Climate Sols. v. State: Washington Clarifies “Horizontal” and “Vertical” Rational Unity Under Article II, Section 19 and Rejects an Omnibus Exception

I. Introduction

In Climate Solutions, et al. v. State of Washington, et al., No. 104240-0 (Wash. Sept. 17, 2026), the Washington Supreme Court reviewed the constitutionality of Initiative 2066 (I-2066), adopted by voters in the November 2024 general election. The initiative broadly sought to protect access to natural gas and to restrain public actions that might “discourage” gas use, while also rolling back or amending elements of Washington’s decarbonization and building-code frameworks.

Respondents (collectively “Climate Solutions”)—a coalition of advocacy organizations, local governments (including King County and the City of Seattle), and other entities—challenged I-2066 under Washington Constitution article II, sections 19 and 37. Appellants (the State of Washington and intervenors the Building Industry Assocation of Washington and Ashli Penner, collectively “BIAW”) defended the initiative.

The key issue ultimately resolved by the court was whether I-2066 violated the single-subject requirement of article II, section 19. The trial court had invalidated the initiative in full on multiple constitutional grounds; the Supreme Court affirmed on narrower grounds, holding the initiative unconstitutional in its entirety because it impermissibly embraced more than one subject and the offending subjects were not severable.

II. Summary of the Opinion

Chief Justice Stephens, writing for the majority, reaffirmed foundational principles governing Washington’s initiative power: initiatives are an exercise of the same sovereign legislative power as statutes and are subject to the same constitutional restraints (citing City of Burien v. Kiga, 144 Wn.2d 819 (2001)).

The court held:

  • I-2066 has a general title, so its provisions are read with some latitude, but must still share a single subject.
  • Article II, section 19 requires “rational unity” both among all provisions and between each provision and the title. The court expressly framed this as requiring both horizontal rational unity (provision-to-provision) and vertical rational unity (provision-to-title).
  • I-2066 contains multiple distinct subjects with “markedly distinct purposes and regulatory effects,” including:
    1. Affirmative requirements ensuring natural gas access (utilities, local governments, and constraints on the UTC).
    2. Broad restraints on local air pollution control agencies under Washington’s Clean Air Act, ch. 70A.15 RCW.
    3. Statewide building and energy code changes constraining the Washington State Building Code Council and removing specific zero-emission goals.
    4. Amendments to the Decarbonization Act, including repeal of its severability clause, which the court treated as legally operative and not mere “policy fluff.”
  • The court rejected the notion that I-2066 could be justified as “omnibus legislation” under In re Boot, 130 Wn.2d 553 (1996), clarifying that article II, section 19 contains no omnibus exception; comprehensiveness may sometimes demonstrate unity, but only when the measure addresses a single problem in a genuinely comprehensive way.
  • Because the distinct subjects were not severable (even with an internal severability clause), the initiative was invalid in its entirety.

Having resolved the case under the single-subject rule, the majority declined to reach additional constitutional claims under the subject-in-title requirement and article II, section 37 (silent amendment).

III. Analysis

A. Precedents Cited

1. Constitutional structure and initiative power

  • State v. Gunwall, 106 Wn.2d 54 (1986): cited for the principle that Washington’s constitution restrains otherwise plenary state power, framing article II, section 19 as a meaningful limit rather than a technicality.
  • City of Burien v. Kiga, 144 Wn.2d 819 (2001): used to equate the people’s initiative power with legislative power, making constitutional restraints fully applicable to initiatives; also foundational to the rational-unity analysis and the “general title” framework.
  • Union High Sch. Dist. No. 1 v. Taxpayers of Union High Sch. Dist. No. 1, 26 Wn.2d 1 (1946): cited for article II, section 19 as a restraint on legislative power, situating the single-subject rule in structural constitutional design.

