Bradley Boardman v. Jay Inslee

FOR PUBLICATION

UNITED STATES COURT OF APPEALS
FOR THE NINTH CIRCUIT

BRADLEY BOARDMAN, a Washington No. 19-35113
Individual Provider; DEBORAH
THURBER, a Washington Family D.C. No.
Childcare Provider; SHANNON BENN, 3:17-cv-05255-
a Washington Family Childcare BHS
Provider; FREEDOM FOUNDATION, a
Washington nonprofit organization,
Plaintiffs-Appellants, OPINION

v.

JAY ROBERT INSLEE, Governor of the
State of Washington; ROBERT
HINES*, Director of the Washington
Department of Social and Health
Services (DSHS); ROSS HUNTER,
Secretary of the Washington

*
Robert Hines is substituted for his predecessor, Patricia Lashway,
as the Director of the Washington Department of Social and Health
Services (“DSHS”). Fed. R. App. P. 43(c)(2).
2 BOARDMAN V. INSLEE

Department of Children, Youth, and
Families (DCYF)**,
Defendants-Appellees,

CAMPAIGN TO PREVENT FRAUD AND
PROTECT SENIORS,
Intervenor-Defendant-Appellee.

Appeal from the United States District Court
for the Western District of Washington
Benjamin H. Settle, District Judge, Presiding

Argued and Submitted February 4, 2020
Seattle, Washington

Filed October 22, 2020

Before: Milan D. Smith, Jr., N. Randy Smith, and
Daniel A. Bress, Circuit Judges.

Opinion by Judge N.R. Smith;
Dissent by Judge Bress

**
On July 6, 2017, the Washington Department of Early Learning
(“DEL”) ceased to exist as an independent state agency and its functions
were transferred to the newly formed Washington Department of Children,
Youth, and Families (“DCYF”). See Act of July 6, 2017, 2017 Wash.
Laws ch. 6, § 802(1) (“The [DEL] is hereby abolished and its powers,
duties, and functions are hereby transferred to the [DCYF].”). Thus, the
DCYF has been substituted for the DEL in this appeal.
 BOARDMAN V. INSLEE 3

SUMMARY**

Civil Rights

The panel affirmed the district court’s summary judgment
in favor of Washington state defendants in an action brought
pursuant to 42 U.S.C. § 1983 challenging the constitutionality
of ballot initiative 1501, passed by Washington voters, which
prohibits public access to certain government-controlled
information, including the personal information of in-home
care providers, but permits that information to be disclosed to
the providers’ certified exclusive bargaining representatives.

Plaintiffs-Appellants are a nonprofit organization and
individual in-home providers who are required by
Washington law to participate in statewide collective
bargaining. Although the individual Appellants are members
of their collective bargaining units, they are not members of
their respective unions and do not pay agency or “fair share”
fees—fees paid to a union by nonmembers to support
activities that are germane to a union’s duties as a collective-
bargaining representative. Appellants seek to inform other
individual in-home providers of their right to opt out of
paying agency fees to their unions. After the passage of
Initiative 1501, however, Appellants were unable to obtain
the contact information of other in-home providers.
Appellants allege that Part III of Initiative 1501 violates the
First Amendment by discriminating among viewpoints and
impairing their freedom of association. Additionally, they
allege that the initiative transgresses the Equal Protection

**
This summary constitutes no part of the opinion of the court. It has
been prepared by court staff for the convenience of the reader.
4 BOARDMAN V. INSLEE

Clause of the Fourteenth Amendment, because it burdens
their fundamental rights and is motivated by animus.

Citing Houchins v. KQED, Inc., 438 U.S. 1, 15 (1978)
(plurality opinion), the panel first acknowledged the well-
settled principle that the First Amendment does not guarantee
a general right of access to government information or
sources of information within the government’s control. The
panel further stated that although the decision whether to
disclose government-controlled information is generally a
task which the Constitution has left to the political processes,
the First Amendment forbids a state from discriminating
invidiously among viewpoints in the provision of information
within its control. However, a state does not engage in
viewpoint discrimination by simply disclosing the personal
information of public or quasi-public employees to the
employees’ certified exclusive collective bargaining
representative, while denying equal access to the public. The
panel concluded that Initiative 1501 did not impose
viewpoint-based conditions on the disclosure of provider
information, did not discriminate between the competing
views of Appellants and their Service Employee Unions, and
did not implicate the individual Appellants’ associational
freedom.

The panel rejected Appellants’ claim that Initiative 1501
violates the First Amendment rights of other in-home care
providers by denying the providers the right to determine for
themselves if they want to hear plaintiffs’ messages. The
panel held that Appellants did not have standing to assert the
rights of other in-home care providers.

Finally, the panel rejected Appellants’ claim that
Initiative 1501 violated the Equal Protection Clause. The
 BOARDMAN V. INSLEE 5

panel concluded that the challenged provisions of Initiative
1501 satisfied rational-basis review. The panel stated that the
State has a legitimate interest in protecting seniors and other
vulnerable individuals—and all of its residents, for that
matter—from identity theft and other financial crimes, and
Washington voters could have rationally decided that
generally prohibiting public access to the personal
information of in-home care providers—many of whom work
within the homes of their clients—would further that interest.
The panel further rejected Appellants’ contention that
Initiative 1501 was motivated by animus. The panel stated
that here was no evidence in the record indicating that the
more than 2.2 million Washington voters who voted in favor
of Initiative 1501 were motivated by an irrational prejudice,
or a bare desire to harm Appellants or their message against
the Unions.

Dissenting, Judge Bress stated that the text and operation
of I-1501 and a troubling documentary record demonstrate
exactly what was going on here: transparent viewpoint
discrimination. The State was effectively using an
information embargo to promote the inherently “pro-union”
views of the incumbent unions, while making it vastly more
difficult for those with opposing views—and particularly
those with views opposite unions—to reach their intended
audience. Judge Bress would hold that I-1501 fails First
Amendment scrutiny.

COUNSEL

Susan K. Stahlfeld (argued), Miller Nash Graham & Dunn
LLP, Seattle, Washington; Caleb Jon F. Vandenbos, Freedom
Foundation, Olympia, Washington; for Plaintiffs-Appellants.
6 BOARDMAN V. INSLEE

Peter B. Gonick (argued) and Callie A. Castillo, Deputy
Solicitors General; Robert W. Ferguson, Attorney General,
Office of the Attorney General, Olympia, Washington; for
Defendants-Appellees.

Gregory J. Wong (argued), Paul J. Lawrence, and Claire E.
McNamara, Pacifica Law Group LLP, Seattle, Washington,
for Intervenor-Defendant-Appellee.

OPINION

N.R. SMITH, Circuit Judge:

Although the decision whether to disclose government-
controlled information is generally a “task which the
Constitution has left to the political processes,” Houchins v.
KQED, Inc., 438 U.S. 1, 12 (1978) (plurality opinion), the
First Amendment forbids a state from discriminating
invidiously among viewpoints in the provision of information
within its control. However, a state does not engage in
viewpoint discrimination by simply disclosing the personal
information of public or quasi-public employees to the
employees’ exclusive collective bargaining representative,
while denying equal access to the public. See Perry Educ.
Ass’n v. Perry Local Educators’ Ass’n, 460 U.S. 37, 49
(1983).

Initiative 1501, a ballot initiative passed by Washington
voters, prohibits public access to certain government-
controlled information, including the personal information of
in-home care providers. Under the initiative, in-home care
providers’ personal information must still be disclosed to
their exclusive bargaining representatives. Appellants, who
 BOARDMAN V. INSLEE 7

have been denied access to this information, challenge
Initiative 1501 under the First Amendment and the Equal
Protection Clause of the Fourteenth Amendment. The district
court rejected Appellants’ claims on summary judgment. So
do we.

I

In this opinion, we are concerned only with the
constitutionality of Initiative 1501. However, in order to
address that issue, we must begin our analysis a few years
prior to the initiative’s passage.

A

Appellants, Bradley Boardman, Deborah Thurber, and
Shannon Benn, provide home-based care services in the State
of Washington. Boardman is an “individual provider” who
provides Medicaid-funded healthcare services for a disabled
family member. See Wash. Rev. Code § 74.39A.240(3)
(“‘Individual provider’ means a person . . . who, under an
individual provider contract with the department [of social
and health services] or as an employee of a consumer directed
employer, provides personal care or respite care services to
persons who are functionally disabled or otherwise eligible
under [various Medicaid- or state-funded programs].”).
Thurber and Benn are “family child care providers” who
provide state-subsidized childcare services for low-income
families. See id. § 41.56.030(7) (“‘Family child care provider’
means a person who: (a) Provides regularly scheduled care
for a child or children in the home of the provider or in the
home of the child or children for periods of less than
twenty-four hours or, if necessary due to the nature of the
parent’s work, for periods equal to or greater than twenty-four
8 BOARDMAN V. INSLEE

hours; (b) receives child care subsidies; and (c) . . . is either
licensed by the state or is exempt from licensing.”).

As in-home care providers, Appellants are required by
Washington law to participate in statewide collective
bargaining. Id. §§ 41.56.028, 74.39A.270. Individual
providers and family child care providers compose two
separate collective bargaining units, each of which is
represented by an exclusive bargaining representative elected
by a majority of providers within the unit. Id.
§§ 41.56.028(2), 74.39A.270(2)(a), (7). Currently, Service
Employees International Union Healthcare 775NW (“SEIU
775”) represents individual providers, and Service Employees
International Union Local 925 (“SEIU 925”) represents
family child care providers (collectively, the “Unions”). As
the exclusive bargaining representatives of in-home care
providers, the Unions must represent all the providers within
each of their collective bargaining units without regard to
union membership. Id. § 41.56.080.

Although they are members of their collective bargaining
units, Boardman, Thurber, and Benn are not members of the
Unions. They do not pay agency or “fair share” fees—fees
paid to a union by nonmembers to support “activities that are
‘germane to the union’s duties as collective-bargaining
representative,’” see Janus v. Am. Fed’n of State, Cty., and
Mun. Emps., Council 31, 138 S. Ct. 2448, 2460 (2018)
(alteration adopted) (quoting Abood v. Detroit Bd. of Educ.,
431 U.S. 209, 235 (1977), overruled by Janus, 138 S. Ct. at
2486), nor do they otherwise support the Unions’ collective
bargaining efforts on their behalf. In fact, Thurber and Benn
serve on the board of directors of Pacific Northwest Family
Child Care Association (“PNFCCA”), a rival union of SEIU
925. With the help of Appellant Freedom Foundation (a
 BOARDMAN V. INSLEE 9

§ 501(c)(3) non-profit organization), Thurber and Benn have
campaigned for several years to replace SEIU 925 with
PNFCCA as the exclusive bargaining representative for
family child care providers. To do so, PNFCCA must first
obtain 30% support among family child care providers to
force an election, a milestone they have yet to achieve. See
Wash. Rev. Code § 41.56.070. The Freedom Foundation has
also assisted Boardman in his efforts to inform other
individual providers of their right to opt out of paying agency
fees to SEIU 775.1

To promote their various causes, Appellants have had to
obtain the contact information of in-home care providers.
This has not been a simple task. More than 40,000 providers
work in homes dispersed throughout Washington. They do
not share workplaces, supervisors, or clients, and they have
a notably high turnover rate. Consequently, to collect in-home
care providers’ contact information, Appellants have relied
largely on Washington’s Public Records Act (“PRA”), which
requires state agencies to “make available for public
inspection and copying all public records,” unless a specific
exemption applies. Id. § 42.56.070(1); see also Doe ex rel.
Roe v. Wash. State Patrol, 374 P.3d 63, 67 (Wash. 2016)
(“Despite the PRA’s presumption of openness and
transparency, the legislature has made certain public records
exempt from production.”). Through PRA requests,
Appellants have obtained lists of in-home care providers’
personal information that had been maintained in the custody

1
In Harris v. Quinn, 573 U.S. 616 (2014), the Supreme Court held
that, under the First Amendment, quasi-public employees (like in-home
care providers in the State of Washington) could not be compelled to pay
agency fees to a union representative who they “do not want to join or
support.” Id. at 656.
10 BOARDMAN V. INSLEE

of the Department of Social and Health Services (“DSHS”)
and the Department of Early Learning (“DEL”) (now the
Department of Children, Youth, and Families
(“DCYF”))—the state agencies that administer home-care
programs. See Wash. Rev. Code § 74.39A.240; Serv. Emps.
Int’l Union Local 925 v. Dep’t of Early Learning, 450 P.3d
1181, 1182 (Wash. 2019). With this information, Appellants
were able to identify in-home care providers and contact them
through the mail and by door-to-door canvassing, which
proved to be a relatively efficient way of spreading their
messages and voicing their opposition to the Unions.

Naturally, the Unions opposed Appellants’ efforts to
replace SEIU 925 with a rival union and to otherwise
diminish the Unions’ support among in-home care providers.
In 2014, the Unions began challenging Appellants’ PRA
requests in state court, seeking to enjoin disclosures of
DSHS’s and DEL’s lists of in-home care providers’
information. For a time, Appellants were successful in
obtaining in-home care providers’ information, despite the
litigation spawned by their PRA requests. But due to in-home
care providers’ high turnover, these lists soon became
outdated. Therefore, Appellants continued to request updated
lists from DSHS and DEL to facilitate in spreading their
messages among a wide audience of in-home care providers.

B

In November 2016, Washington voters passed Initiative
1501, a ballot initiative designed to “protect the safety and
security of seniors and vulnerable individuals” (e.g., the
elderly and individuals with developmental disabilities, see
Wash. Rev. Code § 9.35.005(7)) from various financial
crimes. Seniors and Vulnerable Individuals’ Safety and
 BOARDMAN V. INSLEE 11

Financial Crimes Prevention Act, 2017 Wash. Legis. Serv.,
ch. 4, I.M. No. 1501, § 2 (West). Initiative 1501 was
comprised of three parts. Part I amended the definition of
first-degree identity theft under Washington’s criminal code
to include “knowingly target[ing] a senior or vulnerable
individual in” committing an identity crime. Id. § 6 (codified
at Wash. Rev. Code § 9.35.020(2)). Part II enacted a treble-
damages provision applicable in any civil cause of action in
which the plaintiff “was victim to consumer fraud that
targeted him or her as a senior or vulnerable individual.” Id.
(codified at Wash. Rev. Code § 9.35.060).

Part III of Initiative 1501—which forms the basis of this
dispute—exempted a new category of public records from
disclosure under the PRA. It provided:

To protect vulnerable individuals and their
children from identity crimes and other forms
of victimization, neither the state nor any of
its agencies shall release sensitive personal
information of vulnerable individuals or
sensitive personal information of in-home
caregivers for vulnerable populations, as
those terms are defined in [Revised Code of
Washington section 42.56.640].

Id. § 10 (codified at Wash. Rev. Code. § 43.17.410)
(emphasis added). Additionally, Part III directly amended the
PRA to exempt the sensitive personal information of
vulnerable individuals and their in-home care providers from
“inspection and copying under [the PRA].” Id. § 8 (codified
at Wash. Rev. Code § 42.56.640(1)). “Sensitive personal
information” was defined to cover an individual’s name,
addresses, contact information, and “other personally
12 BOARDMAN V. INSLEE

identifying information.” Id. (codified at Wash. Rev. Code
§ 42.56.640(2)(b)).

With the passage of Initiative 1501, the sensitive personal
information of vulnerable individuals and in-home care
providers joined hundreds of other categories of public
records that have been exempted from disclosure under the
PRA. “The general purpose of the exemptions to the [PRA’s]
broad mandate of disclosure is to exempt from public
inspection those categories of public records most capable of
causing substantial damage to the privacy rights of citizens or
damage to vital functions of government.” Limstrom v.
Ladenburg, 963 P.2d 869, 875 (Wash. 1998). “There are three
sources of PRA exemptions: (1) enumerated exemptions
contained in the PRA itself, (2) any ‘other statute’ that
exempts or prohibits disclosure . . . , and (3) the Washington
Constitution.” SEIU 775 v. State Dep’t of Soc. & Health
Servs., 396 P.3d 369, 372 (Wash. Ct. App. 2017) (quoting
Wash. Rev. Code § 42.56.070(1)). These exemptions cover a
wide range of information, from certain law enforcement and
crime data, public employees’ personal information, and
proprietary records, see, e.g., Wash Rev. Code §§ 42.56.240,
42.56.250, 42.56.270, to library records, id. § 42.56.310,
commercial fishing catch and shellfish harvest data, id.
§ 42.56.430(1), (8), and maps of archaeological sites, id.
§ 42.56.300(1).

