. § 5303] in the rates of pay under the General Schedule.

2 U.S.C. § 4501(1)–(2). Thus, Congressional COLAs are determined by the formula found in Section 704(a)(1) of the ERA, “rounded to the nearest multiple of $100.” 2 U.S.C. § 4501(2)(A). That formula is tied to the Employment Cost Index (“ECI”) 5 subject to the limitation that no COLA can exceed five percent. See ERA, Pub. L. No. 101-194, § 704(a)(1), 103 Stat. 1716, 1769 (1989) (codified at 5 U.S.C. § 5318 note).

That is not the only limitation, however, because Section 4501 also provides that no congressional COLA can exceed the adjustment in rates of pay under the General Schedule (“GS”), applicable to most federal civilian employees, which is determined by a similar ECI-based formula under 5 U.S.C. § 5303. 2 U.S.C § 4501(2)(B). In effect, Section 4501 dictates that members of Congress will receive an automatic COLA every year based on the lesser of the percentage increase set by the formula found in Section 704(a)(1) of the ERA (codified at 5 U.S.C. § 5318 note) and the percentage increase taking effect under 5 U.S.C. § 5303. 6 This is important because, in addition to its ECI-based formula, Section 5303 grants the President the authority to change the pay adjustment that would otherwise take effect under that section, even down to zero, if, in the exercise of his discretion, the President determines that such adjustment is inappropriate due to a “national emergency or serious economic conditions affecting the general welfare.” 5 U.S.C. § 5303(b). The minute differences between the formula dictated by 2 U.S.C. § 4501 and that found in 5 U.S.C. § 5303 are

5 The Employment Cost Index (“ECI”) is a measure of employee wages and benefits on a national scale and is reported quarterly by the Department of Labor. 6 Because Section 4501 calls for a COLA equal to the lesser of the percentage increase under Section 704(a)(1) of the ERA and the percentage increase taking effect under 5 U.S.C. § 5303, and because Section 704(a)(1) of the ERA caps any COLA at five percent, congressional COLAs under Section 4501 will never exceed five percent.

5 not particularly germane. What is relevant for our purposes, however, is that, assuming the ECI increases, and unless the President of the United States intervenes under 5 U.S.C. § 5303(b), a congressional COLA will take effect every year by operation of law pursuant to a definite formula.

Concerning timing, Congressional COLAs take effect “at the beginning of the first applicable pay period commencing on or after the first day of the month in which an adjustment takes effect under [5 U.S.C. § 5303] in the rates of pay under the General Schedule.” 2 U.S.C § 4501(2)(A). “The official table of GS rates (reflecting the changes under § 5303) is announced in an executive order in the December preceding the January in which the new rates take effect.” Davis, 174 Fed. Cl. at 568. This increase takes effect on the “first day of the first applicable pay period beginning on or after January 1 of each calendar year.” 5 U.S.C. § 5303(a). Therefore, unless Congress changes the law, or the president intervenes to manually determine the GS rate increase, congressional COLAs will automatically take effect every year pursuant to the process explained above.

C. COLA-Blocking Legislation

The object of the ERA was, in large part, the laudable goal of taking politics “out of the messy process for determining pay for Congress, the Judiciary, and all civilian federal employees.” Davis, 174 Fed. Cl. at 568; see also Boehner v. Anderson, 809 F. Supp. 138, 141 (D.D.C. 1992) (explaining that implementing this “incredible piece of legislation . . . was truly a courageous act for the Congress and the President”). By exchanging the opportunity to earn most outside income with an immediate pay raise and the promise of future COLAs, the idea was to “remove[] the necessity of members casting public votes to set their own salary because the ERA would automatically adjust member pay each year according to its COLA formula.” Davis, 174 Fed. Cl. at 568; see also Boehner, 809 F. Supp. at 141 (describing the ERA as “a comprehensive package of ethics reform, which includes a series of measures which are interrelated, most notably the cost-of-living and the pay provisions on the one hand and the prohibition of honoraria on the other” (citation omitted)).

Unfortunately, the goal was not achieved. In 1993, a mere two years after the ERA went into effect, Congress started what would become a habit of enacting COLA-blocking legislation, denying itself the 1994 COLA that would have otherwise gone into effect automatically. Emergency Unemployment Compensation Amendments of 1993, Pub. L. No. 103-6, § 7,

6 107 Stat. 33, 35 (1993). In the thirty-four calendar years since the Twenty- Seventh Amendment was ratified in 1992, Congress has denied itself COLAs in twenty-three of them. 7 Ida A. Brudnick, Cong. Rsch. Serv., RL 97-1011, Salaries of Members of Congress: Recent Actions and Historical Tables 2 (Updated Jan. 21, 2026), https://www.congress.gov/crs_external_products/RS/PDF/97-1011/97- 1011.100.pdf. Congress has denied every COLA since 2009, which has caused general congressional pay to remain stagnant at $174,000 per year since then. Id. at 1; see also Davis, 174 Fed. Cl. at 569.

Over the years, Congress has used different language in its COLA- blocking legislation. For example, the legislation that prevented the 1995 COLA from taking effect directed the COLA under 2 U.S.C § 4501 to be calculated as if no adjustment to the GS pay rates had taken effect that year (even though the GS rates were, in fact, adjusted). Treasury, Postal Service and General Government Appropriations Act, 1995, Pub. L. No. 103-329, § 630(a)(2), 108 Stat. 2382, 2424 (1994); see also Davis, 174 Fed. Cl. at 569. Thus, because Section 4501 caps congressional COLAs at the amount of the GS pay rate increase, and because the COLA-blocking legislation required the Section 4501 COLA to be calculated as though no GS rate increase had occurred, Congress’s COLA was capped at zero. See Davis, 174 Fed. Cl. at 569.

In later years, instead of artificially tinkering with a variable in Section 4501’s COLA formula, Congress has more directly denied COLAs. For example, the legislation that blocked the 2010 COLA stated that the COLA “scheduled to take effect under [2 U.S.C § 4501] in calendar year 2010 shall not take effect.” Omnibus Appropriations Act, 2009, Pub. L. No. 111-8, Div.

7 In the briefing, plaintiffs assert challenges to 20 of those COLA-blocking pieces of legislation, which blocked COLAs for the following years: 1995– 97, 1999, 2007, 2010, and 2012–25. Pls.’ Mem. Supp. Mot. Summ. J. 2–5, ECF No. 47-1. Plaintiffs’ motion for partial summary judgment was filed in September of 2025. On November 12, 2025, Congress enacted COLA- blocking legislation denying itself the 2026 COLA. Continuing Appropriations, Agriculture, Legislative Branch, Military Construction and Veterans Affairs, and Extensions Act, 2026, Pub. L. No. 119-37, Div. C, § 210, 139 Stat. 495, 584 (2025). Because no election could have intervened in an odd year, we also address in this opinion the 2025 COLA-blocking legislation that blocked the 2026 COLA from taking effect.

7 J, § 103, 123 Stat. 524, 988. And since 2013, COLA-blocking legislation has adopted the language that “no adjustment shall be made under [2 U.S.C § 4501] (relating to cost of living adjustments for Members of Congress) during [the specified] fiscal year.” E.g., Consolidated Appropriations Act, 2023, Pub. L. No. 117-328, § 6, 136 Stat. 4459, 4462 (2022).

As we noted in our prior opinion, the typical COLA-blocking process can be seen with respect to 2024. Davis, 174 Fed. Cl. at 569. The president’s executive order under 5 U.S.C. § 5303 setting the GS rates for 2024 was issued on December 21, 2023. Id. (citing Exec. Order No. 14,113, 88 Fed. Reg. 89,259 (Dec. 21, 2023)). Accordingly, the new GS rates for 2024 took effect on the “first day of the first applicable pay period beginning on or after January 1 of each calendar year,” 5 U.S.C. § 5303(a), which was January 14, 2024, Kiran A. Ahuja, U.S. Off. of Pers. Mgmt., CPM 2023–20, Memorandum for Heads of Executive Departments and Agencies—January 2024 Pay Adjustments 1 (2023), https://chcoc.gov/sites/default/files/January- 2024-Pay-Adjustment-Memo-Attachments.pdf. To prevent the 2024 COLA from automatically taking effect, however, Congress enacted COLA- blocking legislation on March 23, 2024. Further Consolidated Appropriations Act, 2024, Pub. L. No. 118-47, § 7, 138 Stat. 460, 461.

Since 2009, the timing of the executive orders setting the GS rates and the effective date of those rates has followed a consistent pattern. The president’s executive order is typically issued in December, and the new GS rates take effect the following January. The timing of COLA-blocking legislation, however, has varied widely—being enacted as early as March 2009 to prevent the January 2010 COLA, see Omnibus Appropriations Act, 2009, Pub. L. No. 111-8, 123 Stat. 524, or as late as March 2018 to block the January 2018 COLA, see Consolidated Appropriations Act, 2018, Pub. L. No. 115-141, 132 Stat. 348.

