UC Health v. National Labor Relations Board

United States Court of Appeals
FOR THE DISTRICT OF COLUMBIA CIRCUIT

Argued February 18, 2015 Decided September 18, 2015

No. 14-1049

UC HEALTH,
PETITIONER

v.

NATIONAL LABOR RELATIONS BOARD,
RESPONDENT

Consolidated with 14-1193

On Petition for Review and Cross-Application
for Enforcement of an Order
of the National Labor Relations Board

Kerry P. Hastings argued the cause and filed the briefs
for petitioner.

Kellie Isbell, Attorney, National Labor Relations Board,
argued the cause for respondent. With her on the brief were
Richard Griffin, Jr., General Counsel, John H. Ferguson,
Associate General Counsel, Linda Dreeben, Deputy Associate
General Counsel, and Ruth E. Burdick, Supervisory Attorney.
 2

Before: GRIFFITH, Circuit Judge, and EDWARDS and
SILBERMAN, Senior Circuit Judges.

Opinion for the Court filed by Circuit Judge GRIFFITH.

Concurring opinion filed by Senior Circuit Judge
EDWARDS.

Dissenting opinion filed by Senior Circuit Judge
SILBERMAN.

GRIFFITH, Circuit Judge:

In NLRB v. Noel Canning, 134 S. Ct. 2550 (2014), the
Supreme Court determined that the National Labor Relations
Board lacked authority to act during the time that three of its
five members held office via appointments that violated the
Recess Appointments Clause. This petition for review asks
whether a Regional Director of the Board had authority to
conduct a union election and certify its result during that same
time. We conclude that the Regional Director maintained his
authority and therefore deny the petition for review.

I

A

Section 3(a) of the National Labor Relations Act (NLRA)
calls for a National Labor Relations Board made up of five
members who are appointed by the President with the advice
and consent of the Senate. 29 U.S.C. § 153(a). The Board has
two main functions under the NLRA. Its quasi-judicial
function involves deciding whether particular conduct violates
the provisions of the Act that bar unfair labor practices. Id.
§§ 158, 160. The Board also has the primary responsibility for
 3

directing and holding representation elections by which
employees may choose to designate representatives for
purposes of collective bargaining. Id. § 159(b), (c).
Representation proceedings differ from unfair labor practice
proceedings in that they may only be reviewed in a court of
appeals when they are relevant to the court’s review of an
unfair labor practice proceeding. See Am. Fed’n of Labor v.
NLRB, 308 U.S. 401, 409 (1940). The Act authorizes the
Board to delegate to Regional Directors the authority to direct
representation elections and certify the results. 29
U.S.C. § 153(b). The Board first delegated its authority over
representation proceedings to the Regional Directors in 1961.
See 26 Fed. Reg. 3911 (May 4, 1961). Regional Directors
have been responsible for administering and certifying the
results of representation elections in their particular regions
ever since.

In what turns out to be a critical distinction for the
purposes of this challenge, the statute preserves for the Board
the power to review “any action of a regional director” taken
pursuant to that delegation, should a party object. 29 U.S.C.
§ 153(b). Thus, though the Board may empower Regional
Directors to oversee representation elections, the terms of the
delegation authorized under the Act provide that no Regional
Director’s actions are ever final on their own; they only
become final if the parties decide not to seek Board review or
if the Board leaves those actions undisturbed. Id.

The Act separately permits the Board to delegate “any or
all of the powers which it may itself exercise” to panels made
up of three or more of its members. 29 U.S.C. § 153(b). When
such a panel is created, the Act provides that two of its
members make up a quorum of that group. Id. This provision
allows the Board to process cases more quickly by spreading
them across more panels. Moreover, it allows the Board to
 4

continue to function without requiring the attendance of all
members. Should two of the five members’ terms expire, the
Board can continue to act despite the vacancies, while waiting
for Congress to appoint new members. Nevertheless, the
statute mandates that “three members of the Board shall, at all
times, constitute a quorum of the Board.” Id.

Between August 2010 and January 3, 2012, three of the
Board’s five members’ terms expired and the Senate refused
to confirm any of the President’s nominees to fill the
vacancies, leaving the Board without a quorum and therefore
unable to act. Claiming authority under the Recess
Appointments Clause, see U.S. Const. art. II, § 2, cl. 3, the
President named three individuals to the Board during a
three-day break between pro forma Senate sessions, but the
Supreme Court held those appointments unconstitutional in
Noel Canning, 134 S. Ct. 2550. No Senate-confirmed
appointees were sworn in until August 5, 2013. In the interim,
Regional Directors continued to hold elections and certify the
results, relying upon the Board’s previous delegation of
authority.