Initiative 1501’s prohibition on public access to the
personal information of vulnerable individuals and their in-
home care providers was not without exceptions. The
initiative authorized disclosure of information under several
narrow circumstances, including when “disclosure is required
by federal law” or “by a contract between the state and a
third party,” or when “[t]he information is being released
 BOARDMAN V. INSLEE 13

as part of a judicial or quasi-judicial proceeding.”2

2
Part III of Initiative 1501 provided:

Nothing . . . shall prevent the release of public
information in the following circumstances:

(a) the information is released to a governmental
body, including the state’s area agencies on aging,
and the recipient agrees to protect the
confidentiality of the information;

(b) the information concerns individuals who have
been accused of or disciplined for abuse, neglect,
exploitation, abandonment, or other acts involving
the victimization of individuals or other
professional misconduct;

(c) the information is being released as part of a
judicial or quasi-judicial proceeding and subject to
a court’s order protecting the confidentiality of the
information and allowing it to be used solely in
that proceeding;

(d) the information is being provided to a
representative certified or recognized under RCW
41.56.080, or as necessary for the provision of
fringe benefits to public employees, and the
recipient agrees to protect the confidentiality of the
information;

(e) the disclosure is required by federal law;

(f) the disclosure is required by a contract between
the state and a third party, and the recipient agrees
to protect the confidentiality of the information;

(g) the information is released to a person or entity
under contract with the state to manage,
administer, or provide services to vulnerable
14 BOARDMAN V. INSLEE

Seniors and Vulnerable Individuals’ Safety and Financial
Crimes Prevention Act, 2017 Wash. Legis. Serv., ch. 4, I.M.
No. 1501, § 11(c), (e), (f) (West) (codified at Wash. Rev.
Code § 42.56.645(1)(c), (e), (f)). Information could also be
lawfully disclosed if it were “being provided to a
representative certified or recognized under [Wash. Rev.
Code section] 41.56.080,” id. § 11(d) (codified at Wash. Rev.
Code § 42.56.645(1)(d)), which is an exclusive bargaining
representative of a collective bargaining unit, Wash. Rev.
Code § 41.56.080 (“The bargaining representative which has
been determined to represent a majority of the employees in
a bargaining unit shall be certified by the commission as the
exclusive bargaining representative of, and shall be required
to represent, all the public employees within the unit
without regard to membership in said bargaining
representative . . . .”).3

residents, or under contract with the state to engage
in research or analysis about state services for
vulnerable residents, and the recipient agrees to
protect the confidentiality of the information; or

(h) information about specific public employee(s)
is released to a bona fide news organization that
requests such information to conduct an
investigation into, or report upon, the actions of
such specific public employee(s).

Seniors and Vulnerable Individuals’ Safety and Financial Crimes
Prevention Act, 2017 Wash. Legis. Serv., ch. 4, I.M. No. 1501, § 11
(codified at Wash. Rev. Code § 42.56.645(1)).
3
The Unions have had the ability to obtain access to Provider
Information under this exception, because they are both currently
“representative[s] certified or recognized under [Wash. Rev. Code section]
41.56.080.” See Wash. Rev. Code § 42.56.645(1)(d).
 BOARDMAN V. INSLEE 15

Boardman, Thurber, and the Freedom Foundation were
active in their opposition to Initiative 1501. Together, they
prepared the “Argument Against” statement for the initiative
in the “Voter’s Guide” to the 2016 election. In their
statement, they alleged that the Unions drafted and supported
Initiative 1501 “to prevent in-home caregivers and childcare
providers from learning they no longer can be forced to pay
dues to the union” and to “protect[] union bosses’ wallets.”
“If Initiative 1501 passes,” they claimed, “caregivers will not
even be able to contact each other to discuss issues of
common concern.” The “Argument For” statement was
prepared by two in-home care providers, an elder advocate,
the King County Sheriff, and a representative of the Puget
Sound Advocates for Retirement Action. They argued that
Initiative 1501 would help prevent “fraudulent telemarketers”
and “identity thieves” from “targeting seniors and the
vulnerable.”

The chief proponent of Initiative 1501 was the “Campaign
to Prevent Fraud and Protect Seniors” (“Campaign”), a
campaign committee registered with the State of Washington
to “spearhead[] the campaign” in favor of the initiative. The
Campaign was composed of “unions, advocates for seniors,
public safety officials, and community groups.” Its literature
was unequivocal in its support for Initiative 1501: “Groups
like the Freedom Foundation are threatening unions. . . .
Initiative 1501 will keep groups like the Freedom Foundation
from getting our personal information. Vote yes on I-1501 to
keep our unions strong and protect what we’ve fought for.”
The Campaign received substantial contributions from the
Unions. It was also chaired by SEIU 775’s secretary-
treasurer. SEIU 775 echoed the Campaign’s support for
Initiative 1501 in an email to its members: “There’s one more
way you can fight to stop the Freedom Foundation: When you
16 BOARDMAN V. INSLEE

get your ballot in the mail, vote YES on I-1501, which
protects the private information of caregivers and our state’s
most vulnerable.”

In the final tally, nearly 71% of Washington voters
supported the passage of Initiative 1501. The initiative took
effect on December 8, 2016.

C

After Initiative 1501 took effect, Boardman, Thurber, and
Benn made PRA requests to DSHS and DEL for updated lists
of in-home care providers’ personal information (hereafter,
“Provider Information”). All of their requests were denied.
Appellants then filed suit in federal district court seeking
declaratory and injunctive relief under 42 U.S.C. § 1983.4
Among other claims, Appellants alleged that Part III of
Initiative 1501 violated the First Amendment by
discriminating among viewpoints and impairing their
freedom of association. Additionally, Appellants alleged that
the initiative transgressed the Equal Protection Clause of the
Fourteenth Amendment, because it burdened their
fundamental rights and was motivated by animus. The
Campaign intervened in the action to assist in defending the
constitutionality of Initiative 1501. On cross-motions for
summary judgment, the district court granted judgment in
favor of the State and the Campaign on all of Appellants’

4
When this suit was filed, Appellants named Jay Inslee, Governor of
the State of Washington, Patricia Lashway, then-Director of DSHS, and
Ross Hunter, then-Director of DEL, as defendants (collectively, the
“State”).
 BOARDMAN V. INSLEE 17

claims.5 Appellants appealed the district court’s judgment and
renewed the foregoing claims.

II

We have jurisdiction to review the district court’s
decision on cross-motions for summary judgment under 28
U.S.C. § 1291, and we review this decision de novo, Guatay
Christian Fellowship v. County of San Diego, 670 F.3d 957,
970 (9th Cir. 2011). “[W]e view the evidence in the light
most favorable to the nonmoving party, determine whether
there are any genuine issues of material fact, and decide
whether the district court correctly applied the relevant
substantive law.” Animal Legal Def. Fund v. U.S. Food and
Drug Admin., 836 F.3d 987, 989 (9th Cir. 2016) (per curiam).
Summary judgment is appropriate where “there is no genuine
dispute as to any material fact and the movant is entitled to
judgment as a matter of law.” Fed. R. Civ. P. 56(a).

III

Appellants primarily contend that, under the statutory
provisions codifying Initiative 1501, they are denied access
to Provider Information based on the views they espouse on
the subject of collective bargaining. Before we address the
merits of this claim, we must consider the State’s threshold
argument “that laws restricting public access to records do
not implicate the First Amendment.” On this score, the State
argues that the disclosure of government-controlled
information is “wholly within the legislative power,” such

5
Both the State and the Campaign are Appellees in this case. For
ease of reference, we will refer only to the State when discussing the
arguments raised by both parties in their joint answering brief.
18 BOARDMAN V. INSLEE

that the decision to restrict access to Provider Information is
beyond First Amendment scrutiny. The State urges us to
reject Appellants’ claim on this basis alone.

A

Houchins first announced the well-settled principle that
the First Amendment does not guarantee a general “right of
access to government information or sources of information
within the government’s control.” 438 U.S. at 15; id. at 16
(Stewart, J., concurring in the judgment) (“The First and
Fourteenth Amendments do not guarantee the public a right
of access to information generated or controlled by
government . . . .”); see also McBurney v. Young, 569 U.S.
221, 232 (2013) (“[The Supreme] Court has repeatedly made
clear that there is no constitutional right to obtain all the
information provided by [a state’s freedom-of-information]
laws.”).6 As Chief Justice Burger’s opinion in Houchins

6
Of course, we have recognized a qualified First Amendment right
of access to certain government-controlled judicial records. We recently
explained in First Amendment Coalition of Arizona, Inc. v. Ryan, 938 F.3d
1069 (9th Cir. 2019):

As the Supreme Court originally conceptualized it, the
First Amendment right of access to governmental
proceedings refers to the right of the public to attend
and observe those proceedings. In the initial cases
recognizing the right, the Court held that the public has
the right to attend criminal trials, the jury-selection
process, and preliminary hearings. In situations in
which other interests justify the closure of a proceeding,
the Court held that the public has a right to access a
transcript of the proceeding within a reasonable time.
Our court has since extended the right of access to
various documents filed in criminal proceedings. For
 BOARDMAN V. INSLEE 19

explains, the disclosure of government-controlled information
is a “task which the Constitution has left to the political
processes,” such that “a legislative body might appropriately
resolve one way or the other” whether to provide public
access to information within its control. 438 U.S. at 12;
accord Ctr. for Nat’l Sec. Studies v. U.S. Dep’t of Justice, 331
F.3d 918, 934 (D.C. Cir. 2003) (“[D]isclosure of government
information generally is left to the ‘political forces’ that
govern a democratic republic.” (quoting Houchins, 438 U.S.
at 15)); Capital Cities Media, Inc. v. Chester, 797 F.2d 1164,
1167 (3d Cir. 1986) (en banc) (“The founding fathers
intended affirmative rights of access to government-held
information, other than those expressly conferred by the
Constitution, to depend upon political decisions made by the
people and their elected representatives.”).

example, we have held that the public has the right to
access plea agreements, documents filed in pretrial
proceedings, and documents filed in post-conviction
proceedings.

Id. at 1078 (citations omitted); see also Courthouse News Serv. v. Planet,
947 F.3d 581, 585 (9th Cir. 2020) (deciding that “the press has a qualified
[First Amendment] right of timely access to newly filed civil
nonconfidential complaints”). “To determine whether a First Amendment
right of access attaches to a type of judicial proceeding or record,” we
have applied the two-part test from Press–Enterprise Co. v. Superior
Court, 478 U.S. 1 (1986) (Press–Enterprise II), in which we consider: “(1)
whether that proceeding or record ‘has historically been open to the press
and general public’ and (2) ‘whether public access plays a significant
positive role in the functioning of the particular governmental process in
question.’” Courthouse News Serv., 947 F.3d at 590 (alterations adopted)
(quoting Press-Enterprise II, 478 U.S. at 8). In the district court,
Appellants asserted a right of access to Provider Information under Press
Enterprise II’s two-part test. The district court rejected that argument, and
Appellants do not renew it on appeal.
20 BOARDMAN V. INSLEE

Appellants acknowledge (as they must) that they have no
First Amendment right of access to Provider Information and
that Washington lawmakers have the political prerogative to
“decide not to give out [this] information at all without
violating the First Amendment.” See L.A. Police Dep’t v.
United Reporting Publ’g Corp. (United Reporting), 528 U.S.
32, 40 (1999). However, Appellants argue that the First
Amendment constrains the government’s discretion to
selectively disclose government-controlled information,
namely, by forbidding the government from discriminating
among viewpoints in the provision of information. Because
the thrust of their claim is viewpoint discrimination,
Appellants argue that they have presented us with a
cognizable First Amendment claim.

After review of the relevant precedent, we disagree with
the State. Houchins does not control here. Certainly,
Appellants are without a general right of access to Provider
Information under the First Amendment—this much they
concede. But it is a separate question—and one that is not
addressed in Houchins—whether the government may
transgress the First Amendment by imposing viewpoint-based
conditions on access to government-controlled information or
by otherwise discriminating among viewpoints in the
provision of information within its control. This is the
gravamen of Appellants’ claim, and Houchins does not
resolve it.

The separate writings in United Reporting tease out this
distinction. In United Reporting, the Supreme Court reviewed
a California statute that imposed multiple conditions on
public access to arrestee address information, including that
the requester declare that the information would not be used
“to sell a product or service.” Id. at 34–35. The plaintiff (a
 BOARDMAN V. INSLEE 21

company that provided arrestees’ personal information to its
customers) brought a facial challenge to the statute under the
First and Fourteenth Amendments. Id. at 34. The Supreme
Court did not address whether the conditions on access to
arrestee address information were constitutionally
permissible, but held instead that the statute was not
susceptible to a facial challenge. Id. at 40–41. The Court
explained that “California could decide not to give out
arrestee information at all without violating the First
Amendment.” Id. at 40 (citing Houchins, 438 U.S. at 14).
Accordingly, “for purposes of facial invalidation,” the
California statute was “simply a law regulating access to
information in the hands of the police department.” Id.

Nonetheless, eight Justices in United Reporting joined
three separate writings, all of which acknowledged the critical
point that “restrictions on the disclosure of government-held
information” may, under certain circumstances, “transgress
the First Amendment.” See Sorrell v. IMS Health Inc., 564
U.S. 552, 569 (2011) (citing United Reporting, 528 U.S. at
41–42 (Scalia, J., joined by Thomas, J., concurring); id. at
42–44 (Ginsburg, J., joined by O’Connor, Souter, and Breyer,
JJ., concurring); id. at 44–48 (Stevens, J., joined by Kennedy,
J., dissenting)). Six justices specifically agreed that the
government may not impose viewpoint-based conditions on
the provision of government-controlled information without
raising constitutional concerns. See United Reporting, 528
U.S. at 42–44 (Ginsburg, J., concurring); id. at 44–48
(Stevens, J., dissenting).

In her concurrence, Justice Ginsburg (joined by Justices
O’Connor, Souter, and Breyer) agreed with the opinion of the
Court that “California could . . . constitutionally decide not to
give out arrestee address information at all.” Id. at 43
22 BOARDMAN V. INSLEE

(Ginsburg, J., concurring). However, she compared the
government’s “provision of address information [to] a kind of
subsidy to people who wish to speak to or about arrestees.”
Id. (Ginsburg, J., concurring). As such, “once a State decides
to make such a benefit available to the public, there are no
doubt limits to its freedom to decide how that benefit will be
distributed.” Id. (Ginsburg, J., concurring). For example,
“California could not . . . release address information only to
those whose political views were in line with the party in
power,” for surely a state “could not justify limited
disclosures that discriminated on the basis of viewpoint or
some other proscribed criterion.” Id. at 43–44 (Ginsburg, J.,
concurring) (citing Bd. of Cty. Comm’rs v. Umbehr, 518 U.S.
668 (1996)). But as long as “the award of the subsidy is not
based on an illegitimate criterion such as viewpoint,” a state
“is free to support some speech without supporting other
speech.” Id. at 43 (Ginsburg, J., concurring) (citing Regan v.
Taxation With Representation of Wash., 461 U.S. 540
(1983)).

In dissent, Justice Stevens (joined by Justice Kennedy)
reiterated the rule that California was free to completely deny
access to arrestee address information without violating the
First Amendment. Id. at 45 (Stevens, J., dissenting). He also
thought “it equally clear that California could release the
information on a selective basis to a limited group of users
who have a special, and legitimate, need for the information.”
Id. (Stevens, J., dissenting). But he agreed with Justice
Ginsburg that the authority to “withhold [government-
controlled] information from all persons does not insulate [a
state’s] actions from constitutional scrutiny.” Id. at 47
(Stevens, J., dissenting). “[W]hen the State makes
information generally available, but denies access to a small
disfavored class . . . based on their viewpoint, or political
 BOARDMAN V. INSLEE 23

affiliation, for example, the discrimination would clearly be
invalid” and “obviously unconstitutional.” Id. at 45–46
(Stevens, J., dissenting). “For even though government may
withhold a particular benefit entirely, it ‘may not deny a
benefit to a person on a basis that infringes his
constitutionally protected interests—especially his interest in
freedom of speech.’” Id. at 47–48 (Stevens, J., dissenting)
(quoting Perry v. Sindermann, 408 U.S. 593, 597 (1972)). In
Justice Stevens’s view, the California statute had denied
access to the publishing company “based on the fact that [the
company planned] to publish the information to others who,
in turn, intended to use it for a commercial speech purpose
that the State [found] objectionable.” Id. at 46 (Stevens, J.,
dissenting). Accordingly, “because the State’s discrimination
[was] based on its desire to prevent the information from
being used for constitutionally protected purposes,” he would
have required California to “justify[] its conduct.” Id.
(Stevens, J., dissenting).