D. Relevant Cases Interpreting the Twenty-Seventh Amendment

1. Boehner

In 1994, the D.C. Circuit in Boehner faced the question of whether the COLA provisions of the ERA violated the Twenty-Seventh Amendment. 30 F.3d at 158. In that case, former Speaker of the House John Boehner argued that each COLA under the ERA had to be considered a separate law that took effect the year it became payable, thus subjecting each individual COLA to the requirements of the Ascertainment Clause (thereby requiring bicameral

8 passage and presentment to the President) and the Twenty-Seventh Amendment. See id. at 161. The D.C Circuit rejected this argument and held that the ERA’s COLA provisions did not violate the Twenty-Seventh Amendment. Id. at 161–63. Focusing on the plain language of the Amendment, the court emphasized that the Amendment’s operation is simple:

[I]t conditions the operation of a law varying congressional compensation upon an election of Representatives and the expiration of the Congress that voted for it. The law may be enacted at any time; when an election has been held the first condition is fulfilled; when the new Congress is seated the second condition is fulfilled.

Id. at 162. Thus, the court stated that, under the Twenty-Seventh Amendment, a law varying compensation, “although duly enacted pursuant to Article I, does not ‘take effect’ at the earliest until the new Congress has been seated. . . . Accordingly, the present Congress could specify the salary of the next Congress or of any Congress after that.” Id. The court emphasized that the ERA’s COLA provisions would not cause a COLA to “be made until more than a year later, on January 1, 1991.” Id. Therefore, the court held that the COLA provisions of the ERA are constitutional because they became law in 1989 (before the 1990 election) but “did not cause any adjustment to congressional compensation until after the election of 1990 and the seating of the new Congress.” Id.

Mr. Boehner also pressed a new, alternative argument that if the ERA’s COLA provisions are constitutional, then blocking a COLA from taking effect before the intervention of an election violates the Amendment. Id. Although it would have been of great aid to us now, the D.C. Circuit declined to address this argument for the first time on appeal. Id. at 162–63.

2. Beer

In 2012, the Federal Circuit decided Beer. As the government points out, the Beer decision is not directly analogous to our case here because it dealt with the ERA’s COLA provisions as they applied to members of the federal judiciary with their unique protections under Article III’s “Compensation Clause.” 8 Nonetheless, as we explain later, we think the

8 The Compensation Clause provides that federal judges “shall, at stated

9 Beer decision is helpful in our resolution of the third issue in this opinion— whether COLA-blocking laws “vary” Congressional compensation. But to fully understand Beer, it is necessary to briefly explain the circumstances that led to the Beer decision, beginning with the United States Supreme Court’s decision in United States v. Will, 449 U.S. 200 (1980).

In Will, although addressing a COLA framework within an entirely different statutory scheme (the Executive Salary Cost-of-Living Adjustment Act (“1975 Adjustment Act”)) than the ERA, the Supreme Court framed the issue presented to it in broad terms:

[W]hen, if ever, does the Compensation Clause prohibit the Congress from repealing salary increases that otherwise take effect automatically pursuant to a formula previously enacted? We must decide when a salary increase authorized by Congress under such a formula “vests”—i.e., becomes irreversible under the Compensation Clause. Is the protection of the Clause first invoked when the formula is enacted or when increases take effect?

Will, 449 U.S. at 221. In answering that question, the Court pronounced a seemingly equally broadly applicable rule: “[A] salary increase ‘vests’ for purposes of the Compensation Clause only when it takes effect as part of the compensation due and payable to Article III judges.” Id. at 229. In other words, the Court held that the Compensation Clause allowed Congress to block a judicial COLA at any point until “it first was scheduled to become part of judges’ compensation.” Id.

Fast-forwarding to 2001, the Federal Circuit in Williams v. United States, 240 F.3d 1019 (Fed. Cir. 2001), overruled in part by Beer v. United States, 696 F.3d 1174 (Fed. Cir. 2012) (en banc), faced the same question that the Supreme Court did in Will but dealt with the COLA provisions of the ERA instead of the 1975 Adjustment Act. 240 F.3d at 1023, 1028 (“The question for us is no different than the question which the Supreme Court posed for itself to initiate the deliberative process in Will that led to the constitutional rule of vesting that the Court adopted.”). Williams dealt with a challenge by federal judges under the Compensation Clause to Congress’s

Times, receive for their Services, a Compensation, which shall not be diminished during their Continuance in Office.” U.S. Const. art. III, § 1 (emphasis added).

10 decision to block the ERA’s scheduled COLAs for 1995, 1996, 1997, and 1999. Id. at 1024. The Federal Circuit understood Will’s “due and payable” holding as establishing “a rule that ‘vesting,’ for Article III compensation purposes, in effect requires the actual possession of the additional compensation.” Id. at 1032. In other words, “under the vesting rule in Will, pay increases for Article III judges are left constitutionally unprotected until the judges actually begin to accrue compensation under the increased rates.” Id. Importantly, each COLA-blocking law for the years at issue “became law before the January 1 effective date of the COLA pay increases.” Id. at 1024. Therefore, the court stated, “[u]nder Will, put simply, that is the end of our inquiry, and the Judges’ cause must fail.” Id. at 1031. Because it viewed Will as promulgating a broad, generally applicable constitutional rule, the Federal Circuit rejected the judges’ argument that the ERA’s COLA scheme was sufficiently distinguishable from the COLA scheme of the 1975 Adjustment Act so as to require a different result. See id. at 1037–40. Therefore, the court held that the COLA-blocking laws “were permissible . . . constitutional exercises of congressional power.” Id. at 1040.

The Williams decision elicited multiple dissenting opinions throughout its litigation lifecycle. See, e.g., Williams, 240 F.3d at 1040–64 (Plager, J., dissenting); Williams v. United States, 264 F.3d 1089, 1090–93 (Fed. Cir. 2001) (Mayer, C.J., joined by Newman and Rader, JJ., dissenting from denial of reh’g en banc); Williams v. United States, 535 U.S. 911, 911– 22 (2002) (Breyer, J., joined by Scalia and Kennedy, JJ., dissenting from denial of certiorari). While we later cite the dissenting opinions written by Judge Plager and Chief Judge Mayer for collateral propositions related to the purpose of the ERA and its COLA provisions, Justice Breyer’s dissenting opinion is the most relevant to our discussion of Beer. In brief, Justice Breyer argued that the Supreme Court should have granted certiorari to address the question of whether the ERA was sufficiently distinguishable from the 1975 Adjustment Act such that the ERA triggers Article III judges’ expectation interests under the Compensation Clause. See Williams, 535 U.S. at 917–18 (Breyer, J., joined by Scalia and Kennedy, JJ., dissenting from denial of certiorari) (explaining that the 1975 Adjustment Act’s COLAs were “neither definite nor precise” and framing the relevant question as “whether the [ERA’s COLA provisions are] sufficiently precise and definite to have created an ‘expectation’ that the Compensation Clause protects” (citation omitted)).

Turning to the Federal Circuit’s 2012 decision in Beer, the court, sitting en banc, overruled its prior Williams decision and held that the ERA

11 “triggered the Compensation Clause’s basic expectations and protections.” Beer, 696 F.3d at 1177. The court went to great lengths to explain the differences between the 1975 Adjustment Act and the ERA, noting that the 1975 Act’s COLA provisions were “imprecise as to amount and uncertain as to effect,” id. at 1181 (quoting Williams, 535 U.S. at 917 (Breyer, J., joined by Scalia and Kennedy, JJ., dissenting from denial of certiorari)), whereas the ERA calls for COLAs “according to a mechanical, automatic process that creates expectation and reliance when read in light of the Compensation Clause,” id. The court also emphasized that the “automaticity” of the ERA’s COLA scheme “takes on heightened significance” when considering the fact that the ERA “also banned judges from earning outside income and honoraria.” Id. at 1181–82 (citation omitted). Therefore, due to the differences between the 1975 Adjustment Act and the ERA, the court decided that the rule pronounced in Will did “not foreclose the relief that the judges” sought. Id. at 1183. From there, while emphasizing federal judges’ expectancy interests under Article III, the court held that that

when Congress promised protection against diminishment in real pay in a definite manner and prohibited judges from earning outside income and honoraria to supplement their compensation, [the ERA] triggered the expectation-related protections of the Compensation Clause for all sitting judges. A later Congress could not renege on that commitment without diminishing judicial compensation. . . .

Congress committed to providing sitting and prospective judges with annual COLAs in exchange for limiting their ability to seek outside income and to offset the effects of inflation. . . . Instead of fixing compensation relative to a commodity subject to inflationary pressure, Congress pegged the adjustment to a known measure of change to the economy as a whole, thus protecting the real salary of judges from both inflation and from fickle political will. By enacting blocking legislation in 1995, 1996, 1997, and 1999, Congress broke this commitment and effected a diminution in judicial compensation.

Id. 1184–85. While the court’s holding ultimately turned on expectation- related interests specific to the context of Article III judges, critical for our purposes is the court’s explanation of the COLA provisions of the ERA: The ERA’s “COLA provisions were not an increase in judicial pay. . . . Rather,

12 the statute ensured that real judicial salary would not be reduced in the face of the elimination of outside income and the operation of inflation.” Id. at 1182–83 (emphasis added) (citing Williams, 535 U.S. at 916 (Breyer, J., joined by Scalia and Kennedy, JJ., dissenting from denial of certiorari)).