B

UC Health is a nonprofit corporation that operates a
hospital and provides inpatient and outpatient medical care
near the University of Cincinnati in Ohio. In March 2013,
while the Board lacked a quorum, the UC Health Public
Safety Union filed a petition with the Board seeking to
represent a unit of security officers employed by the
company. UC Health and the Union entered into a Stipulated
Election Agreement that identified the appropriate bargaining
unit and established that the Regional Director would
supervise a secret-ballot election following the Board’s
regulations. Under those regulations, if either party files
 5

timely objections to the election, it is entitled to plenary
review by the Board of any decision of the Regional Director
addressing those objections. 29 C.F.R. § 102.67(c). If no
objections are filed, the Regional Director “shall” certify the
results. Id. § 102.69(b).

The Regional Director held the representation election on
April 16, 2013, and the Union prevailed by a small margin.
The Regional Director certified the results without objection
from UC Health or the Union on April 24. Shortly thereafter,
the Union requested that UC Health bargain, but the company
refused. Citing that refusal to bargain, the Acting General
Counsel charged UC Health with an unfair labor practice. The
company defended itself on the ground that the Regional
Director had acted without authority because the Board lacked
a quorum at the time of the election.

The Board granted summary judgment to the Acting
General Counsel, finding that the company’s argument was
untimely because it had not been made during the
representation proceedings. See UC Health and UC Health
Public Safety Union, 360 N.L.R.B. No. 71 (2014). And even
if not waived, the Board concluded that UC Health’s
argument was without merit because the Board had delegated
authority over representational proceedings to the Regional
Directors in 1961; “[p]ursuant to this delegation, NLRB
Regional Directors remain vested with the authority to
conduct elections and certify their results, regardless of the
Board’s composition at any given moment.” Id. at *1 n.2.
Therefore, the Board determined that the election was valid
and UC Health had committed an unfair labor practice by
refusing to bargain with the Union. Id. at *2-3. UC Health
filed a petition for review in this court. We have jurisdiction
under 29 U.S.C. § 160(e), (f). Absent plain meaning to the
contrary, a court is obliged to defer to an agency’s reasonable
 6

interpretation of its statutory jurisdiction pursuant to the
familiar Chevron doctrine. City of Arlington v. FCC, 133 S.
Ct. 1863, 1870-71 (2013).

II

The sole question before us is whether the Regional
Director had authority to hold the representation election and
certify its results when the Board lacked a quorum. We hold
that he did.

A

The Board argues that we need not address whether the
Regional Director had the necessary authority because UC
Health has waived its challenge by failing to raise its
objection to the Regional Director’s authority at the
representation proceeding. “[A]s a general proposition, the
applicable case law emphasizes the need for parties seeking
judicial review of agency action to raise their issues before the
agency during the administrative process in order to preserve
those issues for review.” Advocates for Highway & Auto
Safety v. Fed. Motor Carrier Safety Admin., 429 F.3d 1136,
1148 (D.C. Cir. 2005). The NLRA states that “[n]o objection
that has not been urged before the Board, its member, agent,
or agency, shall be considered by the court, unless the failure
or neglect to urge such objection shall be excused because of
extraordinary circumstances.” 29 U.S.C. § 160(e). And under
the Board’s practice, “any issues that may be presented during
the representation proceeding must be offered there.” Pace
Univ. v. NLRB, 514 F.3d 19, 23 (D.C. Cir. 2008). Thus, the
Board claims, UC Health’s objection to the Regional
Director’s authority comes too late.
 7

We have consistently held, however, that challenges to the
composition of an agency can be raised on review even when
they are not raised before the agency. See Noel Canning v.
NLRB, 705 F.3d 490, 497 (D.C. Cir. 2013), aff’d on other
grounds, 134 S. Ct. 2550 (2014); Mitchell v. Christopher, 996
F.2d 375, 378-79 (D.C. Cir. 1993) (recognizing an exception
allowing parties to “raise[] for the first time on
review . . . challenges that concern the very composition or
‘constitution’ of an agency”). Since this challenge directly
involves the question of whether the Board’s lack of a quorum
stripped the Regional Directors of power, UC Health may
make it and we may review it.