Of course, we are not bound by the separate writings in
United Reporting. Yet, we find the opinions of the six
Justices expressed therein to be persuasive and are therefore
disinclined to “blandly shrug them off” simply because they
did not comprise the holding of the Court. United States v.
Montero-Camargo, 208 F.3d 1122, 1132 n.17 (9th Cir. 2000)
(en banc) (quoting Zal v. Steppe, 968 F.2d 924, 935 (9th Cir.
1992) (Noonan, J., concurring and dissenting)). These
writings illustrate the limited scope of Houchins. Although
the decision whether to disclose government-controlled
information “at all” is well within the prerogatives of the
political branches, United Reporting, 528 U.S. at 40; see also
Houchins, 438 U.S. at 12, when the government selectively
discloses information within its control, a First Amendment
claim will lie if the government denies access to information
24 BOARDMAN V. INSLEE

“based on an illegitimate criterion such as viewpoint,” United
Reporting, 528 U.S. at 43 (Ginsburg, J., concurring); see also
id. at 45–47 (Stevens, J., dissenting). In the view of six
Justices in United Reporting, a contrary conclusion would be
irreconcilable with the constitutional precept that “the
Government may not deny a benefit to a person on a basis
that infringes his constitutionally protected freedom of speech
even if he has no entitlement to that benefit.” See Agency for
Int’l Dev. v. All. for Open Soc’y Int’l, Inc., 570 U.S. 205, 214
(2013) (alteration adopted and internal quotation marks
omitted) (quoting Rumsfeld v. Forum for Acad. and
Institutional Rights, Inc., 547 U.S. 47, 59 (2006)); see also
Perry, 408 U.S. at 597 (“[E]ven though a person has no
‘right’ to a valuable governmental benefit and even though
the government may deny him the benefit for any number of
reasons, there are some reasons upon which the government
may not rely. It may not deny a benefit to a person on a basis
that infringes his constitutionally protected interests—
especially, his interest in freedom of speech.”). As in other
areas where the legislature enjoys broad discretion in
deciding whether and how to confer a benefit or subsidy, the
government is not insulated from First Amendment scrutiny
when it discriminates invidiously in the provision of
government-controlled information. See, e.g., Agency for Int’l
Dev., 570 U.S. at 221 (striking down a regulation because it
“compel[led] as a condition of federal funding the affirmation
of a belief that by its nature [could not] be confined within the
scope of the Government program”); Nat’l Endowment of the
Arts v. Finley, 524 U.S. 569, 587 (1998) (“If the
[government] were to leverage its power to award subsidies
on the basis of subjective criteria into a penalty on disfavored
viewpoints, . . . [or] if a subsidy were ‘manipulated’ to have
a ‘coercive effect,’ then relief could be appropriate.” (quoting
Ark. Writers’ Project, Inc. v. Ragland, 481 U.S. 221, 237
 BOARDMAN V. INSLEE 25

(1987) (Scalia, J., dissenting))); Leathers v. Medlock, 499
U.S. 439, 450 (1991) (“[A] tax scheme that discriminates
among speakers does not implicate the First Amendment
unless it discriminates on the basis of ideas.” (citing Regan,
461 U.S. at 540)).

Two of our sister circuits have come to similar
conclusions in finding that speaker-based or content-based
restrictions on access to government-controlled information
are “susceptible to a First Amendment challenge.” Fusaro v.
Cogan, 930 F.3d 241, 255 (4th Cir. 2019); accord Lanphere
& Urbaniak v. Colorado, 21 F.3d 1508 (10th Cir. 1994). In
Fusaro, the Fourth Circuit reviewed a Maryland statute that
imposed two conditions on access to the state’s voter-
registration list: (1) the requester must be a registered
Maryland voter; and (2) the information may only be used for
purposes “related to the electoral process.” Id. at 244. The
plaintiff—a resident and registered voter of Virginia—was
denied access to Maryland’s voter-registration list. Id. at
245–46. He then challenged the Maryland statute in federal
court, alleging that it burdened speech in violation of the First
Amendment. Id. at 248. Relying on Houchins, the district
court dismissed the plaintiff’s complaint on the grounds that
he had no First Amendment right of access to the voter-
registration list. Id. The Fourth Circuit reversed. Unlike the
district court, the court viewed the plaintiff’s claim as “a free
speech challenge to conditions that a state has imposed on the
release of voter registration data,” id. at 256, not a right-of-
access claim (which would have been foreclosed by
Houchins), id. at 249–50. Therefore, although “neither the
Supreme Court in Houchins nor any appellate court applying
that decision ha[d] been faced with a situation where the
government provided information only to a discrete group for
limited purposes,” id. at 253, the court held that the plaintiff’s
26 BOARDMAN V. INSLEE

claim was “cognizable under the First Amendment,” id. at
249. The court found that the separate writings in United
Reporting “strongly signaled that . . . some conditions on the
disclosure of government information can run afoul of the
Free Speech Clause, giving rise to a viable constitutional
claim.” Id. at 253. Moreover, the court reasoned that
restrictions based on the identity of a speaker or the content
of a speaker’s message like those imposed by the Maryland
statute “are frequently deemed to be constitutionally suspect”
in other contexts. Id. at 252. Accordingly, the court concluded
that “a First Amendment claim that challenges suspect
conditions on access to government information must be
available, at least where the plaintiff alleges circumstances
indicating improper interference with protected speech.” Id.
at 255.

In Lanphere & Urbaniak, the Tenth Circuit considered a
First Amendment challenge to a Colorado statute that
permitted “public access to criminal justice and official action
records” only if the requester signed a statement affirming
that the records would not be used “for the purpose of directly
soliciting business for pecuniary gain.” 21 F.3d at 1510–11.
The plaintiffs argued that the Colorado statute imposed
content-based conditions on access to government records by
denying access to those who wished to use the records to
engage in commercial speech, thus triggering First
Amendment scrutiny. Id. at 1511. On the other hand, the
defendants (like the State in this case) contended that the
plaintiffs had brought “a simple access-to-records case” that
did not raise any First Amendment issues. Id. The Tenth
Circuit agreed that the plaintiffs had no First Amendment
right of access to the government records under Houchins. Id.
at 1511–12. But it disagreed that its inquiry ended there. Id.
at 1512. Although Colorado had the authority to deny access
 BOARDMAN V. INSLEE 27

to the records entirely, it instead “disallow[ed] the release of
records to those wishing to use them for commercial speech,
while allowing the release of the same records to those having
a noncommercial purpose.” Id. at 1512–13. Because this
regulatory scheme imposed “a content-based restriction on
protected speech,” the court concluded that the Colorado
statute did “in fact implicate the First Amendment” and was
subject to the appropriate level of scrutiny. Id. at 1513.

Here, Appellants challenge a law that regulates the
disclosure of government-controlled information. Under
Houchins, they have no First Amendment right of access to
this information. 438 U.S. at 15. But like the plaintiffs in
Fusaro and Lanphere & Urbaniak, Appellants do not rely on
a right-of-access theory. Instead, Appellants argue that
Washington law denies them access to Provider Information
because of the views they express on a particular topic. The
State’s argument that “laws restricting public access to
records do not implicate the First Amendment” relies on a
misapplication of Houchins, which does not stand for the
proposition that invidious viewpoint discrimination in the
provision of government-controlled information is beyond
constitutional scrutiny. We therefore agree with the Fourth
Circuit and Tenth Circuit that “a First Amendment claim . . .
must be available” in this context, Fusaro, 930 F.3d at 255;
accord Lanphere & Urbaniak, 21 F.3d at 1512–13, lest the
government be permitted to “discriminate invidiously . . . in
such a way as to aim at the suppression of dangerous ideas,”
Regan, 461 U.S. at 548 (alteration adopted and internal
quotations marks omitted) (quoting Cammarano v. United
States, 358 U.S. 498, 513 (1959)). Accordingly, we turn to
the merits of Appellants’ claim.
28 BOARDMAN V. INSLEE

B

Appellants articulate two discernable arguments in
support of their claim that Revised Code of Washington
sections 42.56.640 and 43.17.410—which codify Initiative
1501’s prohibition on public access to the sensitive personal
information of vulnerable individuals and their in-home care
providers—deny them access to Provider Information based
on the views they espouse on the subject of collective
bargaining. Neither argument is supported by the text or
operation of the challenged provisions.

Appellants first invoke what is known as the
“unconstitutional conditions” doctrine by asserting that
sections 42.56.640 and 43.17.410 condition receipt of a
government benefit—Provider Information—on a requester’s
views, such that Appellants are denied access to Provider
Information because they do not espouse “a State-preferred
view” on collective bargaining. In the First Amendment
context, this doctrine provides that “the Government may not
deny a benefit to a person on a basis that infringes his
constitutionally protected freedom of speech even if he has no
entitlement to that benefit,” Agency for Int’l Dev., 570 U.S.
at 214 (alteration adopted and internal quotation marks
omitted) (quoting Rumsfeld, 547 U.S. at 59). However, the
challenged provisions do not operate as Appellants suggest,
for they are completely “silent . . . concerning any speaker’s
point of view.” See Members of the City Council of the City
of L.A. v. Taxpayers for Vincent, 466 U.S. 789, 804 (1984)
(emphasis added). First, section 43.17.410 states:

To protect vulnerable individuals and their
children from identity crimes and other forms
of victimization, neither the state nor any of
 BOARDMAN V. INSLEE 29

its agencies shall release sensitive personal
information of vulnerable individuals or
sensitive personal information of in-home
caregivers for vulnerable populations, as those
terms are defined in [Wash. Rev. Code section
42.56.640].

Second, section 42.56.640 provides: “Sensitive personal
information of vulnerable individuals and sensitive personal
information of in-home caregivers for vulnerable populations
is exempt from inspection and copying under [the PRA].” By
their terms, these provisions deny access to Provider
Information irrespective of a requester’s views, ideology, or
message on any particular subject. The implication of
Appellants’ argument is that they could adopt a different
viewpoint regarding collective bargaining in order to obtain
Provider Information. But this is not so. Under the plain
language of sections 42.56.640 and 43.17.410, pro-collective-
bargaining voices and anti-collective-bargaining voices (and
all voices, for that matter) are denied access to Provider
Information unless the information is requested under one of
several narrow circumstances. Therefore, Appellants’
argument that sections 42.56.640 and 43.17.410 condition
access to Provider Information on the views of the requester
is simply unfounded.

Nevertheless, Appellants contend further that the statutory
provisions codifying Initiative 1501 discriminate among
viewpoints, because they permit disclosure of Provider
Information to the Unions (who convey a certain message on
the subject of collective bargaining), while denying equal
access to Appellants (who convey a competing message). We
disagree. “A regulation engages in viewpoint discrimination
when it regulates speech based on the specific motivating
30 BOARDMAN V. INSLEE

ideology or perspective of the speaker.” First Resort, Inc. v.
Herrera, 860 F.3d 1263, 1277 (9th Cir. 2017) (emphasis
added) (internal quotation marks omitted) (quoting Reed v.
Town of Gilbert, 576 U.S. 155, 168 (2015)); see also Reed,
576 U.S. at 163 (“Government regulation of speech is content
based if a law applies to particular speech because of the
topic discussed or the idea or message expressed.” (emphasis
added)). In other words, the government engages in viewpoint
discrimination when it “targets . . . particular views taken by
speakers on a subject.” Ctr. for Bio-Ethical Reform, Inc. v.
City and County of Honolulu, 455 F.3d 910, 921 (9th Cir.
2006) (emphasis added) (internal quotation marks omitted)
(quoting Brown v. Cal. Dep’t of Transp., 321 F.3d 1217, 1223
(9th Cir. 2003)). However, the challenged provisions do not
“draw[] distinctions based on the message[s]” conveyed by
either Appellants or the Unions. See Reed, 576 U.S. at 163
(emphasis added). Under Revised Code of Washington
section 42.56.645(1)(d), the Unions receive access to
Provider Information, because “[t]he information is being
provided to a representative certified or recognized under
[section] 41.56.080.” This provision does not permit the
Unions access to Provider Information based on the views
they espouse on the subject of collective bargaining. Rather,
the Unions’ current access to Provider Information is based
entirely on their legal status as certified exclusive bargaining
representatives under Washington law. See Wash. Rev. Code.
§ 41.56.080.7

7
The Supreme Court in Janus acknowledged that a union’s ability
to “obtain[] information about employees” was one of the many “benefits”
and “special privileges” of being the exclusive bargaining representative
of a collective bargaining unit. 138 S. Ct. at 2467.
 BOARDMAN V. INSLEE 31

The Supreme Court’s decision in Perry Education Ass’n
underscores this distinction. In Perry Education Ass’n, a
collective-bargaining agreement between a school district and
the teachers’ exclusive bargaining representative provided the
representative with exclusive access to the school district’s
“interschool mail system and teacher mailboxes.” 460 U.S. at
39–41. A rival teachers union, who was denied access to the
mail system and mailboxes, challenged the access policy
under the First Amendment. Id. at 41. The Supreme Court
upheld the policy, finding it to be a reasonable regulation of
speech in a nonpublic forum, because of “the special
responsibilities” associated with being the exclusive
bargaining representative for the school district’s teachers. Id.
at 50–55. In doing so, the Court rejected the contention that
the access policy “favor[ed] a particular viewpoint, that of the
[exclusive bargaining representative], on labor relations.” Id.
at 48–49. But there was “no indication that the school board
intended to discourage one viewpoint and advance another,”
and the Court thought it “more accurate to characterize the
access policy as based on the status of the respective unions
rather than their views.” Id. at 49.8

8
Our dissenting colleague suggests “four main reasons” that “Perry
does not save Initiative 1501.” Dissent at 75–79. However, none of the
four undermine Perry. First, the dissent fails to explain why it matters that
Perry “concerned access to physical government property, not access to
government-held information.” See id. at 75–76. Just as the state in Perry
had “the right to make distinctions in access on the basis of subject matter
and speaker identity,” id. at 75 (quoting Perry, 460 U.S. at 49), the State
in this case is free to subsidize “some speech without supporting other
speech” as long as “the award of the subsidy is not based on an
illegitimate criterion such as viewpoint.” United Reporting, 528 U.S. at 43
(Ginsburg, J., concurring). Thus, it does not matter that the Court in Perry
was dealing with a “nonpublic forum” and the “different considerations”
accompanying that context rather than a government subsidy of speech
like that at issue here. The bottom line: the State has the right to pick and
32 BOARDMAN V. INSLEE

choose which speech is subsidized so long as it does not discriminate on
the basis of viewpoint. See Leathers, 499 U.S. at 450.

As to the dissent’s second point, we join the Supreme Court in
refusing to find that “[Initiative 1501] is ‘viewpoint based’ simply because
its enactment was motivated by the conduct of the partisans on one side
of a debate.” Hill v. Colorado, 530 U.S. 703, 724 (2000); accord Frisky
v. Schultz, 487 U.S. 474, 482 (1988) (finding that an ordinance restricting
picketing was “content neutral” despite record evidence that it was
enacted to suppress the activities of antiabortion protestors).

The dissent’s third and fourth arguments—regarding the
“reasonableness” of Initiative 1501—also miss the mark. Perry dealt with
a government restriction on speech in a nonpublic forum. See 460 U.S. at
49. In that context, courts must ask two distinct questions: (1) whether the
restrictions “discriminate against speech on the basis of viewpoint” and
(2) whether the restrictions are “reasonable in light of the purpose served
by the forum.” Good News Club v. Milford Cent. Sch., 533 U.S. 98,
106–07 (2001) (quoting Cornelius v. NAACP Legal Def. & Educ. Fund,
Inc., 743 U.S. 788, 806 (1985)); accord Perry, 460 U.S. at 48–54
(applying a reasonableness analysis after determining that the regulation
did not constitute viewpoint discrimination); Cornelius, 473 U.S. at 806
(“Control over access to a nonpublic forum can be based on subject matter
and speaker identity so long as the distinctions drawn are reasonable in
light of the purpose served by the forum and are viewpoint neutral.”);
Rosenberger v. Rector & Visitors of Univ. of Va., 515 U.S. 819, 829–30
(1995) (“Once it has opened a limited forum, . . . [t]he State may not
exclude speech where its distinction is not ‘reasonable in light of the
purpose served by the forum, nor may it discriminate against speech on
the basis of its viewpoint.” (citation omitted) (quoting Cornelius, 473 U.S.
at 806)). However, while viewpoint neutrality is “a broadly applicable
requirement to all laws implicating First Amendment concerns with a test
that does not vary,” Wisc. Educ. Ass’n Council v. Walker, 705 F.3d 640,
650 n.8 (7th Cir. 2013) (emphasis added), the “reasonableness” inquiry
does not apply here, where we are examining the constitutionality of a
speech subsidy, see Leathers, 499 U.S. at 450 (stating that a speech
subsidy “that discriminates among speakers does not implicate the First
Amendment unless it discriminates on the basis of ideas”); see also United
Reporting, 528 U.S. at 43 (noting that as long as “the award of the subsidy
 BOARDMAN V. INSLEE 33

An example will demonstrate the propriety of applying
Perry Education Ass’n’s rationale in this case. Since 2016,
Thurber, Benn, and the Freedom Foundation have sought to
replace SEIU 925 with PNFCCA as the exclusive bargaining
representative of family child care providers. Appellants’
argument that Initiative 1501 discriminates between their
views and the views of the Unions begs the question: Who
would receive access to Provider Information if PNFCCA
became the new exclusive bargaining representative of family
child care providers? The answer is quite simple under the
plain terms of section 42.56.645(1)(d). First, because SEIU
925 would no longer be “a representative certified or
recognized under [section] 41.56.080,” it would assuredly be
denied access to Provider Information (unless it sought the
information under a different exception). See Wash. Rev.
Code § 42.56.645(1)(d). Second, as the newly certified
exclusive bargaining representative under section 41.56.080,
PNFCCA would have unconditional access to Provider
Information. See id. Under this scenario, the views of the two
competing unions would be irrelevant, because the sole factor
deciding who may lawfully access Provider Information
under section 42.56.645(1)(d) is whether a requester has
achieved the legal status of an exclusive bargaining
representative under section 41.56.080. 9 Therefore,

is not based on an illegitimate criterion such as viewpoint,” a state “is free
to support some speech without supporting other speech” (Ginsburg, J.,
concurring)). Therefore, the existence of reasonable alternatives plays no
role in the present analysis.
9
The dissent argues that “[l]egal status is merely a label that the law
creates,” Dissent at 72 (internal quotation marks omitted), and asserts that
our conclusion “that the law here distinguishes based on legal status . . .
is only to ask the viewpoint discrimination question, not to answer it,” id.
at 72 (internal quotation marks omitted). However, unlike the dissent’s
34 BOARDMAN V. INSLEE

Appellants’ argument that section 42.56.645(1)(d) currently
provides the Unions with access to Provider Information
based on the Unions’ collective-bargaining views finds no
support in the text or operation of the statute. 10

hypothetical examples, see id. at 72, there is no underlying ideological test
that must be met in order to receive the speech subsidy at issue in this
case. Indeed, as explained above, the Provider Information is awarded
based on status alone; the recipients’ adherence to a given viewpoint plays
no role in determining who receives it.