II. Procedural History

On March 7, 2024, four current and former members of Congress filed this suit 9 alleging that the timing of many COLA-blocking laws violated the Twenty-Seventh Amendment. Plaintiffs included three claims in their complaint. Count I asserted a right to “reimbursement for all unconstitutionally unpaid salary” during each respective plaintiff’s tenure in Congress. Second Am. Compl. ¶¶ 54–57. 10 Count II asserted a right to recovery based on a failure to pay the correct retirement benefits, and Count III asserted a failure to credit plaintiffs with the correct Thrift Savings Plan (“TSP”) contributions under the Federal Employees’ Retirement System (“FERS”). Id. at ¶¶ 58–63.

On June 28, 2024, the government moved to dismiss the complaint under Rule 12(b)(1) of the Rules of the United States Court of Federal Claims (“RCFC”). The government put forth three arguments in support of its motion. First, the government argued that this court lacked jurisdiction over plaintiffs’ unpaid salary claims under Count I because the Twenty-Seventh Amendment is not money-mandating as required by the Tucker Act. Second, the government argued that this court lacked jurisdiction over plaintiffs’ Count II and Count III claims (regarding retirement benefits and TSP employer matching) pursuant to 5 U.S.C. § 8461. Third, the government argued that this court’s six-year statute of limitations barred plaintiffs’ pay and retirement claims insofar as such claims accrued before March 7, 2018.

On September 27, 2024, we issued an opinion denying the government’s motion to dismiss Count I, granting the motion to dismiss Counts II and III, 11 and granting the motion to dismiss all claims that were

9 On June 20, 2024, plaintiffs amended their complaint to add one more plaintiff but did not add any new substantive arguments. 10 For simplicity’s sake, and because none of the complaints varied in substance, we cite to plaintiffs’ second amended complaint, which is the most recent. 11 Counts II and III have been dismissed and are entirely irrelevant to the

13 stale pursuant to the statute of limitations. As to Count I, we rejected the government’s argument that we lacked jurisdiction on the basis that the Twenty-Seventh Amendment is not money-mandating as required by the Tucker Act. Davis, 174 Fed. Cl. at 573. In finding that we have jurisdiction, we did not need to “decide whether the Twenty-Seventh Amendment— standing alone—is money-mandating” because the relevant money- mandating statute under the Tucker Act is the ERA. Id. at 571. We rejected the government’s argument that any damage award under the ERA would be “predicated on first obtaining a declaratory judgment that the COLA- nullifying laws violate the Twenty-Seventh Amendment,” id., because “[b]ut for the allegedly unconstitutional effect of the nullifying legislation, the money-mandating source of jurisdiction remains the ERA. We may enforce the ERA COLAs, in other words, and award whatever payment they grant while ‘omitting’ any ‘constitutionally void’ provisions,” id. at 573 (quoting Gentry v. United States, 546 F.2d 343, 346 (Ct. Cl. 1976)). Thus, we noted that “if a plaintiff alleges that a money-mandating statute has been unconstitutionally thwarted, the claim is based on the statute, not the Constitution.” Id. (citing Matsuo v. United States 416 F. Supp. 2d 982, 991. (D. Ha. 2006)).

We also addressed, in dicta, some arguments that are now properly before us on these cross-motions for summary judgment. For example, we opined, albeit indirectly, on the government’s current argument that the Twenty-Seventh Amendment is concerned only with pay increases and not decreases when we said that “[t]he Founders were deeply concerned about how congressional compensation would be fixed, both to prevent self- enrichment and also performative self-impoverishment.” Id. at 574 (emphasis added) (citation omitted). Furthermore, we dismissed the government’s current argument that COLA-blocking legislation does not “vary” congressional compensation when we stated, in passing, that “blocking legislation plainly consists of laws ‘varying the compensation for the services of the Senators and Representatives.’” Id. at 573 (quoting U.S. Const. amend. XXVII). Lastly, we expressed skepticism as to the government’s characterization of the Twenty-Seventh Amendment as merely a timing provision. Id. However, we did not decide (and indeed expressly declined to decide) exactly how the Twenty-Seventh Amendment operates, id. at 574,

pending motions. Therefore, the reasons for granting the government’s motion to dismiss Counts II and III are not discussed herein.

14 574 n.5, a question which is now before us.

Regarding the government’s third argument, we agreed that any claims that had accrued before March 7, 2018, were barred by the statute of limitations. Id. at 579. 12 More importantly, however, we also stated that the continuing claims doctrine did not apply to this case. Id. Thus, we held that “COLAs nullified more than six years earlier are not resurrected for the 2018–2024 period.” 13 Id.

On January 17, 2025, plaintiffs amended their complaint for the second time, adding two new plaintiffs but leaving their substantive arguments unchanged. 14

On February 4, 2025, we granted plaintiffs’ motion to alter judgment and thus reissued our prior opinion with a finding that some of the issues decided therein warranted an immediate interlocutory appeal to the Federal Circuit pursuant to 28 U.S.C. § 1292(d)(2). Davis, 174 Fed. Cl. at 579. Relevant to the pending motions, we flagged two issues worthy of interlocutory appeal: (1) whether the Court of Federal Claims has jurisdiction over plaintiffs’ Count I salary claims, and (2) “[w]hether the ‘continuing claims doctrine’ permits plaintiffs’ claims for salary and backpay within the limitations period to be calculated according to implementation of [COLAs] mandated by the [ERA] from prior to the beginning of the six-year statute of limitations period for plaintiffs’ claims.” Id. The Federal Circuit declined to hear the appeal. Davis v. United States, No. 2025-119, 2025 U.S. App. LEXIS 12153, at *2 (Fed. Cir. May 20, 2025).

Subsequently, on September 4, 2025, plaintiffs filed their pending motion for partial summary judgment. On October 31, 2025, the government

12 This decision, along with our dismissal of Counts II and III, required the dismissal of all of the claims of two of the five plaintiffs, bringing the then- total down to three. 13 In their pending motion for partial summary judgment, plaintiffs ask the court to revisit our prior decision on this point. Because this opinion is limited in scope, we do not address the applicability of the continuing claims doctrine herein. However, as we noted earlier, see supra n.1, we view this question as unresolved and expect to address it in the future pending further briefing. 14 This brought the total number of plaintiffs back up to its current number of five.

15 filed its cross-motion for summary judgment. Briefing wrapped up on February 6, 2026, and the court heard oral argument on February 20, 2026. We limited argument to the four issues previously stated. Thus, those four issues are ready for adjudication.

STANDARD OF REVIEW

Summary judgment is appropriate when the movant “shows that there is no genuine dispute” of “material fact” and that the “movant is entitled to judgment as a matter of law.” RCFC 56(a); Almanza v. United States, 127 Fed. Cl. 521, 525 (2016), aff’d, 935 F.3d 1332, 1336–37 (Fed. Cir. 2019). A fact is material if it might affect the outcome of the suit. Silver State Land, LLC v. United States, 155 Fed. Cl. 209, 212 (2021). A dispute of material fact is genuine when the fact finder may reasonably resolve it in favor of either party. Id. We read all factual inferences in favor of the nonmoving party. Id.

DISCUSSION

As we noted earlier, the pending motions raise a complicated web of issues, many of which are of first impression. Resolution of any one question could easily give rise to a number of downstream issues or just as easily render those downstream issues entirely irrelevant, depending on how we decide the initial question. Thus, we take it from the top. As we noted earlier, we limited oral argument (and thus now limit this opinion) to the following four issues: (1) whether plaintiffs are estopped from recovering their asserted damages; (2) whether the Twenty-Seventh Amendment applies to laws decreasing Congressional compensation; (3) whether COLA-blocking legislation “varies” Congressional compensation; and (4) whether the Twenty-Seventh Amendment renders noncompliant legislation void ab initio or operates merely as a timing provision, in which case the question becomes whether and to what extent COLA-blocking legislation can be retroactively applied. We address each in turn.

I. Plaintiffs Are Not Estopped from Recovering Their Asserted Damages

The government argues that “whatever harm plaintiffs may be said to have suffered, they themselves are the cause of that injury. And having caused their own injury, they cannot now recover for it.” Def.’s Resp. & Mot. Summ. J. 22, ECF No. 51. In making this claim, the government cites no

16 authority. On the contrary, the government appeals to general principles of law, including contributory negligence, failure to mitigate, and broader concepts of waiver and estoppel, to support its assertion that allowing plaintiffs’ recovery here would run contrary to “basic principles of fairness.” Id. at 23.

Plaintiffs respond with a number of counterarguments as to why they are not estopped from and have not waived their right to recover their asserted damages, including: (1) that the Speech or Debate Clause bars “the evidentiary use of legislative votes (or other legislative acts) to a Member’s detriment;” (2) that applying principles of waiver and estoppel would perversely require members of Congress to always vote in their own self- interest; (3) that none of the COLA-blocking laws were stand-alone bills, and “[m]ultiple-subject bills force members of Congress to vote on legislative packages often containing components with which they disagree, or may even believe are unlawful;” (4) that plaintiffs cannot waive the Twenty- Seventh Amendment’s protections because they benefit the nation at large; (5) that plaintiffs also voted no on certain COLA-blocking bills; and (6) that none of the traditional estoppel requirements are satisfied here. Pls.’ Reply & Resp. 9–17, ECF No. 54.