2. Purposes of article II, section 19: logrolling, transparency, scrutiny

  • Patrice v. Murphy, 136 Wn.2d 845 (1998), and State ex rel. Wash. Toll Bridge Auth. v. Yelle, 32 Wn.2d 13 (1948): cited for the “threefold” purposes of article II, section 19, especially preventing logrolling and hidden legislation.
  • Garfield County Transp. Auth. v. State, 196 Wn.2d 378 (2020): heavily relied on for modern application of anti-logrolling principles and the court’s willingness to enforce the mandate even against voter-adopted measures; also referenced for concerns with mixing statewide provisions and localized directives.

3. The rational unity test and “general title” breadth

  • Filo Foods, LLC v. City of SeaTac, 183 Wn.2d 770 (2015): cited for the initial “general vs. restrictive title” step, controlling the degree of liberal construction.
  • Wash. Ass'n for Substance Abuse & Violence Prevention v. State, 174 Wn.2d 642 (2012), and State v. Grisby, 97 Wn.2d 493 (1982): cited for the “some rational unity” formulation; also for the relevance of legislative history and historical pairing of issues to show unity.
  • Amalg. Transit Union Loc. 587 v. State, 142 Wn.2d 183 (2000): central authority for distinguishing the single-subject and subject-in-title prohibitions; emphasized for the proposition that nominal connection is insufficient where operative effects diverge and for the severability framework.
  • Barde v. State, 90 Wn.2d 470 (1978): used to show that provisions can be nominally related (there, “personal property”) yet still lack rational unity when their direct regulatory effects differ.

4. Illustrations of permissible breadth (and limits) under general titles

  • Citizens for Responsible Wildlife Mgmt. v. State, 149 Wn.2d 622 (2003): example of permissible inclusion of related subtopics (multiple methods of trapping and killing mammals) and caution that “broad, long term and continuing” versus narrow provisions may suggest logrolling.
  • Wash. Fed'n of State Emps. v. State, 127 Wn.2d 544 (1995): supports allowing provisions that are a “rational nexus” to the general subject (campaign contribution limits) and illustrates how repeal provisions can be germane when they implement the central purpose.
  • Wash. Ass'n of Neigh. Stores v. State, 149 Wn.2d 359 (2003), abrogated in part on other grounds by Filo Foods, LLC v. City of SeaTac, 183 Wn.2d 770: cited for the idea that provisions necessary to implementing one another are generally rationally related.
  • Pierce County v. State, 150 Wn.2d 422 (2003): cited for treating nonoperative “policy fluff” as not generating additional subjects—important here because the court contrasted nonoperative statements with operative severability language.

5. Illustrations of impermissible combinations of subjects

  • Wash. Toll Bridge Auth. v. State, 49 Wn.2d 520 (1956): used to show that combining a one-time authorization with systemic changes can reflect disunity and possible logrolling.
  • City of Burien v. Kiga, 144 Wn.2d 819: relied upon for invalidating initiatives combining a one-time refund/repeal with broader systemic tax changes and for the proposition that courts cannot know what would have passed absent the other subject.
  • Garfield County Transp. Auth. v. State, 196 Wn.2d 378: used for the “unrelated local and statewide effects” problem and the specific-directive “errant” provision (Sound Transit bond directive) that broke unity.

6. Omnibus legislation and the “single problem comprehensively addressed” concept

  • In re Boot, 130 Wn.2d 553 (1996): the State and BIAW invoked this to justify combining diverse provisions; the majority distinguished it and clarified its limits, emphasizing no categorical “omnibus exception” exists in article II, section 19.
  • Doriot v. State, 32 Wn. App. 2d 770 (2024): cited as another example where diverse provisions were unified by a comprehensive approach to a broad, multifaceted problem (transportation resources).

7. Severability doctrine and the legal effect of severability clauses

  • El Centro De La Raza v. State, 192 Wn.2d 103 (2018) (plurality opinion), and Gerberding v. Munro, 134 Wn.2d 188 (1998): used to explain that severability clauses have legal effect by informing whether valid portions would have been enacted absent invalid ones.
  • League of Women Voters of Wash. v. State, 184 Wn.2d 393 (2015), State v. Broadaway, 133 Wn.2d 118 (1997), and In re Parentage of C.A.M.A., 154 Wn.2d 52 (2005): cited for the modern severability test and the caution against judicial “drastic rewrite.”
  • Lee v. State, 185 Wn.2d 608 (2016): relied upon to reinforce the inability to disentangle voter intent between separate subjects and to reject severance where doing so would require the court to choose among subjects.
  • Sane Transit v. Sound Transit, 151 Wn.2d 60 (2004): cited for the premise that voters are presumed to enact the measure in its entirety, limiting reliance on title alone to infer severability.