The dissent also argues that our approach somehow contravenes
Janus, and would even result in that case “com[ing] out the other way.”
See Dissent at 72–74. This is plainly incorrect. We do not dispute the
Supreme Court’s holding in Janus that union speech “in the context of
collective bargaining is of great public importance.” 138 S. Ct. at 2475.
However, the mere fact that “[u]nions have views on topics of public
concern,” Dissent at 73, does not transform Initiative 1501 into a
viewpoint-based regulation on speech. At bottom, the speech subsidies at
issue in this case are awarded based on status, without regard for
adherence to a specific viewpoint.
10
The dissent disagrees with our assessment, arguing that Initiative
1501 does discriminate on the basis of viewpoint because all unions “have
an obvious, intrinsic view on whether unions should be disempowered:
they understandably disagree with that position. Unions are, in other
words, ‘pro-union.’” Dissent at 67. Similarly, the dissent asserts that
Initiative 1501 discriminates in favor of those in power: “[i]ncumbents
seek to promote incumbency; those in power wish to retain it.” Id. at
69–69; accord id. at 67 (stating that Initiative 1501’s “speaker-based
distinction powerfully favors those views inherent to incumbent unions
while creating obstacles to speech for anyone with opposing views”). On
these bases, the dissent concludes that Initiative 1501 “necessarily
promot[es] the view that the incumbent Unions should stay in power.” Id.
at 69–70. However, it does not matter to our analysis whether a speech
subsidy happens to affect one particular viewpoint more than another.
Indeed, by the dissent’s logic, every selective speech subsidy could be
struck down for viewpoint discrimination. See Wisc. Educ. Ass’n Council,
705 F.3d at 648–49 (noting that this argument cannot be squared with the
 BOARDMAN V. INSLEE 35

We also reject Appellants’ suggestion that these
challenged provisions are viewpoint discriminatory simply
because they disadvantage Appellants’ message. “A facially
neutral statute . . . [with] a legitimate end is not
discriminatory simply because it affects some groups more
than others.” Interpipe Contracting, Inc. v. Becerra, 898 F.3d
879, 900 (9th Cir. 2018); see also Wisc. Educ. Ass’n Council,
705 F.3d at 650 (“That the benefits of [a] subsidy may fall
more heavily on groups with one particular viewpoint does
not transform a facially neutral statute into a discriminatory
one.”). Instead, “[a] regulation that serves purposes unrelated
to the content of expression is deemed neutral, even if it has
an incidental effect on some speakers or messages but not
others.” Christian Legal Soc’y v. Martinez, 561 U.S. 661, 695
(2010) (quoting Ward v. Rock Against Racism, 491 U.S. 781,
791 (1989)); see also Madsen v. Women’s Health Ctr., Inc.,
512 U.S. 753, 763 (1994) (“[T]he fact that the injunction
covered people with a particular viewpoint does not itself
render the injunction content or viewpoint based.”). As we
already explained, “a law affecting entities holding a
particular viewpoint is not viewpoint discriminatory unless it
targets those entities because of their viewpoint.” Interpipe
Contracting, 898 F.3d at 900; see also First Resort, 860 F.3d

Supreme Court’s analysis in Regan and “proves too much: if different
speakers necessarily espouse different viewpoints, then any selective
[speech subsidy] decision would violate the First Amendment as
viewpoint discriminatory”). The relevant question is simply whether
Initiative 1501 “regulates speech based on the specific motivating
ideology or perspective of the speaker.” First Resort, 860 F.3d at 1277
(emphasis added) (internal quotation marks omitted) (quoting Reed, 576
U.S. at 168). As explained above, Initiative 1501 does not tie access to the
State-held records to any particular viewpoint or ideology. Therefore, that
the Unions happen to hold certain viewpoints and also happen to receive
the benefit of this subsidy is irrelevant.
36 BOARDMAN V. INSLEE

at 1277–78 (finding a regulation viewpoint-neutral, even
though it applied only to entities holding a certain viewpoint,
because it did not regulate those entities “based on” those
views). Thus, the mere fact that the challenged provisions
“disproportionately impact” Appellants’ message “does not
transform [their] facially neutral language into an invidiously
discriminatory” regulation of government-controlled
information. Wisc. Educ. Ass’n Council, 705 F.3d at 651.11

In sum, Initiative 1501 did not impose viewpoint-based
conditions on the disclosure of Provider Information, nor did

11
Appellants also rely on Meyer v. Grant, 486 U.S. 414 (1988), to
argue that Initiative 1501 violated the First Amendment because it
“restrict[ed] access to the most effective, fundamental, and perhaps
economical avenue of political discourse, direct one-on-one
communication.” See id. at 424. But Meyer is inapposite to this case. In
Meyer, a Colorado statute prohibited proponents of an initiative from
paying petition circulators to help obtain the necessary signatures to meet
the requirements to place the initiative on the ballot. Id. at 416–17. The
Supreme Court found that the statute restricted the initiative proponents’
political expression by restricting an otherwise available method of
communication. Id. at 422–23. However, Initiative 1501 did not similarly
restrict the means by which Appellants may communicate their message.
As the district court found, Appellants “may canvass, hire paid canvassers,
distribute pamphlets, make speeches, advertise and hold meetings, picket,
or send mailers to distribute their speech.” Initiative 1501 simply refused
to subsidize Appellants’ efforts to identify its targeted audience with
Provider Information, to which Appellants have no entitlement. See
Houchins, 438 U.S. at 15. The Supreme Court has already “reject[ed] the
‘notion that First Amendment rights are somehow not fully realized unless
they are subsidized by the State.’” Regan, 461 U.S. at 546 (quoting
Cammarano, 358 U.S. at 515 (Douglas, J., concurring)). And although the
nature of in-home care providers’ work creates an obstacle for Appellants,
a state has no duty to remove obstacles “not of its own creation.” Id. at
549–50 (quoting Harris v. McRae, 448 U.S. 297, 316 (1980)).
 BOARDMAN V. INSLEE 37

it discriminate between the competing views of Appellants
and the Unions. We reject Appellants’ claim accordingly.12

IV

Boardman, Thurber, and Benn (“Individual Appellants”)
further claim that Initiative 1501 impaired their associational
rights. The First Amendment implicitly protects the freedom
of expressive association, which in turn “presupposes a
freedom not to associate.” Roberts v. U.S. Jaycees, 468 U.S.
609, 623 (1984) (emphasis added); accord Janus, 138 S. Ct.
at 2463 (“The right to eschew association for expressive
purposes is likewise protected.”). Compelled or mandatory
associations are therefore “permissible only when they serve
a ‘compelling state interest that cannot be achieved through
means significantly less restrictive of associational
freedoms.’” Knox v. Serv. Emps. Int’l Union, Local 1000, 567
U.S. 298, 310 (2012) (alterations adopted) (quoting Roberts,
468 U.S. at 623).

In the Individual Appellants’ view, they are hopelessly
stuck in a compelled association with the Unions. However,
the record reflects very little “association” between the
parties. The Individual Appellants are not members of the
Unions, nor do they pay agency fees to the Unions. The only
relationship between the parties arises from the Unions’

12
Appellants have waived any argument that Initiative 1501 enacted
content-neutral regulations of government-controlled information that
incidentally burden speech, because they failed to raise such an argument
in their opening brief. See Smith v. Marsh, 194 F.3d 1045, 1052 (9th Cir.
1999) (“[O]n appeal, arguments not raised by a party in its opening brief
are deemed waived.”).
38 BOARDMAN V. INSLEE

exclusive representation of the Individual Appellants in
collective bargaining.

As we recently held in Mentele v. Inslee, 916 F.3d 783
(9th Cir. 2019), this limited relationship does not implicate
the Individual Appellants’ associational rights. Id. at 789. In
Mentele, two family child care providers challenged Revised
Code of Washington section 41.56.028, which requires
exclusive collective bargaining representation for family
child care providers. Id. at 784–85. The plaintiffs argued “that
their First Amendment right to expressive association was
violated when Washington recognized SEIU [925] as the
exclusive bargaining representative for all childcare
providers.” Id. at 785. Relying on the Supreme Court’s
decision in Minnesota State Board of Community Colleges v.
Knight, 465 U.S. 271 (1984), we held “that Washington’s
authorization of an exclusive bargaining representative does
not infringe [a family child care provider’s] First Amendment
rights.” Mentele, 916 F.3d at 789.

In Knight, a Minnesota statute required the state to “meet
and confer” with its professional employees on matters
outside the scope of collective bargaining. 465 U.S. at
273–74. Under that statute, if professional employees in a
collective bargaining unit had selected an exclusive
representative for collective bargaining negotiations, that
representative would also serve as their exclusive
representative during the “meet and confer” process. Id.
Several faculty members of Minnesota’s community colleges
(who formed a single collective bargaining unit) challenged
the constitutionality of their exclusive representative’s role in
the “meet and confer” process. Id. at 278. In upholding the
statute, the Supreme Court concluded that the faculty
members’ “freedom to associate or not to associate with
 BOARDMAN V. INSLEE 39

whom they please . . . ha[d] not been impaired” by requiring
exclusive representation during the “meet and confer”
process. Id. at 288–89. The faculty members were “free to
form whatever advocacy groups they [would] like,” and were
“not required to become members” of the union who
represented them in the “meet and confer” process. Id. at 289.
Any “pressure” the faculty members may have felt to join the
union in order to have a voice in the “meet and confer”
process was “no different from the pressure they may [have
felt] to join [the union] because of its unique status” during
collective bargaining, “a status the Court . . . summarily
approved.” Id. at 289–90. Thus, the Supreme Court held that
the faculty members’ “associational freedom ha[d] been
wholly unimpaired” by the Minnesota statute. Id. at 290 n.12.

In Mentele, we found that Knight had addressed and
“approved the requirement that bound non-union dissenters
to exclusive union representation.” Mentele, 916 F.3d at 789.
In making this finding, we rejected the argument that Janus
overruled Knight sub-silentio by deciding that public-sector
unions could not constitutionally compel non-union members
to pay agency fees. Id.; see also Janus, 138 S. Ct. at 2486
(“States and public-sector unions may no longer extract
agency fees from nonconsenting employees. . . . This
procedure violates the First Amendment and cannot
continue.”). A plaintiff in Mentele had argued that we were
bound by a passage in Janus, see Mentele, 916 F.3d at 789,
which stated that the “require[ment] that a union serve as
exclusive bargaining agent for its employees . . . [was] itself
a significant impingement on associational freedoms,”
although it is an impingement that is “tolerated” in the
context of collective bargaining, Janus, 138 S. Ct. at 2478
(emphasis added). We held that this passage was mere dictum
that did not “indicat[e] that the Court intended to revise the
40 BOARDMAN V. INSLEE

analytical underpinnings of Knight or otherwise reset the
longstanding rules governing the permissibility of mandatory
exclusive representation.” Mentele, 916 F.3d at 789.13

In light of Knight and Mentele, the Individual Appellants’
relationship with the Unions raises no First Amendment
concerns. Nevertheless, the Individual Appellants argue that
Initiative 1501 impaired their freedom of association,
regardless of the constitutional propriety of exclusive
bargaining representation. They claim that Initiative 1501 has
effectively disabled them from removing the Unions as their
exclusive bargaining representatives by preventing the

13
Our sister circuits who have addressed the constitutional propriety
of exclusive collective bargaining representation on behalf of in-home
care providers have similarly found that Knight forecloses any freedom-
of-association challenge based on this representation. See D’Agostino v.
Baker, 812 F.3d 240, 243–44 (1st Cir. 2016) (“Since [the] non-union
professionals [in Knight] . . . could claim no violation of associational
rights by an exclusive bargaining agent speaking for their entire
bargaining unit when dealing with the state even outside collective
bargaining, the same understanding of the First Amendment should govern
the position taken by the family care providers here, whose objection goes
only to bargaining representation.”); Hill v. Serv. Emps. Int’l Union, 850
F.3d 861, 864–66 (7th Cir. 2017) (holding that Knight “foreclose[d]” the
argument that an Illinois statute requiring exclusive representation in
collective bargaining was “a mandatory association subject[] to heightened
scrutiny”); Bierman v. Dayton, 900 F.3d 570, 574 (8th Cir. 2018) (“On the
merits, the homecare providers contend that [a Minnesota statute] creates
a ‘mandatory agency relationship’ between them and the exclusive
representative that violates their right to free association under the First
and Fourteenth Amendments. This argument, however, is foreclosed by
Knight.”); Jarvis v. Cuomo, 660 F. App’x 72, 74 (2d Cir. 2016)
(unpublished) (“Plaintiffs contend that New York’s recognition of
defendant . . . as the exclusive bargaining representative for their
bargaining unit violates their First Amendment rights because it compels
union association. The argument is foreclosed by [Knight] . . . .”).
 BOARDMAN V. INSLEE 41

Individual Appellants from obtaining Provider Information.
They allege that use of Provider Information is “the only
feasible means” of building the requisite support among in-
home care providers to force an election for a new exclusive
bargaining representative. Without the assistance of Provider
Information, the Individual Appellants argue that they will be
compelled to associate with the Unions in “perpetuity.”

We are unpersuaded for several reasons. First, we must
start with the obvious: the Individual Appellants are not
entitled to Provider Information. See Houchins, 438 U.S. at
15. Therefore, by denying the Individual Appellants access to
this information, Initiative 1501 simply refused to subsidize
their efforts to replace the Unions. This is of no constitutional
concern. See Regan, 461 U.S. at 546 (“We again reject the
‘notion that First Amendment rights are somehow not fully
realized unless they are subsidized by the State.’” (quoting
Cammarano, 358 U.S. at 515 (Douglas, J., concurring))). As
we have already said, “although government may not place
obstacles in the path of a person’s exercise of freedom of
[association], it need not remove those not of its own
creation.” See Regan, 461 U.S. at 549–50 (alterations
adopted) (quoting Harris, 448 U.S. at 316). Second, even if
we were to characterize Initiative 1501 as placing a burden on
the Individual Appellants’ ability to remove the Unions as
exclusive bargaining representatives, we find no authority
(and Appellants cite to none) supporting the proposition that
the First Amendment requires the government to make it
simple enough for a collective bargaining unit to change its
exclusive representative—a relationship that does not itself
impair associational rights. See Mentele, 916 F.3d at 789.
And third, the Individual Appellants rely largely on dictum
from Janus that exclusive bargaining representation is “a
significant impingement on associational freedoms that
42 BOARDMAN V. INSLEE

would not be tolerated in other contexts.” 138 S. Ct. at 2478.
As we found in Mentele, this passage does not undermine
Knight’s conclusion that mandatory exclusive bargaining
representation in no way impinges First Amendment rights.
Mentele, 916 F.3d at 789. Thus, we conclude that Initiative
1501 did not implicate the Individual Appellants’
associational freedom, which has been left “wholly
unimpaired.” See Knight, 465 U.S. at 290 n.12.

V

We need only remark briefly on Appellants’ claim that
Initiative 1501 violates the First Amendment rights of other
in-home care providers by denying “the providers the right to
determine for themselves if they want to hear Appellants’
messages,” because Appellants do not have standing to assert
the rights of other in-home care providers.14

Our jurisdiction “is limited to actual cases and
controversies.” Alaska Right to Life Political Action Comm.
v. Feldman, 504 F.3d 840, 848 (9th Cir. 2007) (internal
quotation marks omitted) (quoting Allen v. Wright, 468 U.S.
737, 750 (1984)). “One of the controlling elements in the
definition of a case or controversy under Article III is
standing.” Hein v. Freedom From Religion Found., Inc., 551
U.S. 587, 598 (2007) (alteration adopted and internal

14
After rejecting Appellants’ right-to-receive-information claim on
the merits, the district court noted that Appellants failed to include this
claim in their complaint. Because the issue was “raised sufficiently for the
trial court to rule on it,” it has been properly preserved for appeal. Yamada
v. Nobel Biocare Holding AG, 825 F.3d 536, 543 (9th Cir. 2016) (quoting
Whittaker Corp. v. Execuair Corp., 953 F.2d 510, 515 (9th Cir. 1992));
see also id. (“[W]hen a party takes a position and the district court rules
on it, there is no waiver.”).
 BOARDMAN V. INSLEE 43

quotation marks omitted) (quoting ASARCO Inc. v. Kadish,
490 U.S. 605, 613 (1989)). To establish standing, a plaintiff
must show that he “(1) suffered an injury in fact, (2) that is
fairly traceable to the challenged conduct of the defendant,
and (3) that is likely to be redressed by a favorable judicial
decision.” Spokeo, Inc. v. Robins, 136 S. Ct. 1540, 1547
(2016). “In addition to these Article III requirements of injury
in fact, causation, and redressibility, prudential standing
concerns require that we consider, for example, . . . whether
the plaintiff is asserting her own rights or the rights of third
parties . . . .” Alaska Right to Life Political Action Comm.,
504 F.3d at 848; see also Hollingsworth v. Perry, 570 U.S.
693, 708 (2013) (“It is, however, a ‘fundamental restriction
on our authority’ that ‘in the ordinary course, a litigant must
assert his or her own legal rights and interests, and cannot rest
a claim to relief on the legal rights or interests of third
parties.’” (alteration adopted) (quoting Powers v. Ohio, 499
U.S. 400, 410 (1991))).