There are really two ways to view the government’s arguments on this point: (1) an individualized waiver or estoppel argument specific to each plaintiff, and (2) a more corporate argument that because Congress caused its own problems, and due to structural concerns, individual Members of Congress can only seek redress through Congress itself. At oral argument, the government endorsed the second option and seemingly abandoned the first. See Oral Arg. Tr. 28–29, Feb. 20, 2026. Because both arguments are arguably encompassed by the briefing and subsequent oral argument, however, we address them both, beginning with the argument that each plaintiff is individually estopped from recovering their asserted damages.

Equitable estoppel “is a judicial remedy by which a party may be precluded, by its own acts or omissions, from asserting a right to which it otherwise would have been entitled.” LaMirage, Inc. v. United States, 44 Fed. Cl. 192, 200 (1999) (citing Hecklerv. Cmty. Health Servs. of Crawford County, Inc., 467 U.S. 51, 59 (1984)). The Federal Circuit has articulated the general elements of equitable estoppel, albeit in a patent case, as follows:

[1] The actor, who usually must have knowledge of the true facts, communicates something in a misleading way, either by

17 words, conduct or silence. [2] The other relies upon that communication. [3] And the other would be harmed materially if the actor is later permitted to assert any claim inconsistent with his earlier conduct.

A.C. Aukerman Co. v. RL Chaides Constr. Co., 960 F.2d 1020, 1041 (Fed. Cir. 1992) (citation omitted), abrogated on other grounds at SCA Hygiene Prods. Aktiebolag v. First Quality Baby Prods., LLC, 580 U.S. 328 (2017). Critically, whether equitable estoppel applies to a given case is left to our discretion. Id. In the exercise of that discretion, we decline to apply the doctrine here.

When members of Congress vote, they are, in principle at least, representing the interests of their constituents, not their own. And as plaintiffs point out, COLA-blocking legislation has historically not been voted on in isolation; rather, each COLA-blocking law has been but one aspect of multi-subject, often omnibus, bills. Therefore, when a Member of Congress opts to vote in favor of one of those bills in toto, it is appropriate to assume that he or she viewed voting in favor of the entirety of the bill, including the inherent political compromises contained therein, as being in the best interests of his or her constituents. The government’s approach, at least insofar as it calls for evaluating each plaintiff individually, calls for us to exert judicial pressure on Members of Congress to vote in their own self- interest rather than in a representative capacity. The equities here do not require that approach, and it would be inappropriate regardless. We thus refuse to go down that road.

There are also other reasons to deny the government’s equitable estoppel argument. First, as we noted, equitable estoppel requires the party asserting the defense to have relied on the other party’s acts. Here, however, that is not the case. The government did not rely on any individual vote by any individual plaintiff. Rather, the government relied on legislation duly passed in accordance with the requirements of bicameralism and presentment, regardless of how these plaintiffs voted on any individual piece of legislation. Thus, it cannot be said that the government relied on any act by these plaintiffs but only, if at all, on the institutional acts of Congress itself.

Second, equitable estoppel requires the party asserting the defense to have been ignorant of the true facts while the party to be estopped must have had knowledge of those facts. That is plainly not the case here, however,

18 because, on the one hand, the existence of the Twenty-Seventh Amendment was obviously known to all parties involved, and, on the other, the “truth” of whether, to what extent, and in what manner the Amendment applies to COLA-blocking legislation was inherently unknown to all parties because none of those issues have yet been resolved. Thus, the common knowledge of the existence of the Twenty-Seventh Amendment and the uncertainty surrounding the Amendment’s substantive application forecloses the application of principles of estoppel against plaintiffs in this case.

Third, each plaintiff before us has voted both for and against multi- subject bills containing COLA-blocking legislation, which shows that individual members’ votes cannot serve as the basis for applying equitable estoppel. The government effectively concedes this point when it expands its argument to include plaintiffs’ failure to show “that they tried to stop specifically the challenged provisions through the legislative process.” Def.’s Reply 8 n.4, ECF No. 60; see also Def.’s Resp. & Mot. Summ. J. 22, ECF No. 51 (“[P]laintiffs have identified no attempts by any of the named plaintiffs to stop those provisions through the legislative process.”). 15

We also note that the potential impact of the Speech or Debate Clause, which plaintiffs explicitly invoke, gives us great pause. 16 However, there are

15 The government’s point here seems to show that its real argument is a structural one whereby all members of Congress, no matter the context, are equitably barred from seeking a judicial remedy because Congress, as an institution, caused the problems they complain of. 16 Although we do not reach the issue, we note that the Speech or Debate Clause states that “for any Speech or Debate in either House, [Senators and Representatives] shall not be questioned in any other Place.” U.S. Const. art. I, § 6, cl. 1. To hold that plaintiffs here, as individuals, either waived their right or are equitably estopped from asserting their right to their alleged damages, we would undoubtedly have to question them about behavior that falls within the purview of the Speech or Debate Clause, namely “legislative acts or the motivation for actual performance of legislative acts.” United States v. Brewster, 408 U.S. 501, 509 (1972). In addition to preventing liability, whether criminal or civil, from being founded on legislative acts, the Speech or Debate Clause also serves as an evidentiary bar by “protect[ing] Members from inquiry into legislative acts or the motivation for actual performance of legislative acts.” Fields v. Office of Eddie Bernice Johnson, 459 F.3d 1, 14 (D.C. Cir. 2006) (en banc) (citation omitted)

19 sufficient reasons apart from the Speech or Debate Clause to justify our exercise of discretion to reject the government’s equitable argument insofar as it targets plaintiffs individually. Indeed, “[i]t has long been the policy of the courts to decide cases on non-constitutional grounds when that is available, rather than reach out for the constitutional issue.” Stockton E. Water Dist. v. United States, 583 F.3d 1344, 1368 (Fed. Cir. 2009) (citing Nw. Austin Mun. Util. Dist. No. One v. Holder, 557 U.S. 193, 206 (2009)). Consistent with that principle, it is unnecessary for us to rely on the Speech or Debate Clause, and we therefore decline to do so here.

(quoting Brewster, 408 U.S. at 508); see also United States v. Helstoski, 442 U.S. 477, 489 (1979) (noting that the Speech or Debate Clause guards “against inquiry into acts that occur in the regular course of the legislative process and into the motivation for those acts” and “precludes any showing of how [a legislator] acted, voted, or decided” (alteration in original) (citation omitted)). The government’s briefing, however, explicitly calls on us to evaluate plaintiffs’ votes, their motives for those votes, and their acts (or inaction) throughout the legislative process in determining whether principles of equity bar their asserted recovery here.

The Speech or Debate Clause’s evidentiary bar applies in civil cases. See, e.g., Fields, 459 F.3d at 4, 14 (applying the evidentiary bar to a civil suit under the Congressional Accountability Act of 1995, 2 U.S.C. §§ 1301– 1438). However, the government points out that plaintiffs fail to cite any case in which the Speech or Debate Clause “was wielded by plaintiff-Members as a sword rather than by defendant-Members as a shield.” Def.’s Reply 7, ECF No. 60. We too cannot find a case supporting (or rejecting) that proposition. See Nunes v. NBCUniversal Media, LLC, No. 22-CV-01633-PKC-SN, 2024 U.S. Dist. LEXIS 48491, at *4 (S.D.N.Y. Mar. 19, 2024) (“The Court has been unable to locate a case where a Member initiated a civil lawsuit against a private party and then asserted the Speech or Debate Clause to withhold discovery.”) But we think there is at least a legitimate question as to whether the Clause applies in a case such as this where the government seeks to use plaintiffs’ quintessentially legislative acts to preclude them “from asserting a right to which [they] otherwise would have [allegedly] been entitled.” LaMirage, Inc., 44 Fed. Cl. at 200 (1999) (citation omitted).

20 We thus turn to the government’s broader argument in equity, which is really a structural one. See Oral Arg. Tr. 29, Feb. 20, 2026 (stating that the government’s argument exists at the “intersection of . . . classic separation of powers concerns” and “common law principles of equity like estoppel”). The crux of the government’s argument, as we understand it, is that the Ascertainment Clause delegates to Congress the authority and responsibility to set its own pay, and the Twenty-Seventh Amendment merely “factors into that process.” Oral Arg. Tr. 30, Feb. 20, 2026. In the exercise of its pay- setting authority, whether done in accordance with the Twenty-Seventh Amendment or not, Congress has consistently decided to reject the ERA’s automatic COLAs. Therefore, as the government sees it, because “both the laws being challenged and the constitutional provision at issue are confined to Congress managing itself,” Members of Congress “have both the privilege . . . and the responsibility to seek redress for whatever harms they believe they have caused themselves through the same legislative process that caused them,” Def.’s Reply 8–9, ECF No. 60, i.e., not through the courts.

We reject this argument. Members of Congress can certainly try to use the legislative process to increase congressional pay. However, it does not follow that Members of Congress cannot seek redress for the alleged injury of not receiving the compensation to which they are entitled. What plaintiffs ask the court to do is award them damages under the ERA (to the extent COLA-blocking legislation violated the Twenty-Seventh Amendment during the relevant years) precisely because whatever the ERA dictates their pay should have been is the compensation that Congress set and to which plaintiffs are entitled. To the extent that plaintiffs can show that the law required them to be paid a sum certain and that they were not paid that amount, we see it as not only appropriate but also our duty to award those damages. See Gentry, 546 F.2d at 346 (explaining that the court must “read the relevant statute and award whatever payment it grants, omitting from our interpretation of its effect any constitutionally void provision contained in it”).