B. Legal Reasoning

1. The opinion’s doctrinal clarification: “horizontal” and “vertical” rational unity

A major contribution of the decision is its explicit articulation that the “rational unity” inquiry has two dimensions:

  • Vertical rational unity: each provision must rationally relate to the general subject expressed in the title.
  • Horizontal rational unity: provisions must rationally relate to one another, not merely to the title.

This matters because defendants often attempt to define an initiative’s “subject” at a high level of abstraction (here, “preserving energy choice” or “protecting natural gas access”) and then argue every provision fits within that abstraction. The court’s approach rejects “title-only unity” as insufficient: even if provisions can be characterized as fitting under a broad label, they still must cohere with one another in operative purpose and regulatory effect.

2. Distinct subjects identified by distinct regulatory effects

The court’s single-subject violation finding rested on its assessment that I-2066 contained provisions whose primary purposes and effects diverged in ways that cannot be treated as merely “incidental subdivisions” of one subject:

  • Natural gas access mandates (sections 2-5) were framed as direct regulation of providers and regulators (utilities, cities/towns in some contexts, and the UTC) to ensure gas availability and prevent rate-plan mechanisms that restrict it.
  • Air quality regulation constraint (section 11) was treated as a separate subject because it rewrote the enforcement capacity of local air pollution control agencies “in any way” to avoid discouraging gas use, creating what the court described as a “glaring loophole” in the Clean Air Act’s regulatory structure. The court emphasized that air agencies do not regulate energy supply/demand; they regulate air quality, with energy impacts only incidental.
  • Building and energy code changes (sections 6-8) were treated as distinct because they reoriented statewide construction standards and constrained the Washington State Building Code Council—regulation that is structurally different from mandating utility service or constraining air regulators.
  • Repeal of the Decarbonization Act’s severability clause (section 12) was deemed legally operative and a separate subject because severability affects judicial treatment of statutory invalidity and legislative intent, unlike nonoperative policy statements (distinguished from Pierce County v. State).

3. Rejection of “omnibus” as a categorical safe harbor

The court responded to the State/BIAW reliance on In re Boot by clarifying that “omnibus” is descriptive, not doctrinal: article II, section 19 does not contain an “omnibus exception.” Instead, In re Boot and Doriot v. State are best read as cases where the breadth of a single problem (violence prevention; transportation resources) justified diverse but still unified provisions.

By contrast, the court described I-2066 as “patchwork,” not comprehensive: even under a broad framing (“energy choice”), the initiative did not comprehensively regulate energy choice statewide; it instead advanced several discrete policy moves (utility mandates, local-government prohibitions, air agency constraints, code changes, and severability repeal).

4. Severability: why the whole initiative fell

Even though I-2066 contained its own severability clause (section 13), the court invalidated the entire measure. Two related ideas drove that result:

  • Anti-logrolling fidelity: when a measure contains multiple distinct subjects, severing one subject risks judicially reconstructing the “deal” that voters accepted as a package.
  • Unknowable voter intent: relying on Power, Inc. v. Huntley, City of Burien v. Kiga, and Lee v. State, the court reasoned that it is “impossible to determine how many people voted for one provision and how many for the other” when the vote was on a bundled measure.

The court nevertheless applied the severability framework from Amalg. Transit Union Loc. 587 v. State, League of Women Voters of Wash. v. State, and State v. Broadaway, emphasizing that courts may sever only when they can presume the enacting body (here, the voters) would have enacted the remainder without the invalid parts, and when severance would not require a “drastic rewrite” (citing In re Parentage of C.A.M.A.). In the majority’s view, selecting which “subject” to preserve would force the judiciary to rank subjects (primary vs. secondary), exceeding permissible judicial construction.