Appellants claim that Initiative 1501 infringed other in-
home care providers’ First Amendment right to receive
information. Surely, the First Amendment “protects the right
to receive information and ideas.” Stanley v. Georgia, 394
U.S. 557, 564 (1969). However, Appellants do not “assert
their own legal rights,” but “those of third parties.” See
Johnson v. Stuart, 702 F.2d 193, 196 (9th Cir. 1983)
(emphasis added). They have therefore failed to establish
standing to bring this claim.

VI

Finally, we address Appellants’ claim that Initiative 1501
violated the Equal Protection Clause. “The Equal Protection
Clause of the Fourteenth Amendment commands that no State
44 BOARDMAN V. INSLEE

shall ‘deny to any person within its jurisdiction the equal
protection of the laws,’ which is essentially a direction that all
persons similarly situated should be treated alike.” Gallinger
v. Becerra, 898 F.3d 1012, 1016 (9th Cir. 2018) (quoting City
of Cleburne v. Cleburne Living Ctr., 473 U.S. 432, 439
(1985)).

“To prevail on [their] equal-protection claim, [Appellants]
‘must [first] show that a class that is similarly situated has
been treated disparately.’” Roy v. Barr, 960 F.3d 1175, 1181
(9th Cir. 2020) (quoting Ariz. Dream Act Coal. v. Brewer,
855 F.3d 957, 966 (9th Cir. 2017)). “[O]ur first step is to
identify the state’s classification of groups.” Gallinger, 898
F.3d at 1016 (alteration adopted and internal quotation marks
omitted) (quoting Country Classic Dairies, Inc. v. Milk
Control Bureau, 847 F.2d 593, 596 (9th Cir. 1988)). Here,
Initiative 1501 exempted certified exclusive bargaining
representatives from the general prohibition on access to
Provider Information. See Wash. Rev. Code
§ 42.56.645(1)(d). Having “identified a classified group, we
look for a control group composed of individuals who are
similarly situated to those in the classified group in respects
that are relevant to the state’s challenged policy.” Gallinger,
898 F.3d at 1016 (citations omitted). Appellants contend that
they are similarly situated to the Unions (who are currently
the certified exclusive bargaining representatives of in-home
care providers), because “both seek to utilize updated
provider lists to engage in protected speech with providers.”
The State contends that Appellants are in no way similarly
situated to an exclusive bargaining representative who is
required under Washington law to represent all members of
a collective bargaining unit in collective bargaining
negotiations. However, we need not resolve this dispute,
because Initiative 1501 satisfies the appropriate level of
 BOARDMAN V. INSLEE 45

scrutiny even if Appellants were similarly situated to this
classified group. See Roy, 960 F.3d at 1181 (“If the two
groups are similarly situated, we determine the appropriate
level of scrutiny and then apply it.” (quoting Gallinger, 898
F.3d at 1016)).

As we have already discussed, Appellants have failed to
demonstrate that Initiative 1501 has burdened their First
Amendment rights. Accordingly, the initiative “need not be
tested by the strict scrutiny applied when government action
impinges upon a fundamental right protected by the
Constitution.”15 Perry Educ. Ass’n, 460 U.S. at 54. Instead,
Initiative 1501 “‘need only rationally further a legitimate
state purpose’ to be valid under the Equal Protection Clause.”
Knight, 465 U.S. at 291 (quoting Perry Educ. Ass’n, 460 U.S.
at 54); accord Armour v. City of Indianapolis, 566 U.S. 673,
680 (2012). Under rational-basis review, “[a] statute is
presumed constitutional, and ‘the burden is on the one
attacking the legislative arrangement to negative every
conceivable basis which might support it,’ whether or not the
basis has a foundation in the record.” Heller, 509 U.S. at
320–21 (alteration adopted and citations omitted) (quoting
Lehnhausen v. Lake Shore Auto Parts Co., 410 U.S. 356, 364
(1973)). “The Equal Protection Clause is satisfied so long as
there is a plausible policy reason for [a] classification, . . . and
the relationship of the classification to its goal is not so
attenuated as to render the distinction arbitrary or irrational.”
Crawford v. Antonio B. Won Pat Int’l Airport Auth., 917 F.3d
1081, 1095–96 (9th Cir. 2019) 1095–96 (citations omitted)
(quoting Nordlinger v. Hahn, 505 U.S. 1, 11 (1992)). “Given
the standard of review, it should come as no surprise that

15
Appellants do not claim that Initiative 1501 “proceed[s] along
suspect lines.” See Heller v. Doe, 509 U.S. 312, 319 (1993).
46 BOARDMAN V. INSLEE

[courts] hardly ever strike[] down a policy as illegitimate
under rational basis scrutiny.” Trump v. Hawaii, 138 S. Ct.
2392, 2420 (2018).

The challenged provisions of Initiative 1501 satisfy
rational-basis review. First, the State has a legitimate interest
in protecting seniors and other vulnerable individuals—and
all of its residents, for that matter—from identity theft and
other financial crimes, and Washington voters “could have
rationally decided” that generally prohibiting public access to
the personal information of in-home care providers—many of
whom work within the homes of their clients—“would further
that interest.” Crawford, 917 F.3d at 1095 (emphasis
removed) (quoting Johnson v. Rancho Santiago Cmty. Coll.
Dist., 623 F.3d 1011, 1031 (9th Cir. 2010)). Second,
Washington voters could have rationally decided that
providing the exclusive bargaining representatives of in-home
care providers with access to Provider Information would
further the “legitimate state [interest] . . . in the special
responsibilities of an exclusive bargaining representative.”
See Perry Educ. Ass’n, 460 U.S. at 54.16

16
Appellants argue that Initiative 1501 has no rational basis for
prohibiting public access to the personal information of family child care
providers in particular, because these providers do not care for “vulnerable
individuals” (as they are defined under Initiative 1501). However, this
argument was not presented to the district court. Because Appellants
“fail[] to address any of the exceptions to the general rule that an argument
raised for the first time on appeal is waived,” we decline to address their
belated argument. See Club One Casino, Inc. v. Bernhardt, 959 F.3d 1142,
1153 (9th Cir. 2020)). “The district court is not merely a way station
through which parties pass by arguing one issue while holding back . . .
others for appeal.” Crawford v. Lungren, 96 F.3d 380, 389 n.6 (9th Cir.
1996).
 BOARDMAN V. INSLEE 47

Nevertheless, Appellants argue that Initiative 1501 fails
rational-basis review, because it was motivated by animus
towards Appellants and their opposition to the Unions.
Certainly, “a bare desire to harm a politically unpopular
group cannot constitute a legitimate governmental interest”
under the Equal Protection Clause. Romer v. Evans, 517 U.S.
620, 634 (1996) (alterations adopted) (quoting U.S. Dep’t of
Ag. v. Moreno (Moreno), 413 U.S. 528, 534 (1973)); accord
City of Cleburne, 473 U.S. at 446–47. But where the affected
parties are not members of a “traditionally suspect class” (i.e.,
classes defined by race or national origin), we “may strike
down” a challenged law on the basis of animus only “if the
statute serves no legitimate governmental purpose and if
impermissible animus toward an unpopular group prompted
the statute’s enactment.’” Animal Legal Def. Fund v. Wasden
(Wasden), 878 F.3d 1184, 1200 (9th Cir. 2018) (quoting
Mountain Water Co. v. Mont. Dep’t of Pub. Serv. Regulation,
919 F.2d 593, 598 (9th Cir. 1990)); accord Gallinger, 898
F.3d at 1021. We have already found that Initiative 1501
serves legitimate state interests. That “conclusion, on its own,
prevents [Appellants] from succeeding on their Equal
Protection claim.” Gallinger, 898 F.3d at 1021; see also
Wasden, 878 F.3d at 1200 (upholding a state statute under the
Equal Protection Clause, because, although it was motivated
by animus, it served a legitimate state interest, and therefore
did “not rest exclusively on an ‘irrational prejudice’” (quoting
City of Cleburne, 473 U.S. at 450)).

That being said, we further reject Appellants’ contention
that Initiative 1501 was motivated by animus. A plaintiff
demonstrates animus by showing “that an ‘invidious
discriminatory purpose was a motivating factor’ in the
relevant decision.” Dep’t of Homeland Sec. v. Regents of the
Univ. of Cal., 140 S. Ct. 1891, 1915 (2020) (quoting
48 BOARDMAN V. INSLEE

Arlington Heights v. Metro. Hous. Dev. Corp., 429 U.S. 252,
266 (1977)). “Possible evidence includes disparate impact on
a particular group, ‘departures from the normal procedural
sequence,’ and ‘contemporary statements by members of the
decisionmaking body.’” Id. (quoting Arlington Heights, 429
U.S. at 266–68). As evidence that Initiative 1501 was
motivated by animus, Appellants present the Unions’
motivations in supporting the ballot measure. However, there
is no evidence in the record (and Appellants certainly cite to
none) indicating that the more than 2.2 million Washington
voters who voted in favor of Initiative 1501 were motivated
by “an irrational prejudice,” see City of Cleburne, 473 U.S. at
450, or “a bare desire to harm” Appellants or their message
against the Unions, see Romer, 517 U.S. at 634 (alteration
adopted) (quoting Moreno, 413 U.S. at 534). We refuse to
impute upon Washington voters the allegedly invidious
motivations of the Unions. Appellants’ claim of animus is
therefore unfounded.

AFFIRMED

BRESS, Circuit Judge, dissenting:

What if the State of Washington passed a law that gave
the reigning political party access to certain State-controlled,
speech-enabling information, but denied that information to
everyone else? It is hard to imagine anyone believing such a
law would be constitutional under the First Amendment. So
should it matter if the State enacted the same law, but instead
of giving the information to the incumbent political party, it
gave it to an incumbent public-sector union that serves as the
exclusive bargaining representative for certain employees
 BOARDMAN V. INSLEE 49

paid with public funds? That is what happened here when
Washington voters enacted I-1501. This ballot initiative gave
the incumbent unions access to critical identifying
information for difficult-to-locate State-paid home care
workers, but disallowed anyone else from obtaining it
through a public records request—including an organization
that sought the information to contact these State employees
and inform them they had a constitutional right not to pay
union dues. See Janus v. AFSCME, Council 31, 138 S. Ct.
2448 (2018); Harris v. Quinn, 134 S. Ct. 2618 (2014).

The text and operation of I-1501 and a troubling
documentary record demonstrate exactly what is going on
here: transparent viewpoint discrimination. The State is
effectively using an information embargo to promote the
inherently “pro-union” views of the incumbent unions, while
making it vastly more difficult for those with opposing
views—and particularly those with views opposite
unions—to reach their intended audience. The First
Amendment does not permit this. There are competing
perspectives in this country about the role and efficacy of
public-sector unions. And in Washington State, there are
divergent views as to whether the incumbent unions are best
serving the in-home care providers they represent. I-1501
gives government-held information to persons on only one
side of these important debates—the incumbent unions
themselves—while denying that information to everyone on
the other side. Persons who oppose public-sector unions
cannot get the information, nor can persons who wish to
replace the incumbent unions with a rival union. The State-
held information here is not otherwise available. And the
information at issue is critical: the very means by which each
side can feasibly locate the in-home care providers with
whom they wish to engage in core political speech about
50 BOARDMAN V. INSLEE

topics of vital concern—including whether public-sector
employees should stop paying union fees.

We should have recognized I-1501 for what it is: a
powerful deterrent to the open discourse of ideas that is the
hallmark of a free people. The First Amendment requires that
winners and losers in political debates win or lose on their
own accord, without government favoritism. Here, the whole
point of I-1501 was to block opposition to public-sector
unions, entrench the incumbent unions, and prevent the
exercise of First Amendment rights that Harris and Janus
validated. That the State’s effort has seemingly worked only
confirms the unacceptable viewpoint discrimination that is
afoot. The State cannot accomplish through a viewpoint-
discriminatory disclosure of government-controlled
information what it could never achieve through a direct
restriction on speech. And when the government stacks the
decks in a political debate through grossly uneven access to
information, courts should recognize that the free
marketplace of ideas has been unacceptably compromised.

Whatever informational advantage Washington could
properly give an incumbent union over others, Washington’s
extreme approach for in-home care providers reflects
impermissible viewpoint discrimination. I would have held
that I-1501 fails First Amendment scrutiny. I respectfully
dissent.

I

A

I-1501 is a response to recent Supreme Court decisions
recognizing the First Amendment rights of public-sector
 BOARDMAN V. INSLEE 51

employees. To understand how and why I-1501 came about
and the viewpoint discrimination that was its objective, it is
first necessary to understand underlying developments in
First Amendment law.

In Abood v. Detroit Board of Education, 431 U.S. 209
(1977), the Supreme Court had held that public-sector
employees who decline to join a public-sector union could,
consistent with the First Amendment, be required to pay
“agency fees” for that portion of union dues used for union
activities “germane to [the union’s] duties as [a] collective
bargaining representative.” Id. at 235. The justifications for
allowing these fees to be charged to objecting employees
were the promotion of “labor peace” and “the risk of ‘free
riders.’” Id. at 224.

Over time, Abood was criticized for requiring public-
sector employees to subsidize union activities with which
they disagreed, in asserted violation of the First
Amendment’s prohibition on compelled speech. See, e.g.,
Knox v. SEIU, Local 1000, 567 U.S. 298, 311 (2012). In
Harris v. Quinn, 134 S. Ct. 2618 (2014), the Supreme Court
considered whether state-paid home healthcare workers in
Illinois could be required to pay union agency fees. These
persons were not “full-fledged public employees” but were
only “deemed to be public employees solely for the purpose
of unionization and the collection of an agency fee.” Id. at
2627–28.

Harris explained at length how Abood rested on
“questionable foundations.” Id. at 2638. But rather than
overrule Abood, Harris held that Abood’s rationale did not
extend to the quasi-public employees at issue. Id.
Unconstrained by Abood, the Supreme Court concluded that
52 BOARDMAN V. INSLEE

Illinois’s requirement that state-funded home healthcare
workers pay agency fees violated the First Amendment. Id.
2639–44. Such a requirement, the Court held, ran afoul of the
“bedrock principle” that “no person in this country may be
compelled to subsidize speech by a third party that he or she
does not wish to support.” Id. at 2644.

Several years later, in Janus v. AFSCME, Council 31, 138
S. Ct. 2448 (2018), the Supreme Court overruled Abood. The
Court explained that “Abood was poorly reasoned” and failed
to account for the “[f]undamental free speech rights [that] are
at stake.” Id. at 2460. And in a point that has special
relevance to this case for reasons I will explain, the Supreme
Court squarely rejected the theory that the “union speech paid
for by agency fees” concerns only “private” matters and not
ones of “public concern.” Id. at 2474. Instead, Janus held
that union speech in collective bargaining is “overwhelmingly
of substantial public concern” because unions speak out on a
range of important political topics, and the wages and benefits
of public employees are themselves a matter of “great public
concern.” Id. at 2475, 2477. It therefore “violate[d] the free
speech of nonmembers [to] compel[] them to subsidize
private speech on matters of substantial public concern.” Id.
at 2460.

B

In Washington, there are two types of in-home care
providers: individual providers, who care for disabled adults,
and family child care providers, who provide child care for
low-income families. Wash. Rev. Code §§ 41.56.030(7),
74.39A.240(3) (Supp. 2020). Although hired by private
individuals and families, these care providers receive
compensation through government-funded programs. See id.
 BOARDMAN V. INSLEE 53

Washington’s in-home care providers are thus equivalent to
the workers in Harris: they are treated as public employees
only for purposes of collective bargaining. Id.
§§ 41.56.028(1), 74.39A.270(1) (2013 & Supp. 2020); see
also Harris, 134 S. Ct. at 2635. Washington’s Department of
Early Learning (“DEL”) (now the Department of Children,
Youth, and Families (“DCYF”)) and Department of Social
and Health Services (“DSHS”) administer the State’s
programs for family child care providers and individual
providers.

Uniquely relevant to this case, effectively communicating
with care providers is essentially impossible without
certain State-held information. Unlike typical public-sector
employees, Washington’s approximately 45,000 care
providers are geographically dispersed throughout the State
and are hard to identify. Most of them do not work in what
we would regard as typical workplaces, as they are often one
family member caring for another in the privacy of their
homes. They could be our neighbors, but we would have
little reason to know they were quasi-public employees. Care
providers do not gather in any centralized location or
typically share clients, supervisors, or any other contacts with
one another. Care providers also experience a high turnover
rate of 20% to 40% each year, so contact information quickly
becomes outdated. By statute, Washington care providers are
subject to statewide collective bargaining representation and
are required to be represented by one exclusive bargaining
representative. Wash. Rev. Code §§ 41.56.026, 41.56.028(1),
(2)(a), (b), 74.39A.270(2)(a) (2012 & Supp. 2020).
Currently, Service Employees International Union Local
775NW (“SEIU 775”) is the exclusive bargaining
representative for individual providers, while Service
Employees International Union Local 925 (“SEIU 925”) is
54 BOARDMAN V. INSLEE

the exclusive bargaining representative for family child care
providers. I will refer to these incumbent unions collectively
as “the Unions.”