The government’s approach would have us give force to what are allegedly unconstitutional enactments. Even if every relevant COLA- blocking law violated the Twenty-Seventh Amendment, and even if plaintiffs are correct that laws violative of the Twenty-Seventh Amendment are void ab initio, the government would nonetheless give them effect so as not to step into the realm of “Congress managing itself.” Def.’s Reply 8, ECF No. 60. In other words, we would render the Twenty-Seventh Amendment no more than a paper tiger.

21 The government also argues that if we allow plaintiffs to recover their asserted damages, we will set a precedent that would allow future Members of Congress to repeatedly posture to their constituents that they rejected (or at least tried to reject) automatic COLAs while simultaneously receiving those COLAs through the judicial process. See Oral Arg. Tr. 34–35, Feb. 20, 2026. This is clearly a political concern that the voters (and thus future Congresses) can sort out as the Twenty-Seventh Amendment contemplates.

In sum, whether viewed corporately or individually, we hold that principles of equity do not bar plaintiffs from recovering damages in this case.

II. The Twenty-Seventh Amendment Properly Applies to Laws Decreasing Congressional Compensation

We turn to the second question: whether the Twenty-Seventh Amendment applies to laws decreasing Congressional compensation. In its briefing, the government argues that judgment in favor of these plaintiffs would undermine the political accountability purposes of the Twenty- Seventh Amendment. Def.’s Resp. & Mot. Summ. J. 6, 9–10, ECF No. 51. The government devotes much of its briefing on this point to arguing that the Twenty-Seventh Amendment was meant to target pay increases, not decreases. Id. at 6–9. At oral argument, however, the government disavowed any reliance on this argument. Oral Arg. Tr. 41, Feb. 20, 2026 (“I don’t think we actually make that argument, and I don’t think the Court should address that question.”). 17 Instead, while characterizing the issue as an “open question,” the government urged the court not to address it because, under the government’s theory of the case, there are sufficient reasons to find in the government’s favor such that we need not reach the question of whether the

17 In its briefing, the government ultimately grounds its argument in the similar political posturing concern we addressed earlier. Def.’s Resp. & Mot. Summ. J. 6, 9–10, ECF No. 51. The focal point of the argument is whether the purpose of the Twenty-Seventh Amendment as the government sees it (political accountability) is served if the Amendment applies to pay decreases. See id. Because (in the government’s view) the Amendment’s purpose is thwarted by applying it to pay decreases, the government argues we should not grant plaintiffs their asserted damages. Ultimately, we see this as no more than an argument as to why the Twenty-Seventh Amendment should not apply to pay decreases.

22 Twenty-Seventh Amendment speaks to pay decreases. Id.

Plaintiffs argue that the Twenty-Seventh Amendment speaks to both pay increases and decreases for two reasons. First, the Twenty-Seventh Amendment uses the term “varying,” which, by its plain meaning, encompasses changes in compensation both upward and downward. Pls.’ Reply & Resp. 4–5, ECF No. 54. In plaintiffs’ view, to hold otherwise would require rewriting the Twenty-Seventh Amendment to replace “varying” with “increasing.” Id. Second, plaintiffs argue that “experience with the British Parliament and the Revolutionary Era informed the Founders’ desire to prevent Congress from lowering its own pay for immediate political gain at the expense of institutional judgment, the public interest, and permitting individuals of modest means to serve.” Id. at 5.

As a preliminary matter, we cannot avoid addressing the question of whether the Twenty-Seventh Amendment encompasses legislation that decreases Congressional pay because we do not resolve any of the other three comprehensive issues in a manner sufficient to dispose of the entirety of plaintiffs’ claims. Although we appreciate the principles of judicial minimalism and constitutional avoidance for which the government advocates, they are not applicable here.

We thus now address the merits of the question. In so doing, our analysis “begins by examining the constitutional language at issue.” Botello v. United States, 173 Fed. Cl. 26, 34 (2024) (citing United States v. Rahimi, 602 U.S. 680, 715 (2024) (J. Kavanaugh, concurring) (“The first and most important rule in constitutional interpretation is to heed the text—that is, the actual words of the Constitution[.]”)). The Twenty-Seventh Amendment states that “[n]o law, varying the compensation for the services of the Senators and Representatives, shall take effect, until an election of Representatives shall have intervened.” U.S. Const. amend. XXVII. The question we address here, therefore, is whether a law that decreases congressional compensation is a law “varying the compensation of the Senators and Representatives.” Plainly it is.

“The framers of the Constitution employed words in their natural sense; and where they are plain and clear, resort to collateral aids to interpretation is unnecessary and cannot be indulged in to narrow or enlarge the text . . . .” McPherson v. Blacker, 146 U. S. 1, 27 (1892). Only “where there is ambiguity or doubt,” id., may we look to other means of interpreting a constitutional provision, such as its history. Rahimi, 602 U.S. at 718, 718

23 n.2 (2024) (J. Kavanaugh, concurring) (“The historical approach applies when the text is vague. But the text of the Constitution always controls. So history contrary to clear text is not to be followed.”).

Resorting to collateral aids, including the Twenty-Seventh Amendment’s relevant history, is neither necessary nor proper in interpreting the meaning of “varying” because the meaning of that term is unambiguous. The Twenty-Seventh Amendment was originally proposed on June 8, 1789. Richard B. Bernstein, The Sleeper Wakes: The History and Legacy of the Twenty-Seventh Amendment, 61 Fordham L. Rev. 497, 498 (1992). In the Twenty-Seventh Amendment, the word “varying” is used in a participial phrase that modifies the noun, “law.” In contemporaneous founding-era dictionaries, “varying,” when functioning as a verb or an adjective (as it is in the Twenty-Seventh Amendment), is defined as simply meaning “[c]hanging” or “altering.” E.g., 2 J. Ash, The New and Complete Dictionary of the English Language (1775); see also 2 S. Johnson, A Dictionary of the English Language (4th ed. 1773) (defining the root word, to vary, as “[t]o change; to make unlike itself” and “[t]o change to something else”).

The precise question we address here is whether a law that decreases Congressional compensation falls within the Twenty-Seventh Amendment’s purview. The answer to that question is clearly yes. If Congressional compensation is reduced, there is no doubt that it has been changed and altered. Therefore, a law decreasing Congressional compensation obviously “varies” that compensation. Because the text is clear, we have no need to look to history to discern the meaning of the Twenty-Seventh Amendment on this point. 18 We thus have no difficulty in holding that “varying,” as that term is used in the Twenty-Seventh Amendment, plainly encompasses laws that increase and those that decrease Congressional compensation.

III. Cola-Blocking Laws Are Laws “Varying” Congressional Compensation Under the Twenty-Seventh Amendment

Now to a more difficult question: whether COLA-blocking legislation amounts to laws “varying” congressional compensation for purposes of the Twenty-Seventh Amendment. We note at the outset that this is a question of first impression. Plaintiffs’ argument that COLA-blocking legislation

18 Indeed, due to the clear meaning of “varying,” the only reason to look to history would be to improperly “narrow . . . the text.” McPherson, 146 U. S. at 27.

24 “varies” congressional compensation can be distilled as follows: The Ascertainment Clause, found in Article I, Section 6, Clause 1 of the United States Constitution, requires that congressional compensation be set “by Law.” Relying on the interplay between Boehner, Beer, and Justice Breyer’s opinion dissenting from the denial of certiorari in Williams v. United States, 535 U.S. 911 (2002), plaintiffs argue that the ERA is the law setting congressional compensation and has been since it became effective. This includes the ERA’s automatic COLA provisions. Thus, because the ERA’s pay formula is the congressional compensation set by law, any subsequent law deviating from that formula, whether up or down, is a law “varying” congressional compensation that must adhere to the requirements of the Twenty-Seventh Amendment.

The government sees things differently. It argues that the Federal Circuit’s decision in Beer is inapplicable here because that case applied only to Article III federal judges. The government further argues that the D.C. Circuit’s decision in Boehner does not speak to the question at issue here at all. In the government’s view, COLA-blocking legislation does not trigger the Twenty-Seventh Amendment because, by blocking an automatic COLA from going into effect, such legislation prevents a variation from taking place. In other words, the government focuses on the nominal quantum of congressional pay—currently $174,000—in that, because a COLA-blocking law maintains congressional compensation at $174,000, such a law prevents congressional compensation from “varying.”

In our prior decision, we said in passing that “[b]locking legislation plainly consists of laws ‘varying the compensation for the services of the Senators and Representatives.’” Davis, 174 Fed. Cl. at 573 (quoting U.S. Const. amend. XXVII). As the government points out, this was stated in dicta, and perhaps we were premature in characterizing that conclusion as being apparent. Regardless, we reach the same conclusion today and hold that when Congress enacts a piece of COLA-blocking legislation, it enacts a law “varying” congressional compensation for purposes of the Twenty- Seventh Amendment.