5. The dissent’s competing approach (Justice Mungia)

Justice Mungia agreed that subsection 12(5) (repeal of the Decarbonization Act severability clause) was not rationally related, but would have severed it and upheld the remainder. The dissent also argued that sections 9-11 should be severed as silent amendments under article II, section 37 (an issue the majority did not reach). The majority rejected this severance-oriented approach as requiring an impermissible judicial choice among “equal and distinct” subjects and as inconsistent with Lee v. State.

C. Impact

1. Stronger, clearer enforcement of single-subject limits in initiatives

This decision reinforces that Washington’s single-subject rule is not satisfied by defining a broad label (“energy choice,” “natural gas access”) and aggregating heterogeneous regulatory programs beneath it. By formalizing “horizontal” and “vertical” rational unity, the court provides litigants and drafters a clearer roadmap for both challenging and defending initiatives:

  • Even a general title does not permit bundling distinct regulatory regimes with separate institutional targets and public-policy aims.
  • Courts will focus on operative effect and primary purpose, not only semantic overlap.

2. Narrowing the practical reach of “omnibus” arguments

By clarifying that In re Boot does not create an “omnibus exception,” the court limits a common defense strategy: characterizing a contested initiative as a comprehensive response to a problem. Future proponents should expect to demonstrate genuine comprehensiveness toward a single problem, not merely a collection of coordinated policy preferences.

3. Drafting consequences for energy and climate-related initiatives (and beyond)

The opinion signals heightened vulnerability for initiatives that simultaneously:

  • mandate service or entitlements (e.g., utility obligations),
  • constrain regulatory enforcement authorities (e.g., air agencies), and
  • rewrite statewide codes or standards (e.g., building codes),

unless those components are tightly integrated and mutually implementing. Initiatives that use expansive “discourage” prohibitions—particularly when tethered to public-health regulatory programs—may be viewed as introducing distinct subjects.

4. Severability clause repeal as a substantive “subject”

The court’s treatment of repealing another act’s severability clause as legally meaningful may have broader ripple effects. Drafters who include “structural” or “litigation posture” provisions (like severability manipulation) should anticipate scrutiny under article II, section 19 as to whether such provisions cohere with the initiative’s operative regulatory program.

IV. Complex Concepts Simplified

  • Single-subject rule (article II, section 19): a bill/initiative must not combine separate topics into one package vote. The point is to prevent “take the good with the bad” bundling.
  • Logrolling: attaching an unpopular proposal to a popular one so the whole package passes, even if the unpopular part would fail alone.
  • General vs. restrictive title: a general title is broad and is read more liberally; a restrictive title is narrow and limits what may be included.
  • Rational unity: the required logical connection among provisions and between provisions and the title. This decision clarifies:
    • Vertical rational unity = each section must relate to the title’s general subject.
    • Horizontal rational unity = each section must relate to the others; it is not enough that each relates to the title.
  • Severability clause: a provision stating that if part of a law is invalidated, the rest should remain effective. The court stressed it has legal effect because it influences how courts decide whether remaining sections can stand.
  • Severance: a court’s act of cutting out unconstitutional pieces while leaving the rest. The court refused severance here because it could not know what voters would have passed if the subjects had been separated.
  • Omnibus legislation: a large bill covering many subtopics. This case clarifies it is not automatically allowed; the question remains whether all parts address one problem comprehensively and share rational unity.

V. Conclusion

Climate Solutions, et al. v. State of Washington, et al. is a significant article II, section 19 decision for Washington initiative practice. The court (1) crystallizes the rational-unity inquiry into “horizontal” and “vertical” requirements, (2) emphasizes that distinct regulatory effects and purposes can create distinct “subjects” even under a general title, (3) rejects any categorical “omnibus exception” derived from In re Boot, and (4) underscores the judiciary’s limited role in salvaging multi-subject initiatives through severance where voter intent cannot be reliably disentangled. The practical message to initiative drafters is straightforward: if policy goals implicate materially different regulatory regimes, they must be separated into distinct measures or integrated so tightly that each provision genuinely implements the same unified subject.