If one wants to communicate with Washington’s in-home
care providers, it is first necessary to identify them. This task
is extremely difficult without information that the State
maintains. Specifically, DCYF and DSHS maintain lists of
the identities and contact information for in-home care
providers. The Unions use these lists regularly to
communicate with care providers through direct mail, email,
and door-to-door canvassing. The content of the Unions’
communications bears out the Supreme Court’s
understanding in Janus that public-sector unions engage in
speech on topics of public concern.

In their written communications to Washington’s care
providers, the Unions have urged care providers to vote on
certain measures related to their employment, such as ballot
initiatives concerning minimum wage and sick leave. But the
Unions have also encouraged care providers to vote for
certain candidates for public office, including United States
Senator and Governor of Washington. For example, the
Unions urged in-home care providers to vote for Governor
Jay Inslee because, among other things, he “led the way to
implement the Affordable Care Act and expand Medicaid
coverage.” The Unions in communications to care providers
have advocated other policy positions that are seemingly less
related to collective bargaining. This includes recommending
that care providers vote in favor of gun control measures and
against a proposal that the Unions described as “giv[ing] huge
tax breaks to corporations.” My point is not to comment on
any position the Unions may wish to take, but to emphasize
that the Unions are communicating with care providers about
 BOARDMAN V. INSLEE 55

matters of substantial public concern—i.e., core political
speech.

Before the passage of I-1501, other persons also relied on
the State-maintained lists, obtained through public record
requests, to communicate with in-home care providers.
Appellant Freedom Foundation is a non-profit organization
that, by its description, “seeks to advance individual liberty,
free enterprise, and limited, accountable government.” After
the Supreme Court decided Harris, the Foundation created an
initiative to inform in-home care providers that they were not
required to pay agency fees. The Foundation claims that
“[p]roviders are often grateful to the Foundation because they
were previously unaware they were unionized, unaware of
their rights, and had unwittingly paid hundreds of dollars to
the Unions each year.” The Foundation used the lists
obtained through public records requests to identify and
communicate with care providers.

Appellant Bradley Boardman is an individual provider
who cares for his disabled sister-in-law. Once a member of
SEIU 775, Boardman left the union due to its “heavy
involvement in partisan politics.” After Harris, he exercised
his right to stop paying union fees and authored a letter to
other individual providers explaining how they could do the
same. Appellants Sharon Benn and Deborah Thurber are
family child care providers. They stopped paying union fees
to SEIU 925 after Harris and now seek to replace the union
with a different union by triggering an election, which
requires signatures from 30% of their fellow bargaining unit
members. Wash. Rev. Code § 41.56.070 (Supp. 2020);
Wash. Admin. Code § 391-25-110(1) (2009). Benn and
Thurber disagree with how SEIU 925 has represented its
constituents. Thurber explains that SEIU 925 “has not
56 BOARDMAN V. INSLEE

adequately represented the interests of Childcare Providers.”
She wants a new union installed that “would be fully
voluntary, fight for higher wages and insurance benefits, and
oppose unreasonable regulations.” Both Benn and Thurber
used public records requests to obtain contact information for
their fellow child care providers and send them information
about their policy positions.

In 2014, the Freedom Foundation submitted a records
request to DEL and obtained identifying information for care
providers that the Foundation then used to contact them. The
Foundation asserts that its efforts to educate in-home care
providers about the Supreme Court’s decision in Harris led
to a dramatic drop in union membership. As of January 2017,
63.2% of family child care providers are reported to have left
SEIU 925 post-Harris. As of April 2018, that number
reportedly had climbed to 65.5%.

Alarmed at these developments, the Unions quickly
moved to block them. Between 2014 and 2016, the Unions
repeatedly sued the Freedom Foundation, Boardman, and
Benn to prevent the release of the information. These
lawsuits for the most part failed or were dismissed, although
at least in some instances, the care provider lists were
outdated by the time Appellants finally received them.

C

With Appellants gaining traction in the courts, the Unions
lobbied the Washington legislature to make amendments to
the Public Records Act. When those efforts did not pan out,
the Unions turned to Washington’s ballot initiative process.
SEIU 775’s Secretary-Treasurer, Adam Glickman, founded
and chaired a political action committee called “Campaign to
 BOARDMAN V. INSLEE 57

Prevent Fraud and Protect Seniors.” The Unions formed the
Campaign to advocate for I-1501’s inclusion on
Washington’s 2016 ballot. Virtually all the funding for the
ballot initiative (more than $2 million) came from the Unions.

As relevant here, I-1501 amends Washington’s Public
Records Act to bar the release of care providers’ “sensitive
personal information.” Wash. Rev. Code § 43.17.410(1)
(2018). “Sensitive personal information” includes care
providers’ names, addresses, and “other personally
identifying information.” Id. § 42.56.640(2)(b). Importantly,
the incumbent Unions were specifically exempted from these
restrictions: the State can provide this information “to a
representative certified or recognized under RCW
41.56.080,” id. § 42.56.645(1)(d), a reference to the “[t]he
bargaining representative which has been determined to
represent a majority of the employees in a bargaining unit,”
id. § 41.56.080 (Supp. 2020). Here, those bargaining
representatives are SEIU 775 and SEIU 925.1

1
State agencies and State contracting parties may also obtain care
providers’ “sensitive personal information.” See Wash. Rev. Code
§ 42.56.645(1)(a), (f), (g) (2018). Other provisions allow targeted
disclosures for information about individual in-home care providers, in
some cases providers whose identities the requestor would already know.
See, e.g., id. § 42.56.645(1)(b) (permitting disclosure of information
concerning “individuals who have been accused of or disciplined for
abuse, neglect, exploitation, abandonment, or other acts involving the
victimization of individuals or other professional misconduct”); id.
§ 42.56.645(1)(c) (permitting disclosure when “[t]he information is being
released as part of a judicial or quasi-judicial proceeding and subject to a
court’s order protecting the confidentiality of the information and allowing
it to be used solely in that proceeding”). There is no suggestion that these
various provisions solve the constitutional problem here, and they do not.
58 BOARDMAN V. INSLEE

I-1501 was ostensibly based on protecting in-home care
providers and the persons for whom they care from identity
theft and fraud. The “Argument For” I-1501 in Washington’s
2016 “Voter’s Guide” urged passage of I-1501 to prevent
“criminals” from “steal[ing] an identity, causing emotional
distress, devastating personal finances[,] and ruining credit.”
As the district court in this case acknowledged, the
“evidence” supporting these claimed harms is “thin.” Indeed,
the district court went on, one could instead “rationally infer
that the predominate motivating factor for the Initiative and
the Campaign’s support for the Initiative was animus toward
the Freedom Foundation and outside entities with
prerogatives similar to the Foundation.” As the district court
explained, Appellants’ “most compelling argument may be
that the true, or at least primary motivation of the Initiative’s
drafters and promoters was to restrict the Foundation’s ability
to communicate with caregivers about their right to withhold
financial support from the unions.”

Extensive record evidence supports the district court’s
observations:

• Literature from the Campaign promoting I-1501
stated that “[g]roups like the Freedom Foundation are
threatening unions. They tell us to stop paying
dues—but that would weaken our unions and rollback
what we’ve won, like raises, health care, retirement,
and paid time off. Initiative 1501 will keep groups
like the Freedom Foundation from getting our
personal information. Vote yes on I-1501 to keep our
unions strong and protect what we’ve fought for.”

• A mass letter to SEIU 775 members ostensibly from
another member, but on SEIU letterhead, warned that
 BOARDMAN V. INSLEE 59

“[t]he Freedom Foundation is at it again” and is “bent
on tearing down everything we’ve won.” It criticized
the Freedom Foundation’s “anti-union agenda” and
the Foundation’s efforts to “weaken[] the union we’ve
worked so hard to build.” The letter warned members
that with the Union’s then-recent litigation losses over
the State-held lists, the Foundation may start
contacting them again, but that care providers should
tell the Foundation “NO WAY am I leaving the union
that’s done so much for caregivers.” The letter
concluded: “There’s one more way you can fight to
stop the Freedom Foundation: When you get your
ballot in the mail, vote YES on I-1501, which protects
the private information of caregivers and our state’s
most vulnerable.”

• In correspondence to a Seattle newspaper, Adam
Glickman, the Campaign’s chairman, argued in
favor of I-1501 because it would prevent care
providers’ information from being “made available
to the Freedom Foundation or any other
advocacy/political/religious group with an agenda.”

The Unions expressly urged their members to vote for I-
1501 because it would shut down the Foundation’s advocacy.
One SEIU 775 email to members explained that “[b]y voting
Yes [on I-1501] we protect caregivers in our union from anti-
union bullying of the Freedom Foundation.” Another Union
communication to members stated: “Groups like the Freedom
Foundation are threatening our union. They tell us to stop
paying dues—but that would weaken our union . . . . A vote
for I-1501 is a vote to protect our union, ourselves and our
clients.” Fuse Vote, an advocacy organization supported by
the Unions, likewise urged passage of I-1501 to end the
60 BOARDMAN V. INSLEE

Foundation’s advocacy efforts. It warned that the “right-
wing” Foundation had “been working to acquire the names
and contact information of home health care workers and
child care providers as part of a deceptive campaign to
destroy the unions, and this initiative would prevent them
from acquiring that private personal data from the state.”

Several internal communications to the office of
Governor Inslee produced in discovery in this case also
indicated that I-1501’s intended objective was to hinder anti-
union advocacy. One internal communication from the
Governor’s General Counsel to other staff in the Governor’s
office explained that “[I-1501] ostensibly deals with ID theft,
but is aimed at preventing the state from releasing public
records.” Another internal email from the Governor’s office
reveals that SEIU 775 provided the Governor’s office with
draft remarks to deliver at an SEIU 775 town hall in support
of I-1501. These draft remarks urged the Governor to say
that “[w]hat the Freedom Foundation is doing is wrong,” that
“the Freedom Foundation is just spreading lies and garbage
about your union,” and that the Governor would “work to
make sure that groups like the Freedom Foundation do not get
access to lists that they are not entitled to.”

This same document contains a draft script for Adam
Glickman, the Campaign’s chairman. In the script, after a
union vice president says that members should vote for I-
1501 because the Freedom Foundation “is trying to take away
everything we’ve won,” Glickman’s draft remarks urge
passage of I-1501 because “the Freedom Foundation has been
aggressively contacting our members, trying to convince
them to give up their union rights.” Notably, this portion of
the agenda is entitled: “1501 and the Freedom Foundation.”
 BOARDMAN V. INSLEE 61

Several publications took notice of I-1501’s evident non-
privacy objective and urged voters to vote against the
measure. For example, the Seattle Times Editorial Board
cautioned,

Don’t be fooled by I-1501’s pitch to close
scary loopholes and block the release of
records that enable identity theft. . . . Voters
should be aware that I-1501 is the result of a
spat between the powerful Service Employees
International Union and the conservative
Freedom Foundation. They are fighting over
whether the [F]oundation can contact state-
employed care providers to inform them that
they no longer are required to pay union dues
or fees to SEIU, following a U.S. Supreme
Court ruling in 2014.

After I-1501 passed, SEIU 775 sent a congratulatory
email to its members stating that the new law would
“protect[] caregivers from the Freedom Foundation or other
groups getting access to their personal information.” SEIU
925 similarly sent an email to its members touting I-1501’s
approval, decrying “extremist groups like the anti-union
Freedom Foundation.”

I now turn to the legal principles that render I-1501
unconstitutional.2

2
I agree with the majority that Appellants’ First Amendment
freedom of association claim fails. See Minn. State Bd. for Cmty. Colls.
v. Knight, 465 U.S. 271, 288 (1984) (holding that an exclusive bargaining
arrangement does not restrain members’ “freedom to associate or not to
associate with whom they please, including the exclusive representative”).
62 BOARDMAN V. INSLEE

II

The majority and I agree on how the analysis must begin:
a claim that the State has discriminated in access to
information based on viewpoint is a cognizable First
Amendment theory. Washington’s argument otherwise is
seriously mistaken, and the court today properly rejects it.

Eight Justices writing separately in Los Angeles Police
Department v. United Reporting Publishing Corp., 528 U.S.
32 (1999), made clear that the First Amendment would
forbid a State from conditioning access to government-held
information based on viewpoint. See id. at 42 (Scalia, J.,
joined by Thomas, J., concurring); id. at 43 (Ginsburg, J.,
joined by O’Connor, Souter, and Breyer, JJ., concurring); id.
at 46 (Stevens, J., joined by Kennedy, J., dissenting). As the
Supreme Court later explained, these separate writings all
“recognized that restrictions on the disclosure of government-
held information can facilitate or burden the expression of
potential recipients and so transgress the First Amendment.”
Sorrell v. IMS Health, Inc., 564 U.S. 552, 569 (2011); see
also Fusaro v. Cogan, 930 F.3d 241, 254 (4th Cir. 2019)
(“United Reporting provides affirmative support for a First
Amendment challenge to certain types of conditions being
placed on the dissemination of government information.”).

But see Thompson v. Marietta Educ. Ass’n, 972 F.3d 809, 814 (6th Cir.
2020) (explaining how “Knight’s reasoning conflicts with the reasoning
in Janus”). Appellants also purport to bring a freestanding claim based
on the alleged right of in-home care workers to receive information. My
preferred holding that I-1501 reflects viewpoint discrimination would
resolve that claim on its own. The critical First Amendment issue here,
and the one that fits the facts, is I-1501’s viewpoint-discriminatory denial
of information.
 BOARDMAN V. INSLEE 63

It is important to see why the First Amendment permits a
claim based on discriminatory access to government-
controlled information. Government discrimination against
speech based on viewpoint is the highest form of First
Amendment offense. Viewpoint discrimination is “an
egregious form of content discrimination” violating the
foundational principle that “[i]n the realm of private speech
or expression, government regulation may not favor one
speaker over another.” Rosenberger v. Rector and Visitors of
Univ. of Va., 515 U.S. 819, 828–29 (1995). A State’s
complete withholding of information would not present this
issue, because there is no general “First Amendment
guarantee of a right of access to all sources of information
within government control.” Houchins v. KQED, Inc., 438
U.S. 1, 9 (1978) (plurality opinion). Washington thus was
free not to make care provider information available to
anyone. United Reporting, 528 U.S. at 40 (citing Houchins,
438 U.S. at 14).

But the government’s selective disclosure of information
is different. Information facilitates speech. Sorrell, 564 U.S.
at 569. Not having information can burden speech or
effectively restrict it altogether. Id.; United Reporting, 528
U.S. at 42 (Scalia, J., concurring); id. at 47–48 (Stevens, J.,
dissenting). Because State-held information can be used to
enable speech, a State cannot withhold such information on
a viewpoint-discriminatory basis and thereby accomplish
through a restriction on information what it could never
achieve through a direct regulation of speech. Sorrell, 564
U.S. at 569.

Imagine, for example, that the government held certain
speech-enabling information and gave it only to members of
one political party. That would subsidize and promote one
64 BOARDMAN V. INSLEE

party’s views while muzzling the contrary views of other
parties. There is no sense in which that could be
constitutional. See United Reporting, 528 U.S. at 43
(Ginsburg, J., concurring) (“California could not, for
example, release address information only to those whose
political views were in line with the party in power.”).

Although the majority and I agree on the availability of
the First Amendment theory here, we disagree on whether I-
1501 reflects viewpoint discrimination. In my view, it plainly
does. This is a situation where a law “den[ying] access to
persons who wish to use the information for certain speech
purposes[] is in reality a restriction upon speech” that violates
the First Amendment. Id. at 42 (Scalia, J., concurring)
(emphasis omitted).

A

To see why this is so, we need to move a few analytical
pieces into place. The first is that I-1501 draws a distinction
based on the speaker: it gives the incumbent Unions unique
access to care provider information. The Supreme Court’s
“precedents are deeply skeptical of laws that ‘distinguish
among different speakers, allowing speech by some but not
others.’” Nat’l Inst. of Fam. & Life Advocs. v. Becerra, 138
S. Ct. 2361, 2378 (2018) (quoting Citizens United v. FEC,
558 U.S. 310, 340 (2010)) (alteration in original omitted).
The reason is that “[s]peaker-based laws run the risk that ‘the
State has left unburdened those speakers whose messages are
in accord with its own views.’” Id. (quoting Sorrell, 564 U.S.
at 580); see also Reed v. Town of Gilbert, 576 U.S. 155, 170
(2015) (“Speech restrictions based on the identity of the
speaker are all too often simply a means to control content.”
 BOARDMAN V. INSLEE 65

(quoting Citizens United, 558 U.S. at 340) (alteration
omitted).).

I-1501 is not a direct restriction on speech. But it is a
restriction on critical State-held information that in turn
enables speech. See United Reporting, 528 U.S. at 42 (Scalia,
J., concurring); id. at 43 (Ginsburg, J., concurring). And in
this case, that speech consists of core First Amendment
material—speech on vital matters of public concern, such as
collective bargaining and candidates for public office. Janus,
138 S. Ct. at 2474–77. The risks that a speaker-based law
creates for free expression are not materially different when
the government engages in a selective disclosure of State-
controlled information that is used for speech purposes.
Sorrell, 564 U.S. at 569. Today’s majority opinion accepts
this same premise.