We begin with the Constitution. As stated previously, the Ascertainment Clause states that “[t]he Senators and Representatives shall receive a Compensation for their Services, to be ascertained by Law, and paid out of the Treasury of the United States.” U.S. Const. art. I, § 6, cl. 1. The Twenty-Seventh Amendment, which “modifies the Ascertainment Clause,” Massie v. Pelosi, 590 F. Supp. 3d 196, 220 (D.D.C. 2022), states that “[n]o

25 law, varying the compensation for the services of the Senators and Representatives, shall take effect, until an election of Representatives shall have intervened.” U.S. Const. amend. XXVII.

The ERA is the law that has “ascertained” congressional compensation since it became effective. The ERA “is a comprehensive codification of ethical rules (§§ 301–03), financial reporting requirements (§ 202), . . . prohibitions on outside income and honoraria (§ 601), a twenty- five percent pay raise (§ 703), and future [COLAs] (§ 704).” Williams, 264 F.3d at 1090 (Mayer, C.J., joined by Newman and Rader, JJ., dissenting from denial of reh’g en banc) (citation omitted). In exchange for restricting the ability to earn outside income, the ERA “created essentially a structured payment plan with a twenty-five percent immediate pay increase coupled with automatic COLAs based on the ECI to maintain the value of the salaries during inflation.” Id. at 1091. In other words, as part of the ERA and the tradeoffs contained therein, Congress granted itself a twenty-five percent raise and included automatic COLA provisions to ensure that its real salary remained at that new level. See Beer, 696 F.3d at 1177 (explaining that the ERA “provided for self-executing and non-discretionary [COLAs] to protect and maintain a judge’s real salary”). 19

As noted previously, the D.C. Circuit in Boehner addressed the question of whether the COLA provisions of the ERA complied with the Twenty-Seventh Amendment. Boehner, 30 F.3d at 158. The appellant in Boehner argued that each COLA was a separate law that took effect the year it was payable, thereby requiring each individual COLA to comply with the Ascertainment Clause and the Twenty-Seventh Amendment. See id. at 161. The Boehner court rejected this argument and instead held that the phrase “shall take effect” in the Twenty-Seventh Amendment referred to the operative date of the ERA’s COLA provisions (January 1, 1991), not the date any particular COLA is actually paid to Members of Congress. Id. at 161– 62. The court held that, under the Twenty-Seventh Amendment, a law

19 See also Williams, 240 F.3d at 1042 (Plager, J., dissenting) (“Congress, when it enacted the [ERA], clearly had in mind a dual purpose—to remove from itself, from the Executive Branch, and from the Judiciary, the right to earn outside income through particular activities that were considered to have the potential for conflicts of interest, and to replace those sources of income with guaranteed [COLAs] whenever standard inflation measures so indicated.”).

26 varying congressional compensation can be enacted at any time but cannot take effect until an election intervenes and “the new Congress has been seated.” Id. at 162. “Accordingly, the present Congress could specify the salary of the next Congress or of any Congress after that. . . . We see no reason whatsoever why the Congress cannot, for convenience, instead specify an index or formula with the same effect.” Id. The court summarized that

it has been the law since 1989 that a COLA would be made on January 1, 1991 and each year thereafter pursuant to a specified formula. Therefore, . . . the COLA provision of the [ERA] is constitutional because it did not cause any adjustment to congressional compensation until after the election of 1990 and the seating of the new Congress.

Id. Unfortunately, as noted earlier, the Boehner court specifically declined to address the exact question before us today. See id. at 162–63.

On the other hand, in Beer, the Federal Circuit addressed the constitutionality of COLA-blocking laws but did so only in the context of Article III judges. The government is in large part correct that the reasoning of Beer does not directly translate to the question before us, especially considering the Beer court’s emphasis on federal judges’ expectation interests under Article III. However, one aspect of the Beer decision that we view as both controlling and dispositive here, and which is not limited to the Article III context, is the Federal Circuit’s conclusion that the ERA’s “COLA provisions were not an increase in judicial pay. . . . Rather, the statute ensured that real judicial salary would not be reduced in the face of the elimination of outside income and the operation of inflation.” Beer, 696 F.3d at 1182–83 (emphasis added) (citing Williams, 535 U.S. at 916 (Breyer, J., joined by Scalia and Kennedy, JJ., dissenting from denial of certiorari)). In other words, the ERA did not set up a system for periodic pay raises; rather, the ERA’s COLA adjustments are simply Congress’s chosen method of ensuring that real judicial pay would not decrease due to the effects of inflation. The ERA was intended to do the same thing for Members of Congress. Indeed, “the essence of the bargain struck by Congress” through the ERA, both for its own Members and for judges, was that “it would provide regular cost-of-living adjustments to maintain the value of the concomitant twenty-five percent salary increase in exchange for the limitations” imposed by the ERA. Williams, 264 F.3d at 1092 (Mayer, C.J., joined by Newman and Rader, JJ., dissenting from denial of reh’g en banc).

27 In exchange for Members’ services and the requirement to give up external income to so serve, the ERA provided an immediate twenty-five percent pay raise and the promise that, through the COLA process (subject to the occurrence of a GS rate adjustment), the level of congressional pay will be maintained against the tide of inflation. In other words, the immediate adjustment and the promise of subsequent COLAs were the pay raise. Because, according to Beer, the automatic COLA provisions of the ERA are not pay increases (and clearly are not pay decreases), they can only be one thing: the congressional compensation ascertained by law pursuant to the Ascertainment Clause. From here, the answer to our ultimate question is straightforward. COLA-blocking legislation, which prevents the promised maintenance of real congressional salary, decreases and, therefore, “varies” congressional compensation, thereby triggering the Twenty-Seventh Amendment.

To be sure, it is entirely within Congress’s authority under the Ascertainment Clause to renege on its commitment to maintain real congressional salary against the threat of inflation. And nothing in the Constitution requires Congress to insulate congressional compensation from the effects of inflation at all. Through the ERA, however, Congress statutorily committed to maintain the real value of congressional compensation (as set following the one-time twenty-five percent increase) through automatic, annual COLAs in exchange for the relinquishment of the right to outside income (among other ERA limitations). That is the bargain the ERA struck. 20 Due to the ERA creating a compensation scheme that is insulated against inflation, Congress’s later decision to block an automatic COLA thereby results in a decrease in real compensation due to the operation of inflation and the missed COLA adjustment and thus “varies” congressional compensation even though the nominal value of that compensation remains constant. As such, although it is entirely within the power of Congress to enact COLA-blocking legislation, it must do so in a manner consistent with the requirements of the Twenty-Seventh

20 “That later Congresses, perhaps with different members and different political agendas, may choose to pursue different goals does not change the nature of the compromise struck and the resulting purpose that the 1989 Congress had when it enacted the [ERA].” Williams, 240 F.3d at 1057 (Plager, J., dissenting).

28 Amendment.

This conclusion is further supported by the framework of the Twenty- Seventh Amendment itself, including the alleged “political accountability” aspect of the Amendment that the government stresses. The Twenty-Seventh Amendment “conditions the operation of a law varying congressional compensation upon an election of Representatives and the expiration of the Congress that voted for it.” Boehner, 30 F.3d at 162. As the Boehner court held, the ERA, including its COLA provisions, satisfied the Amendment’s requirements. In other words, after its enactment, the voters weighed in through the intervening election of 1990 and the next Congress was seated before the ERA took effect. The same cannot be said, however, for each of the allegedly constitutionally deficient COLA-blocking laws. Each COLA- blocking law, to the extent it did not comply with the Twenty-Seventh Amendment, prevented the congressional compensation ascertained by law (the ERA) from being paid even before the voters could weigh in through an intervening election and before the next Congress was in place. The Twenty- Seventh Amendment exists to prevent this exact situation.

In sum, we hold that the ERA, including its COLA provisions, is and has been the law ascertaining congressional compensation since it was enacted in 1989. See id. at 162. The ERA’s COLA provisions do not increase pay themselves but instead set congressional compensation in such a manner so as to prevent the degradation of real value through inflation. See Beer, 696 F.3d at 1182–83 Thus, because the congressional compensation ascertained by law pursuant to the Ascertainment Clause includes the right to automatic COLAs (subject to the occurrence of a GS rate adjustment) in order to maintain real congressional salary at the level set by the ERA, any deviation from the adjustment that the ERA’s COLA provisions would have otherwise required varies congressional compensation. Therefore, COLA-blocking laws, which, by their terms, block the operation of the ERA’s automatic COLAs, decrease and therefore “vary” congressional compensation, thereby triggering the Twenty-Seventh Amendment.

IV. Application of the Twenty-Seventh Amendment

The fourth and final issue we address in this opinion is the most difficult: whether the Twenty-Seventh Amendment renders noncompliant legislation void ab initio or operates as a timing provision, i.e., merely delaying when pay-changing laws take effect. Plaintiffs assert that any law that is subject to and violative of the Twenty-Seventh Amendment is void ab

29 initio, i.e., has no effect (and can never have any effect) even after an election intervenes. The government counters that, based on the text of the Twenty- Seventh Amendment, the Amendment can never void any law at all but instead operates as a timing provision, merely governing when a law varying congressional compensation takes effect. Resolution of this top-line question has the potential to lead to a series of cascading issues. Before we explain our decision on the top-line issue and the collateral questions that flow from it, we think it is helpful to explain the approaches the parties have raised and the various consequences of each approach.