In this case, the nature of the speaker that is singled
out—public-sector unions—is central. I will have more to
say about the majority’s mistaken determination that I-1501
fashions a distinction based merely on the incumbent Unions’
“legal status” as exclusive bargaining representatives. But
what is critical to appreciate up front is that unions both have
views and espouse them. The foundation of the Supreme
Court’s decision in Janus was that “unions express views on
a wide range of subjects.” 138 S. Ct. at 2475. Requiring
nonmembers to subsidize public-sector union speech
therefore violated the First Amendment because
“[c]ompelling individuals to mouth support for views they
find objectionable violates th[e] cardinal constitutional
command” against “forc[ing] citizens to confess by word or
act their faith” in particular ideas. Id. at 2463 (quoting W. Va.
Bd. of Educ. v. Barnette, 319 U.S. 624, 642 (1943)) (emphasis
66 BOARDMAN V. INSLEE

omitted). That I-1501 isolates incumbent Unions for special,
preferential treatment is therefore highly relevant.

The final set piece is I-1501’s distinctive features. It is a
unique law in at least two key respects. The first is that it
gives the incumbent Unions the full range of State-held
information about in-home care providers but prevents
everyone else from obtaining that information through a
public records request. Wash. Rev. Code §§ 42.56.645(1)(d),
43.17.410(1) (2018). The “sensitive personal information”
that the law protects from public disclosure (but makes
available to the incumbent Unions) is the care providers’
“names, addresses, GPS coordinates, telephone numbers,
email addresses, social security numbers, driver’s license
numbers, or other personally identifying information.” Id.
§ 42.56.640(2)(b). The incumbent Unions can get all of this
information; for purposes relevant here, no one else can get
any of it. Other states have enacted laws that give unions
who are exclusive bargaining agents more modest
informational advantages over others, as I will discuss further
below. But the parties have not identified any state with a
law like Washington’s, which gives the incumbent Unions a
complete monopoly over the information, including the very
identity of the employees.

The second unique aspect of I-1501 is that it involves
quasi-public-sector employees who are very difficult to
identify or locate. This point is not disputed. More
traditional public-sector employees are not hard to find.
Teachers, for example, show up to a school each day. Police
officers are based out of a precinct. In-home care providers
are quite different. They operate in the privacy of homes,
often their own. Their labors can be of a deeply private
nature.
 BOARDMAN V. INSLEE 67

The State’s information about the identities and contact
information for in-home care providers is thus the golden
ticket to communicating with them—which explains the
extensive litigation over the State-held lists and the expensive
ballot initiative process that produced the law before us.
Absent the State-held contact information, it is extremely
difficult to locate the care providers, especially en masse.
Indeed, the entire asserted rationale for I-1501 is to ensure the
privacy of in-home care providers, and, by extension, those
for whom they care. If Washington’s care providers could be
easily identified through other means, I-1501 and the privacy
protection it supposedly confers would be pointless.

B

With these various pieces in place, the fundamental
problem with I-1501 reveals itself. It is this: through extreme
favoritism as to who may receive critical and otherwise
unavailable speech-enabling information about in-home care
providers, I-1501’s speaker-based distinction powerfully
favors those views inherent to incumbent unions while
creating significant obstacles to speech for anyone with
opposing views. The information disparity that I-1501
creates in the First Amendment’s political speech heartland
is so severe that the inference of viewpoint discrimination is
inescapable.

Whatever differing views unions may have among
themselves on some issues, unions—and especially
incumbent unions—have an obvious, intrinsic view on
whether unions should be disempowered: they
understandably disagree with that position. Unions are, in
other words, “pro-union.” The issue is one of great public
significance. Some people wish to promote the activities of
68 BOARDMAN V. INSLEE

public-sector unions and believe they are vitally important to
American society. Others dislike the work of public-sector
unions and believe, for various reasons, that these unions do
more harm than good. The passions on both sides of this
issue are well known. I-1501 gives critical State-controlled
information to powerful actors on only one side of this
important public debate, while denying everyone on the other
side the same information.

It does not matter to the constitutional analysis that under
I-1501, other persons or groups who may also be “pro-union”
cannot get the same State-held information as the incumbent
Unions. The reason is that by giving the information to the
incumbent Unions only, I-1501 promotes only one side of an
overall debate, favoring a positive view of unions while
imposing a significant burden on everyone who wishes to
promote an opposing perspective. “The First Amendment
forbids the government to regulate speech in ways that favor
some viewpoints or ideas at the expense of others.” Matal v.
Tam, 137 S. Ct. 1744, 1757 (2017) (quoting Lamb’s Chapel
v. Ctr. Moriches Union Free Sch. Dist., 508 U.S. 384, 394
(1993)) (internal quotation marks and alteration omitted).
That is what Washington is doing here, except through the
discriminatory denial of the critical information that enables
speech rather than through a prohibition on the core political
speech that the State-held information later facilitates. This
selective release of information violates the First Amendment
no less. Sorrell, 564 U.S. at 569; United Reporting, 528 U.S.
at 42 (Scalia, J., concurring); id. at 43 (Ginsburg, J.,
concurring).

That I-1501 gives total access to incumbent unions while
denying all access to rival ones, like the one Benn and
Thurber support, makes the viewpoint discrimination even
 BOARDMAN V. INSLEE 69

more conspicuous. Incumbents seek to promote incumbency;
those with power wish to retain it. John Hart Ely, Democracy
and Distrust: A Theory of Judicial Review 106 (1980). An
incumbent union has a natural and entirely expected view on
the question whether the union is properly serving the
interests of the represented workers: the incumbent believes
it should continue in the role. Persons seeking to promote a
rival union, by contrast, strongly disagree. In this case, Benn
and Thurber have various policy disagreements with an
incumbent Union and would like to see it replaced. By
discriminating among different unions through the sharply
unequal disclosure of information vital for effective
communication, the State necessarily subsidizes the
incumbent Unions and, with it, those Unions’ viewpoint that
they should stay in power. See, e.g., NLRB v. Magnavox Co.
of Tenn., 415 U.S. 322, 325 (1974) (“[I]t is difficult to assume
that the incumbent union has no self-interest of its own to
serve by perpetuating itself as the bargaining
representative.”).

It is, once again, no answer to say that persons who
believe the incumbent Unions are doing a fine enough job
also cannot get the State-held information under I-1501. The
problem here is the same one we observed with “pro-union”
and “anti-union” groups, except another level down. See
Rosenberger, 515 U.S. at 831–32 (“[T]hat debate is not
skewed so long as multiple voices are silenced is simply
wrong . . . .”). In the intensely political debate about whether
an incumbent union should stay in power, Washington
provides valuable State-held information to one side of this
debate—represented by the incumbent Union itself—and
everyone with a different view of the incumbent Union gets
nothing. By conferring on the incumbent Unions a powerful
70 BOARDMAN V. INSLEE

information advantage, the State is necessarily promoting the
view that the incumbent Unions should stay in power.

The majority nonetheless holds that I-1501 “den[ies]
access to Provider Information irrespective of a requester’s
views, ideology, or message on any particular subject.” That
is not the right way to look at this. Given the absolute bar on
public records requests, the only relevant entities who can
obtain the information are the incumbent Unions. Wash. Rev.
Code § 42.56.645(1)(d) (2018). And the only entities that are
incumbent unions are ones who have (1) a certain “pro-
union” view of unions and (2) a positive view as to whether
the incumbent union should remain in power. Nobody who
has an “anti-union” view can get the information. Nor can
anyone who opposes the incumbent union. The viewpoint
discrimination built into I-1501 is patent.

Treating I-1501 as “facially neutral,” as the majority does,
fails to appreciate I-1501’s obvious design. The First
Amendment disallows government regulation “that is in
reality a facade for viewpoint-based discrimination.”
Cornelius v. NAACP Legal Def. & Educ. Fund, Inc., 473 U.S.
788, 811 (1985). If I-1501 had stated on its face that no in-
home care provider information would be available to
(1) anyone who opposed public-sector unions and (2) anyone
who sought to oppose the incumbent Unions in an election,
nobody would regard this as “facially neutral.” See
Rosenberger, 515 U.S. at 828–29. Washington has
accomplished the same result through the expedient (or
subterfuge) of granting exclusive access to critical
information to the incumbent Unions, which those Unions
then use to engage in core political speech with in-home care
workers, including speech that promotes the Unions
themselves.
 BOARDMAN V. INSLEE 71

What we have here is thus not “facial neutrality,” but a
statutory framework that promotes one set of pro-union and
pro–incumbent union views over others in the First
Amendment’s most hallowed ground of core political speech.
It is therefore too simple to maintain, as the majority does,
that “the Unions happen to hold certain viewpoints and also
happen to receive” certain benefits under I-1501. Nothing
about I-1501 turns on mere happenstance. The powerful,
built-in informational advantage given to the incumbent
Unions is so extreme as to warrant the inference of
impermissible viewpoint discrimination. United Reporting,
528 U.S. at 42 (Scalia, J., concurring); id. at 43 (Ginsburg, J.,
concurring).

The majority’s repeated characterization of I-1501 as a
“subsidy” thus obscures what is at stake here. I-1501 can
hardly be equated with a traditional subsidy like a tax benefit.
See Regan v. Tax’n with Representation of Wash., 461 U.S.
540 (1983). Instead, and through the discriminatory non-
disclosure of critical, speech-enabling information that is not
otherwise readily available, I-1501 confers an enormous
advantage to incumbent unions who necessarily hold views
intrinsic to such organizations. I-1501 is much more than a
mere subsidy: it is the key to the incumbent Unions’ ability
to communicate with in-home care providers, and the barrier
that prevents anti-union groups and rival unions from doing
so effectively. I-1501 is a direct limitation on (and promotion
of) expression in a way that a bare financial subsidy is not.
Even so, the majority agrees that a subsidy cannot be awarded
“based on an illegitimate criterion such as viewpoint.”
United Reporting, 528 U.S. at 43 (Ginsburg, J., concurring)
(citing Regan, 461 U.S. at 540). Because I-1501 operates in
that manner, the majority’s characterization of I-1501 as a
subsidy is ultimately irrelevant.
72 BOARDMAN V. INSLEE

C

The majority concludes there is no viewpoint
discrimination here because, in its view, “the Unions’ current
access to Provider Information is based entirely on their legal
status as certified exclusive bargaining representatives under
Washington law.” This is mistaken, and it is a mistake with
serious First Amendment consequences.

“Legal status” is merely a label that a law creates. To
know whether a distinction based on “status” reflects
viewpoint discrimination, one needs to know more about the
nature of the “status” and its origins. Imagine a law
providing that information could only be given to the political
party that had achieved the “status” of winning the election.
Such a law would plainly discriminate based on viewpoint.
See United Reporting, 528 U.S. at 43 (Ginsburg, J.,
concurring). The reason is that by its very nature, a speaker-
based law that turned on political affiliation would have an
obvious valence to viewpoint.

Take as another example a contractor who won a
government project through a competitive bid process. A law
that gave this contractor state-held information based on its
“status” as the bid-winner might seem neutral. But what if
we learned that the bid process involved consideration of the
contractors’ political views so that only contractors with a
particular view on a political issue could secure the work? At
that point, a “status” that initially seemed innocuous would
raise a serious inference of viewpoint discrimination. To say
that the law here distinguishes based on “legal status,” as the
majority does, is only to ask the viewpoint discrimination
question, not to answer it.
 BOARDMAN V. INSLEE 73

In this case, given I-1501’s distinctive features, the
“status” of being an incumbent public-sector union cannot be
divorced from the reality that public-sector unions’ speech
reflects political viewpoints, including about unions
themselves. This was the basis for Janus’s holding that it
violated the First Amendment to require nonmembers to
subsidize public-sector union speech. Janus leaves no doubt
on this point. There, the Supreme Court considered whether
public-sector union speech could be analogized to speech on
matters of “private concern,” so that the speech public-sector
workers were required to fund through agency fees was more
analogous to speech that “forms part of the official duties of
the union officers who engage in speech.” Janus, 138 S. Ct.
at 2474.

The Supreme Court squarely rejected this sanitized view
of union speech. It held that “union speech in collective
bargaining, including speech about wages and benefits,” is
“‘a matter of great public concern.’” Id. (quoting Harris, 134
S. Ct. at 2642–43). Moreover, unions “speak out in collective
bargaining on controversial subjects such as climate change,
the Confederacy, sexual orientation and gender identity,
evolution, and minority religions,” topics that are “sensitive
political topics” and “undoubtedly matters of profound value
and concern to the public.” Id. at 2476 (internal quotation
marks and footnotes omitted).

Based on this conception of “what actually occurs in
public-sector collective bargaining,” id., it cannot be
maintained that a law that gives otherwise unavailable
information to incumbent public-sector unions and no one
else turns merely on a formal “legal status” untethered to any
viewpoints. Unions have views on topics of public concern,
which include—critically—the claimed benefits of unions
74 BOARDMAN V. INSLEE

themselves and the claimed benefits of one union over
another. All of this explains why, under Janus, an objecting
nonmember cannot be compelled to fund union speech under
the First Amendment. If what matters to the analysis is solely
the State-created “status” of being an exclusive bargaining
agent, Janus should have come out the other way. And if the
State had enacted a direct restriction of speech premised on
the incumbent Unions’ State-created “status,” I am hard-
pressed to see how anyone would regard the law as insulated
from searching First Amendment review.

For its “legal status” theory, the majority relies almost
exclusively on Perry Education Ass’n v. Perry Local
Educators’ Ass’n, 460 U.S. 37 (1983). But Perry is not a safe
haven for I-1501’s clear viewpoint discrimination. In Perry,
a public-sector collective bargaining agreement allowed an
incumbent teacher’s union access to an interschool mail
system and teacher mailboxes but largely denied rival unions
the same access. Id. at 39. A rival union sued, claiming that
this differential access constituted impermissible viewpoint
discrimination in violation of the First Amendment. Id. at
48–49. The Court disagreed. It noted (among other things)
that on the record before it, there was “no indication that the
School Board intended to discourage one viewpoint and
advance another,” so that “it is more accurate to characterize
the access policy as based on the status of the respective
unions rather than their views.” Id. at 49.

Perry, decided in 1983, sits uncomfortably with the
Supreme Court’s modern jurisprudence concerning public-
sector unions. Among other things, Perry relied on Abood,
concluding that the “exclusion of the rival union may
reasonably be considered a means of insuring labor peace
within the schools.” Perry, 460 U.S. at 52. But Janus
 BOARDMAN V. INSLEE 75

overruled Abood and specifically rejected Abood’s “labor
peace” rationale. Janus, 138 S. Ct. at 2466. As Janus
explained, “Abood cited no evidence that the pandemonium
it imagined would result if agency fees were not allowed, and
it is now clear that Abood’s fears were unfounded.” Id. at
2465. Perry’s Abood-based foundations are now in serious
question. Perry of course remains binding on us, regardless
of its evident tension with Janus. But Perry does not save I-
1501, for four main reasons.

First, Perry presented a fundamentally different First
Amendment question because it concerned access to physical
government property, not access to government-held
information. Central to Perry was the Supreme Court’s
determination that the school mail system and mailboxes
were neither public property that is traditionally open for
expressive activity nor a so-called “limited public
forum”—First Amendment parlance for public property that
“the State has opened for use by the public as a place for
expressive activity.” Perry, 460 U.S. at 45–46. Because the
school mail system was a “nonpublic forum,” the state had
“the right to make distinctions in access on the basis of
subject matter and speaker identity.” Id. at 49.

As the Supreme Court later explained, Perry was a case
about a claimed “right of access to government property for
use in speaking to potentially willing listeners.” Minn. State
Bd. for Cmty. Colls. v. Knight, 465 U.S. 271, 282 (1984).
While the State claims the mantle of Perry in this case, it
nonetheless concedes in its briefing in this court that “[u]nlike
the present case, the [Perry] decision primarily involved an
analysis as to what extent a public school’s mailbox system
was a public forum,” and “[a] forum analysis is not relevant
here, where the issue is access to government information.”
76 BOARDMAN V. INSLEE

There are plainly different considerations at stake in the
context of a request for access to government property that,
as in Perry, the government does not otherwise open to
everyone for expressive purposes. Perry was concerned with
“prevent[ing] the District’s schools from becoming a
battlefield for inter-union squabbles.” 460 U.S. at 52. One
can imagine various examples of outside groups trying to
gain physical access to locations where public-sector
employees work, which could prove highly disruptive. There
may thus be valid reasons for limiting such access to certain
groups, so that it could be permissible in some circumstances
to allow access to a union that is the exclusive bargaining
agent but to exclude others. These reasons simply do not
apply in this case, which involves a request for access to
information and does not present the same burdens for the
government.