We begin with plaintiffs’ argument that the Twenty-Seventh Amendment renders all non-compliant laws, in their entirety, per se void ab initio. This approach is the simplest to apply. If the Twenty-Seventh Amendment renders all noncompliant legislation void ab initio, all that is left for us to do is evaluate every COLA-blocking law for the relevant years 21 and determine which of them failed to comply with the Twenty-Seventh Amendment. From there, we would then simply ignore every non-complying COLA-blocking law. See Gentry, 546 F.2d at 346. Importantly, the “void ab initio” approach would allow COLAs in relevant years to “accumulate and compound,” Pls.’ Mem. Supp. Mot. Summ. J. 19, ECF No. 47-1, in calculating the appropriate damages for years 2018 on.

Below is a chart illustrating the void ab initio approach with a few simplistic assumptions. First, the chart assumes that each COLA-blocking law from 2018 on did not comply with the Twenty-Seventh Amendment. Second, the chart assumes that the continuing claims doctrine does not apply. 22 Third, rather than applying the ERA’s specific, formula-driven calculation, the chart assumes that each COLA came out to be a two percent increase. Fourth, the chart assumes that the proper way to calculate the 2018 COLA is by using the actual rate paid during 2017, although there is an argument that we should use the pay that Congress should have received in

21 We use the phrase “relevant years” because the relevant years in this case in large part depends on our resolution of the applicability of the continuing claims doctrine question, which we do not address in this opinion. 22 If the continuing claims doctrine did apply in this void ab initio scenario, then there would be a large gap from the outset in 2018 between the “pay received” and “pay due” categories reflecting the accumulation and compounding of the allegedly improperly blocked COLAs beginning in 1995.

30 2017 as the baseline in computing the 2018 COLA. See Pls.’ Reply & Resp. 27, ECF No. 54. The gap between the “pay due” category and the “pay received” category for every year beginning in 2018 illustrates the approximate quantum of damages for that year (except that plaintiffs here cannot recover the entire spread for 2018 but can only recover those damages beginning March 7, 2018). 23

Void ab initio Approach

210 Thousands

206 202 198 194 190 186 182 178 174 170 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026

Pay Received Pay Due

As the above chart demonstrates, under the void ab initio approach, the damages due to plaintiffs would start small in 2018 and grow exponentially thereafter as each COLA accumulates and compounds year over year.

The government’s ultimate argument is that the Twenty-Seventh Amendment is merely a timing provision and, as such, laws that “violate” the Amendment initially nevertheless come into full effect retroactively upon the satisfaction of the Amendment’s intervening election requirement. In other words, “although the 119th Congress cannot vary its own compensation while it is in session,” Def.’s Resp. & Mot. Summ. J. 20, ECF No. 51, the 119th Congress is free to pass a law that attempts to immediately vary its own compensation; while the Twenty-Seventh Amendment prevents such a law from immediately taking effect, the Amendment nevertheless allows the

23 For 2026, the chart depicts what the appropriate amount of damages would be under our assumptions if Congress continues to receive a salary of $174,000 through the end of 2026.

31 law to retroactively vary the compensation of the 119th Congress when the law takes effect after the next election and the seating of the 120th Congress. Because an election has intervened every two years up until the most recent election in 2024, all COLA-blocking laws enacted prior to the 2024 election have now taken effect under that approach.

This does lead to a rather peculiar issue with respect to the laws blocking the 2025 and 2026 COLAs, which plaintiffs assert did not comply with the Twenty-Seventh Amendment. The passage of time has not yet rendered effective the COLA-blocking provisions through an intervening election. Thus, if we calculated damages currently, the government’s approach would require the inclusion of the lost COLAs for 2025 and 2026, but then those COLAs would be reversed after the 2026 election renders the COLA-blocking laws effective, presumably requiring the recipients of those payments to pay them back to the government. The government responds that we should simply recognize that the 2026 election will intervene and render the laws blocking the 2025 and 2026 COLAs retroactively valid, thereby justifying the court in withholding damages. Below is a chart reflecting this approach (with the same assumptions as the previous chart) and the potential confusion with respect to how to treat years 2025 and 2026.

Timing Provision: Full Retroactvity Approach

182 Thousands

170 2017 2018 2019 2020 2021 2022 2023 2024 2025? 2026?

Things get a bit more complicated, however, under the government’s alternative timing provision argument, which we call the partial retroactivity approach. Under this approach, we would recognize that, whenever a COLA-

32 blocking law violates the Twenty-Seventh Amendment, Members of Congress are entitled to receive as damages the COLA that would otherwise have gone into effect for the first year. However, if, as the government argues, the Twenty-Seventh Amendment operates as merely a timing provision, the COLA-blocking law can still go into effect retroactively but only for the purpose of “prospectively setting the baseline to which future COLAs are applied with every new Congress.” Def.’s Resp. & Mot. Summ. J. 26, ECF No. 51. This approach is also best illustrated with a chart, which we include below. The chart uses the same assumptions as the first two.

Timing Provision: Partial Retroactivity Approach

170 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026

As we can see from the chart, the 2018 COLA is calculated based on 2017’s baseline of $174,000, leading to a “pay due” value of $177,480 ($174,000 increased by two percent). However, upon the intervention of the 2018 election, the law meant to block the 2018 COLA went into effect upon the seating of the new Congress in 2019. See Boehner, 30 F.3d at 162. Therefore, under the partial retroactivity approach, we would pretend that the 2018 COLA never existed for purposes of calculating the 2019 COLA. Thus, the 2019 COLA (which we again assume was two percent) is calculated against the 2017 baseline of $174,000, again leading to a “pay due” value of $177,480 ($174,000 increased by two percent).

It is a different story, however, for the 2020 COLA. At the time the 2020 COLA came due, no election had yet intervened to allow the retroactive application (prospectively resetting the baseline) of the law blocking the

33 2019 COLA. Thus, in calculating the proper value for the 2020 COLA, we would use 2019’s “pay due” rate of $177,480. Therefore, to calculate 2020’s COLA, we multiply the two percent rate by $177,480, not $174,000. That is why the “pay due” rate for 2020 is greater than the “pay due” rate for 2019.

When it comes time to calculate the proper COLA for 2021, the 2020 election has intervened. Therefore, we would then need to ignore both the 2019 and 2020 COLAs (and their respective “pay due” rates) and return to 2017’s baseline compensation of $174,000. Therefore, applying again the presumed two percent COLA, 2021’s “pay due” rate returns to $177,480 ($174,000 increased by two percent). Under the partial retroactivity approach, then, the resulting ping pong effect would continue, in theory at least, in perpetuity (as can be seen in the chart).

Plaintiffs respond that, even if the Twenty-Seventh Amendment conceptually allows for the partial retroactivity approach, we must still evaluate whether each particular COLA-blocking law is amenable to partial retroactive application at all (i.e., prospective application beginning after the next election intervenes). Plaintiffs put forth two arguments under this idea. First, plaintiffs assert that when a COLA-blocking law is part of a one-year appropriations bill (which is the case for all COLA-blocking laws for the relevant years), it is “no longer available for delayed application” because the one-year window of operation for the appropriations bill as a whole has passed. Pls.’ Reply & Resp. 25, ECF No. 54 (citation omitted). This first approach would essentially lead to the same result and calculation of damages as the void ab initio approach, including under a scenario where we determine that the continuing claims doctrine applies.

Second, plaintiffs argue that we must look at the language of each COLA-blocking law to determine whether it is even compatible with retroactive application (prospectively resetting the baseline). Specifically, plaintiffs assert that “the explicit terms of each COLA-blocking law from 2013 to present state that the law’s effect was to be applied only ‘during’ the identified calendar year.” Id. at 28. Therefore, “[b]y its own terms, each COLA-blocking law had effectively expired prior to the point (i.e., upon the next Congress’s seating) when it would otherwise be applied retroactively.” Id. If the continuing claims doctrine does not apply, then this approach would lead to the same result and calculation of damages as the void ab initio approach. However, if the continuing claims doctrine does apply, it is unclear whether this approach would lead to the same result as the void ab initio approach because we would need to evaluate the language of each COLA-

34 blocking law for the relevant years, and, as we previously stated, COLA- blocking legislation has not always used the same language throughout the years.

We begin our analysis of the fourth issue with the language of the Amendment. See Botello, 173 Fed. Cl. at 35 (citation omitted). Recall that the Twenty-Seventh Amendment states that “[n]o law, varying the compensation for the services of the Senators and Representatives, shall take effect, until an election of Representatives shall have intervened.” U.S. Const. amend. XXVII. Or, to put it more succinctly, the Twenty-Seventh Amendment provides that any law that varies congressional compensation cannot take effect until an election of Representatives intervenes. Thus, it seems to us plain that any law is ineffective to the extent it purports to vary congressional compensation prior to an intervening election. Indeed, this plain reading is entirely consistent with the Amendment’s primary purpose, which, “[a]ccording to Madison, and to all the ratifying states that stated their understanding, . . . is to ensure that a” law varying congressional pay “cannot be for the particular benefit of those who are concerned with determining the value of the service.” Boehner, 30 F.3d at 159 (citation omitted) (quoting James Madison, Speech in the House of Representatives (June 8, 1789), in The Congressional Register, June 8, 1789, reprinted in Creating the Bill of Rights: The Documentary Record from the First Federal Congress 84 (Helen E. Veit et al., eds., 1991)).

This does not mean, however, that all non-compliant laws are, in their entirety, per se void ab initio as plaintiffs argue. The Twenty-Seventh Amendment is concerned with both the substance of certain laws (applying only to laws varying congressional compensation) and the timing of when such laws can become operative. Connecting the two, the Amendment, by its terms, prevents a law from taking effect (and thus effectuating a change to congressional compensation) “until” an election intervenes. U.S. Const. amend. XXVII. But plaintiffs’ void ab initio approach would require us to read “until” out of the Amendment entirely. Properly read, the Amendment renders subject laws ineffective up to the next intervening election but no further. At a conceptual level, therefore, this leaves the door open for the Twenty-Seventh Amendment to render part of a law ineffective and yet part of that same law effective once an election intervenes. Thus, to the extent a law violates the Twenty-Seventh Amendment (by seeking to effectuate a change in congressional compensation before the next election intervenes), it is ineffective. But to the extent that same law does not violate the Amendment (by seeking to effectuate a change in congressional

35 compensation after the next election intervenes), it simply takes effect when the Twenty-Seventh Amendment is satisfied, i.e., once an election intervenes.

Some examples are helpful to illustrate this point. If Congress were to hypothetically pass a law in 2027 changing congressional compensation from $174,000 to a permanent $200,000 effective immediately, such a law would clearly violate the Twenty Seventh Amendment by varying congressional compensation before the 2028 election intervened. Thus, congressional compensation for 2027 and 2028 would remain $174,000 (ignoring potential COLAs). However, once the 2028 election intervenes and the next Congress is seated, Boehner, 30 F.3d at 162, that law is free to take effect and vary congressional compensation from that point moving forward. Thus, while purporting to immediately change congressional compensation to $200,000 in 2027, the law, by operation of the Twenty-Seventh Amendment, would set compensation at $200,000 but only beginning in 2029, the first year that the Amendment allows the law to take effect following an election.

Or perhaps more helpful, imagine the new law instead specified that congressional compensation would be $200,000 in 2027, $210,000 in 2028, $220,000 in 2029, $230,000 in 2030, and so on. Under our understanding of the Twenty-Seventh Amendment, any variation in congressional compensation that the law purported to effectuate prior to the intervention of the next election is ineffective. Therefore, for years 2027 and 2028, congressional compensation would remain $174,000. However, the Twenty- Seventh Amendment would not render the entire law itself void ab initio. Instead, upon the intervention of the 2028 election and the seating of the next Congress, the law would be free to go into effect and apply by its terms. Thus, according to the terms of the law, congressional compensation would be $220,000 for 2029, $230,000 for 2030, and so on, but the provisions purporting to change congressional compensation for 2027 and 2028 would be of no effect. In other words, while the Twenty-Seventh Amendment does not allow Congress to alter the salary due it during its term, the Amendment does allow Congress to “specify the salary of the next Congress or of any Congress after that.” Id. at 162.

Under the government’s full retroactivity approach, however, a sitting Congress would be able to alter its own pay because any effort to do so that otherwise violates the Twenty-Seventh Amendment would simply take effect retroactively once the intervening election requirement is satisfied. Using the second example above, because the law facially violates the Twenty-Seventh

36 Amendment as to years 2027 and 2028 (by purporting to vary congressional pay before an election intervenes), congressional compensation would remain $174,000 during 2027 and 2028. However, once the 2028 election intervenes, the law would retroactively deem congressional compensation to have been $200,000 during 2027 and $210,000 during 2028, presumably requiring backpay to the Members of that prior Congress. In other words, as plaintiffs alluded to at oral argument, under the government’s view, Congress can do retroactively what it cannot do directly. See Oral Arg. Tr. 86, Feb. 20, 2026.

Perhaps recognizing that this would render the Twenty-Seventh Amendment arguably pointless, the government argues that the Amendment works by “ensuring that two consecutive congresses align on any law that varies congressional compensation: the Congress that enacted the law, and the Congress that—having been elected in the face of that enactment— allows it to go into effect.” Def.’s Resp. & Mot. Summ. J. 9, ECF No. 51. Thus, the government argues that the Twenty-Seventh Amendment conditions the operation of a law varying congressional compensation on the acceptance of that law by the next Congress. While the government argues that this is what the text of the Twenty-Seventh Amendment commands, see Def.’s Reply 3 n.2, ECF No. 60, we do not see how this can be so. The Amendment, by its terms, conditions the operation of a law varying congressional compensation on only one thing: the intervention of an election. Thus, as the D.C. Circuit has made clear, although a law varying congressional compensation “may be enacted at any time,” the law cannot take effect until an election intervenes, which occurs when two conditions are met: “when an election has been held the first condition is fulfilled; when the new Congress is seated the second condition is fulfilled.” Boehner, 30 F.3d at 162 (emphasis added). Thus, there is no third condition requiring the acquiescence of the next Congress after it has been seated, nor could there be, because, as the court in Boehner explained, laws varying congressional compensation can take effect under the Twenty-Seventh Amendment the moment the next Congress is seated. See id. The government’s full retroactivity approach, however, would require us to essentially rewrite the Twenty-Seventh Amendment as though it read as follows: “No law, varying the compensation for the services of the Senators and Representatives, shall take effect, until an election of Representatives shall have intervened and the next Congress allows it to go into effect.” There is no basis for doing so.

Adopting the government’s full retroactivity approach would negate the Amendment’s purpose almost entirely. This is because, without rewriting

37 the amendment to include an acquiescence requirement, the government’s appeal to the broad principle of political accountability in justification of its full retroactivity approach, see Def.’s Reply 3 n.2, ECF No. 60, entirely overlooks the fact that the Amendment’s primary purpose is to ensure that no variance in congressional compensation “be for the particular benefit of those who are concerned with determining the value of the service.” Boehner, 30 F.3d at 159 (citations omitted); see also Bernstein, supra, at 522 (quoting James Madison, Speech in the House of Representatives (June 8, 1789), in The Congressional Register, June 8, 1789, reprinted in Creating the Bill of Rights: The Documentary Record from the First Federal Congress 84 (Helen E. Veit et al., eds., 1991)) (It was the “seeming indecorum” of any one Congress setting its own pay that led to the proposal of the Twenty-Seventh Amendment.). Interpreting the Twenty-Seventh Amendment in such a way that would allow Congress to pass a law today that takes effect in the future and thus retroactively changes its current compensation would thwart the purpose of the Twenty-Seventh Amendment.

We thus reject the parties’ ultimate arguments (the void ab initio approach and the full retroactivity approach) and hold that laws varying congressional compensation are ineffective to the extent such laws seek to effectuate a change in congressional compensation before an election intervenes. But to the extent that a law is passed before the relevant intervening election and can be read to vary congressional compensation after the “expiration of the Congress that voted for it,” Boehner, 30 F.3d at 162, the law is free to take effect, at least in part, when the new Congress is seated.

The remaining question under this fourth issue, therefore, is whether and to what extent each COLA-blocking law during the relevant years can be applied prospectively following the intervention of the next election, i.e., whether the government’s partial retroactivity approach is tenable. There are two relevant questions specific to each COLA-blocking law: (1) whether the specific language of each COLA-blocking law is amenable to prospective application without judicial reformation, i.e., whether the law as it is written seeks, even if only in part, to vary the compensation “of the next Congress or of any Congress after that,” id. at 162; and, if so, (2) whether the appropriations bill to which each specific COLA-blocking law belongs allows for the prospective application of the COLA-blocking law following the intervention of the next election, see Beer, 696 F.3d at 1185 (citation omitted) (noting that Section 140 of Pub. L. 97-92 “expired in 1982” and could therefore have no effect after that date because “[t]he appropriations

38 act containing Section 140 expired by its terms on September 30, 1982”). While these issues were briefed to some extent, particularly by plaintiffs, we do not believe they are sufficiently briefed to render a decision thereon (especially considering the outstanding continuing claims doctrine question) and thus decline to do so today.

CONCLUSION

In sum, we hold (1) that plaintiffs have not waived and are not estopped from bringing these claims, (2) that the Twenty-Seventh Amendment applies to laws decreasing congressional compensation, (3) that COLA-blocking laws are laws “varying” congressional compensation under the Amendment, and (4) that laws varying congressional compensation are ineffective to the extent they seek to effectuate a change in congressional compensation before an election intervenes. The pending motions for summary judgment are denied in all other respects.

Looking forward, there are multiple other questions that we must answer pending further briefing. These include whether the continuing claims doctrine applies (which will help define the relevant years); whether the specific language of each COLA-blocking law for the relevant years is amenable to any prospective application, and, if so, whether each COLA- blocking law nonetheless expired before it could have taken effect under the Twenty-Seventh Amendment by virtue of being part of an appropriations bill; when an election “intervenes” for purposes of the Amendment; and, potentially, how to treat years 2012 and 2013 where the President “set GS COLAs to zero,” which would have capped the congressional COLA at zero even without Congress’s COLA-blocking legislation. The parties are directed to confer and propose by June 22, 2026, jointly if possible, a briefing schedule to address the issues remaining.

s/Eric G. Bruggink ERIC G. BRUGGINK Senior Judge