Second, in upholding the policy limiting mail system
access to the incumbent union, Perry repeatedly made clear
that no evidence of viewpoint discrimination was to be found.
This was central to the Court’s result. In Perry, there was “no
indication that the School Board intended to discourage one
viewpoint and advance another.” 460 U.S. at 49. Later,
when responding to Justice Brennan’s dissent, the Court
again emphasized that “there is no indication in the record
that the policy was motivated by a desire to suppress [the
rival union’s] views.” Id. at 50 n.9. The same cannot be said
here, given the extreme information disparity inherent in I-
1501’s basic design and the extensive evidence showing that
an objective of I-1501—if not the objective—was to stop
anti-union organizations from disseminating an anti-union
message to care providers.
 BOARDMAN V. INSLEE 77

Third, Perry upheld the school mail system policy as
“reasonable[]” based on “the substantial alternative channels
that remain[ed] open for union-teacher communication to
take place.” 460 U.S. at 53. The teachers worked in school
buildings, and the rival union still retained access to school
“bulletin boards” and “meeting facilities” and could (with a
principal’s approval) “make announcements on the public
address system.” Id. at 41, 53. Critically, there was “no
showing [t]here that [the rival union’s] ability to
communicate with teachers [was] seriously impinged by the
restricted access to the internal mail system.” Id. at 53.

This case is completely different. There are no
“substantial alternative channels” to reach in-home care
providers in Washington. Id. It is undisputed that there are
approximately 45,000 care providers who work in private
homes throughout the State, so that “the only meaningful and
practical way to communicate with Providers is through using
public records.” The high turnover rate among care providers
only adds to the communication problem. Unlike the teachers
in Perry who worked in one of thirteen schools in a single
Indiana township, 460 U.S. at 39, Appellants here cannot
easily know where the in-home care providers work or even
who they are. And if it were the case that care providers
could be readily located and contacted, then the entire
privacy-based rationale that supposedly justifies I-1501
would collapse.

Finally, there is one last key difference between the
mailbox access restriction in Perry and I-1501: Perry’s
preferential access policy did not apply during periods of
union elections, when rival unions are vying to be the
exclusive bargaining agent. See Perry, 460 U.S. at 41
(“[U]nder Indiana law, the preferential access of the
78 BOARDMAN V. INSLEE

bargaining agent may continue only while its status as
exclusive representative is insulated from challenge. While
a representation contest is in progress, unions must be
afforded equal access to such communication facilities.”)
(citation omitted); id. at 53 (“During election periods, [the
rival union] is assured of equal access to all modes of
communication.”). In fact, this was one of the factors the
Court cited in affirming “the reasonableness of the
limitations” that Indiana law imposed. Id.

I-1501 once again differs considerably. At no point in the
union electoral process can rival unions obtain care provider
identifying information. As noted above, to trigger an
election and displace the incumbent bargaining
representative, a proponent must submit a representation
petition supported by 30% or more of the care providers in
the bargaining unit. See Wash. Rev. Code § 41.56.070 (Supp.
2020). Benn and Thurber explain that “[w]ithout a current
list of Childcare Providers, it will be impossible for us to get
30% of Childcare Providers to sign cards signifying an
interest to hold a new election.” Unlike in Perry, I-1501
creates no window in the overall electoral process during
which rival unions may gain access to care provider lists. 3

3
The majority contends that Perry’s reliance on the “substantial
alternative channels” for union-teacher communications and the rival
unions’ access to the mail system during union elections reflected
considerations unique to a public forum analysis. But I read Perry as
stressing these points as part of showing overall why the Indiana law at
issue was not viewpoint discriminatory. And if the majority is correct that
Perry’s analysis on these two points was instead limited to the nonpublic
fora context, that only further underscores my point that this case presents
a very different First Amendment question than Perry.
 BOARDMAN V. INSLEE 79

In light of all of this, it is quite difficult to credit the
majority’s “example” that supposedly “demonstrate[s] the
propriety” of I-1501. The majority reasons that I-1501 does
not discriminate based on viewpoint because if Appellants
succeed in replacing the incumbent Unions with rival unions,
then those rival unions would secure the “legal status”
entitling them to the care provider contact information on an
exclusive basis. But what the majority overlooks is that I-
1501 is designed to make it as difficult as possible for that
transfer of power to ever happen.

By denying rival unions critical State-held information,
Washington law ensures that incumbent Unions have easy
access to the care providers who vote in union elections,
while simultaneously making it extremely difficult for rival
unions to have the same communications. As compared to
the incumbent Unions, rival unions must compete in a
Kafkaesque election process where they cannot easily
identify who the voters even are or how they can be
contacted. And if it were really that easy for rival unions to
win an election in this rigged regime, would the incumbent
Unions have spent so much time and money securing the
passage of I-1501, only for it to be used against them?

I do not think the First Amendment either allows or
requires us to ignore the obvious political realities of I-1501’s
basic design. That law’s extreme incumbent advantage—and
the viewpoint discrimination it necessarily reflects—cannot
be regarded as a “neutral” law.

D

The highly troubling documentary record in this case also
demonstrates that the majority’s narrow focus on the
80 BOARDMAN V. INSLEE

incumbent Unions’ “legal status” is artificial. See Sorrell,
564 U.S. at 563–64 (observing that “[o]n its face, Vermont’s
law enacts content- and speaker-based restrictions,” and then
noting that “[a]ny doubt . . . is dispelled by the record and by
formal legislative findings”). The documentary record is not
necessary to my analysis because I-1501’s text and operation
more than sufficiently demonstrate viewpoint discrimination.
But the record evidence is certainly corroborative on that
point, and it should be cause for serious concern.

I-1501 arose from warring viewpoints about the
incumbent Unions. After the Freedom Foundation
successfully used State-provided lists to persuade many in-
home care providers to stop paying union dues in the wake of
Harris, the Unions sued the Foundation and even Benn in her
personal capacity to stop them from obtaining the
information. When that failed, the Unions pursued a ballot
initiative process and made public and private statements,
demonstrating that the obvious motivation behind I-1501 was
to silence anti-union voices and entrench in office the
incumbent Unions. There is no real dispute that this plan has
worked as planned, as Appellants face serious impediments
in identifying in-home care providers and promoting their
contrary views. For example, there is evidence in the record
that as of January 2017—by which point the Unions’
litigation tactics and I-1501 had successfully prevented the
Foundation from obtaining up-to-date lists of individual
providers—11% of individual providers had left SEIU 775.
Compare this to over 60% of family child care providers
leaving SEIU 925 between July 2014 and January 2017.

Government efforts to engage in viewpoint discrimination
are usually better concealed. In this case, and strikingly, the
viewpoint discrimination motivating I-1501 was not just
 BOARDMAN V. INSLEE 81

poorly hidden, it was actually touted as a principal selling
point of the law. The Union-driven Campaign supporting I-
1501 prepared materials explicitly urging voters to support I-
1501 because “[g]roups like the Freedom Foundation are
threatening unions” and “threatening our union.” SEIU 775
urged its members to vote for I-1501 “to stop the Freedom
Foundation” and its “anti-union agenda.” For a town hall
supporting I-1501, SEIU 775’s draft remarks called for
Washington’s Governor (whose candidacy the Unions
actively supported) to urge passage of I-1501 because “[w]hat
the Freedom Foundation is doing is wrong.” And there are
many other examples of similar statements in the record. The
Unions and their supporters are of course entitled to their
views about the Freedom Foundation. But operationalizing
those views through the coercive force of law is an affront to
First Amendment ideals.

The majority’s apparent response is that “there is no
evidence in the record . . . indicating that the more than 2.2
million Washington voters who voted in favor of Initiative
1501 were motivated by an irrational prejudice.” But ballot
initiatives have been invalidated on First Amendment
grounds, see, e.g., Cal. Democratic Party v. Jones, 530 U.S.
567, 570, 586 (2000); Citizens Against Rent Control/Coal. for
Fair Housing v. City of Berkeley, 454 U.S. 290, 292, 300
(1981); Mont. Chamber of Com. v. Argenbright, 226 F.3d
1049, 1052, 1058 (9th Cir. 2000), without any suggestion that
courts or litigants were somehow required to probe the hearts
of the millions of persons who enacted a viewpoint
discriminatory law to confirm they had nefarious anti-speech
intent. In any event, that I-1501 was passed as a ballot
proposition is beside the point. As the Supreme Court has
explained, “[i]t is irrelevant that the voters rather than a
legislative body enacted [the challenged law], because the
82 BOARDMAN V. INSLEE

voters may no more violate the Constitution by enacting a
ballot measure than a legislative body may do so by enacting
legislation.” Citizens Against Rent Control, 454 U.S. at 295.

The issue here is decidedly not the good faith of the
Washington voters who enacted I-1501. The casebooks are
full of decisions rejecting under the First Amendment
potentially well-meaning efforts to restrict speech that was
regarded as contrary to shared values. See, e.g., R.A.V. v. City
of St. Paul, 505 U.S. 377, 396 (1992); Texas v. Johnson, 491
U.S. 397, 420 (1989); Nat’l Socialist Party of Am. v. Village
of Skokie, 432 U.S. 43, 44 (1977) (per curiam). Whether
Washington’s voters believed in I-1501’s privacy-based
rationales or thought that anti-union forces should be
disempowered, or both, the result is the same: I-1501, through
its text and operation, discriminates based on viewpoint.

III

“Discrimination against speech because of its message is
presumed to be unconstitutional.” Rosenberger, 515 U.S. at
828. As a result, laws reflecting viewpoint discrimination,
such as I-1501, must be subject to strict scrutiny, the highest
level of First Amendment review. Reed, 576 U.S. at 170;
First Resort, Inc. v. Herrera, 860 F.3d 1263, 1277 (9th Cir.
2017).

To withstand strict scrutiny, I-1501 must be “narrowly
tailored to serve compelling state interests,” Reed, 576 U.S.
at 163, and “the least restrictive means of achieving” those
interests, McCullen v. Coakley, 573 U.S. 464, 478 (2014).
The burden is on the State to make this showing. See Ark.
Writers’ Project, Inc. v. Ragland, 481 U.S. 221, 231 (1987).
“If a less restrictive alternative would serve the state’s
 BOARDMAN V. INSLEE 83

compelling interest with the same level of effectiveness, the
state must use that alternative.” Victory Processing, LLC v.
Fox, 937 F.3d 1218, 1228 (9th Cir. 2019). “Furthermore,
when the plaintiff offers ‘a plausible, less restrictive
alternative . . . , it is the Government’s obligation to prove
that the alternative will be ineffective to achieve its goals.’”
Id. (quoting United States v. Playboy Ent. Grp., Inc., 529 U.S.
803, 816 (2000)). “To meet this burden, the state must
provide ‘more than anecdote and supposition;’ it must point
to evidence in the legislative record or present other evidence
that demonstrates why the challenged restriction, rather than
a less restrictive alternative, is necessary to further its
significant interests.” Id. (quoting Playboy Ent. Grp., 529
U.S. at 822).

I-1501 fails strict scrutiny. Assuming that protecting the
privacy of in-home care providers and the persons they
support is a compelling state interest, I-1501 is not narrowly
tailored toward that objective. As an initial matter, there is
scant evidence as to how restricting access to the names of in-
home care providers will protect either them or the vulnerable
persons for whom they care from identity theft. The State has
not brought forward any evidence that previous public
records requests of care providers’ identities have led to
identity theft. As the State Director of AARP Washington
(which declined to endorse I-1501) explained, identity thieves
“do not fill out public records requests to get their victims.
That’s not how it works.” Indeed, the district court below
itself noted that the evidence of identity theft “is thin,” and
“[t]he Initiative does not explicitly articulate how withholding
caregiver identities will protect vulnerable individuals.”

The State also fails to show that I-1501 is the least
restrictive means for achieving its stated goals. See Victory
84 BOARDMAN V. INSLEE

Processing, 937 F.3d at 1228. Among other things, the State
has not explained why, to avoid identity theft, it needs to
prevent the disclosure of all identifying information of in-
home care providers, as opposed to merely some of it. The
State also has not explained why it could not protect against
identity theft by allowing some greater disclosure of care
provider information while instituting some confidentiality
restrictions surrounding its receipt. To this point, the
incumbent Unions, State agencies, and State contracting
parties are given complete access to care provider
information, with “the recipient agree[ing] to protect the
confidentiality of the information.” Wash. Rev. Code
§ 42.56.645(1)(a), (d), (f)–(g) (2018). If that protection is
sufficient to protect privacy interests, including when the
information is given to the incumbent Unions, it is unclear
why this would not be sufficient for Appellants as well. By
failing to explain why less-restrictive alternatives “will be
ineffective to achieve its goals,” Victory Processing, 937 F.3d
at 1228, the State fails to carry its burden.

Comparing I-1501 to the laws of other states confirms it
is not narrowly tailored. While some states have laws that
give incumbent unions some informational advantages when
it comes to the identities and contact information of the
worker class, the parties have not identified a state law that is
nearly as extreme as Washington’s. The State has cited laws
from Illinois, Maryland, Wisconsin, and California. But in
the examples of Illinois, Maryland, and Wisconsin, and as the
State here does not dispute, the statutes give incumbent
unions exclusive access to public-sector workers’ contact
information, but not the workers’ names, which is the whole
ballgame. See 5 Ill. Comp. Stat. §§ 140/7(1)(b), 315/3(n),
(o), 315/6(c) (West, Westlaw through P.A. 101-651)
(prohibiting disclosure of public employees’ “[p]rivate
 BOARDMAN V. INSLEE 85

information” but requiring such information to be provided to
incumbent unions); id. §§ 140/2(c-5), 315/6(c-5) (omitting
employee names from the definition of “private information”
and related non-disclosure provisions); Md. Code Ann., State
Pers. & Pens. § 3-208 (West, Westlaw through legislation
effective July 1, 2020) (mandating that public employees’
names and contact information be disclosed to their
representative unions); Md. Code Ann., Gen. Prov. § 4-331
(prohibiting public disclosure of employees’ home addresses
or phone numbers, but not names); Wis. Stat. Ann.
§ 19.36(10)(a) (West, Westlaw through 2019 Act 186)
(omitting names from the list of information to which
incumbent unions have exclusive access); see also 19 Or.
Rev. Stat. §§ 192.355(3), 243.804(4)(a) (2019) (requiring
public agencies to provide public employees’ contact
information to incumbent unions but omitting from other
non-disclosure rules the names of public employees). On
their face, these statutes allow rival unions and other groups
to at least identify who the employees are, so that they may
be located by other means.

An Illinois statutory provision referenced in Janus was
likewise much more narrowly tailored than I-1501. In Janus,
the Supreme Court noted that “[e]ven without agency fees,
designation as the exclusive representative confers many
benefits,” including “obtaining information about
employees.” 138 S. Ct. at 2467 (citing 5 Ill. Comp. Stat.
315/6(c) (West 2016)). But the Illinois statute that Janus
cited at the time only required public employers upon request
to “furnish the exclusive bargaining representative with a
complete list of the names and addresses of the public
employees in the bargaining unit.” 5 Ill. Comp. Stat.
315/6(c). Nothing in this statute purported to preclude other
entities from obtaining public employees’ names through
86 BOARDMAN V. INSLEE

other means, such as public records requests. And while
Illinois changed its law post-Janus (an evident reaction to
Janus), the new law, as noted above, on its face does not
prohibit disclosure of public-sector employees’ names. See
5 Ill. Comp. Stat. § 315/6(c-5).4

Washington also points to California law. But while
California bars from public release both the names and the
contact information of in-home care providers, it does make
this information available to “any labor organization seeking
representation rights.” Cal. Gov. Code Ann. § 6253.2(b)
(West 2019). That is a critical difference compared to I-1501.
Washington does not dispute that under California law, Benn
and Thurber’s rival union could obtain the information, which
they cannot do under Washington law. In fact, under
California law, rival unions can even obtain workers’ cell
phone numbers and personal email addresses. Id. I provide
these examples from California and other states not to suggest
they are necessarily immune from constitutional challenge,
but to show that by comparison, Washington’s law is an
outlier. Other states’ laws show that Washington could have
more narrowly tailored I-1501 in furthering any supposed
privacy interest.

I imagine there could be laws that give incumbent unions
a somewhat greater informational advantage over others, but

4
Washington argues here that Janus’s reference to the Illinois
provision implies that I-1501 is constitutional. Given the significant
differences between I-1501 and Illinois law (as it existed at the time of
Janus), the State’s position is not tenable. Indeed, it is somewhat
remarkable for Washington to claim that Janus somehow provides support
for I-1501 when I-1501 reflects an obvious effort to make an end-run
around Janus by preventing in-home care providers from knowing they
have a Harris/Janus right not to pay union agency fees.
 BOARDMAN V. INSLEE 87

which pass constitutional muster because the more modest
information disparity does not raise such an obvious inference
of viewpoint discrimination. But here, I-1501 gives the
incumbent Unions all the identifying information, and anti-
union groups and rival unions get none of it. I-1501 thus
maximizes to the greatest extent the incumbent Unions’
ability to convey viewpoints inherent to incumbent unions.
And it does so in the context of a population of public-sector
workers who cannot be easily identified otherwise given the
unique nature of their work. This whole regime is designed
to promote pro-union views and stem the tide of workers
leaving the unions in the wake of Harris and Janus.

In my view, these distinctive features cause I-1501 to fail
First Amendment scrutiny and thereby violate Appellants’
fundamental rights under the Equal Protection Clause. See
OSU Student Alliance v. Ray, 699 F.3d 1053, 1067 (9th Cir.
2012) (“The equal protection claims rise and fall with the
First Amendment claims.”). Because the court does not
vindicate Appellants’ constitutional rights, I respectfully
dissent.