The Board also asserts that UC Health may not challenge
the Regional Director’s authority because the company
voluntarily entered into the Stipulated Election Agreement
with the Union, and therefore agreed to let the Regional
Director supervise the election. According to the Board, the
agreement is a contract binding on both parties: UC Health
accepted the Regional Director’s authority to oversee the
election and, in exchange, received important procedural
benefits, including a prompt election. Because UC Health
explicitly agreed to the terms of the election, the Board insists
that the company cannot challenge one of those terms now.
We reject this argument. UC Health did not expressly give up
the challenge it brings now when it executed the Agreement;
it merely signed a form agreement providing that the Board’s
regulations would govern the election. Indeed, when UC
Health entered the Stipulated Election Agreement, no one
knew whether Congress might confirm the President’s
appointments and obviate the quorum issue by the time the
representation election in this case took place. And for that
matter, UC Health could not have known with any certainty
that the Board had no quorum even without Senate approval
for the President’s appointments until the Supreme Court
 8

handed down its decision in Noel Canning fourteen months
after the election. We will not hold UC Health responsible for
failing to see the future. And as we have already said,
“challenges that concern the very composition or
‘constitution’ of an agency” can “be raised for the first time
on review,” even if the objecting party failed to make that
objection at the appropriate time below. Mitchell, 996 F.2d at
378-79. Perhaps some objections to agency action could be
abandoned by explicit acceptance of the agency’s authority to
act under the statute. But we need not decide that here
because UC Health did not expressly abandon anything at all
in the Stipulated Election Agreement, and we will not hold it
responsible for failing to preserve expressly an argument the
substance of which had not yet arisen. See San Miguel Hosp.
Corp. v. NLRB, 697 F.3d 1181, 1187 n.7 (D.C. Cir. 2012).

B

UC Health’s challenge to the Board’s decision that the
Regional Director had authority to conduct the election fails
on the merits.

1

The Board interprets the relevant provision of the NLRA
to permit Regional Directors to continue exercising their
delegated authority while the Board lacks a quorum. We
consider the validity of the Board’s interpretation of the Act
under “the familiar two-step Chevron test.” 1 Int’l Alliance of

1
The dissent contends that Chevron is inapplicable in this case
because the Board opinion never relied on Chevron nor stated
explicitly that the statute is ambiguous. Dissent at 5-6. But our
precedent does not require such statements. In Arizona v.
 9

Theatrical & Stage Emps. v. NLRB, 334 F.3d 27, 31 (D.C.
Cir. 2003). At step one we ask “whether Congress has directly
spoken to the precise question at issue.” Chevron, U.S.A., Inc.
v. Natural Res. Def. Council, Inc., 467 U.S. 837, 842 (1984).
If Congress has addressed whether Regional Directors may
continue to act in the absence of a Board quorum, “that is the
end of the matter[,] for the court, as well as the agency, must
give effect to the unambiguously expressed intent of
Congress.” Id. at 842-43. If the statute is ambiguous, we move
to Chevron’s second step and ask whether the Board’s
interpretation is “a permissible construction of the statute” to
which we must defer. Id. at 843.

Thompson, 281 F.3d 248 (D.C. Cir. 2002), and other cases the
dissent cites, we refused to defer to an agency when the agency
itself made clear that it believed the interpretation on which it relied
was compelled by Congress and did not represent its own view of
an ambiguous statute. See id. at 254. Here, although the Board’s
explanation of its interpretation of the statute is brief, it contains
nothing suggesting that it viewed its interpretation as reflecting
Congress’s unambiguously expressed intent. Instead, the Board’s
reasoning rested on its interpretation of the extent of its prior
delegation of authority to the Regional Director, backed by its
reading of New Process Steel, L.P. v. NLRB, 560 U.S. 674 (2010),
and other case law. Of course, “[l]ike other administrative agencies,
the NLRB is entitled to judicial deference when it interprets an
ambiguous provision of a statute that it administers.” ITT Indus.,
Inc. v. NLRB, 251 F.3d 995, 999 (D.C. Cir. 2001) (internal
quotation marks and citations omitted). And we note that even if we
agreed with the dissent that the Board did not exercise its own
judgment here, we would not in any case rule for UC Health;
because we conclude that the statute is ambiguous, “[t]he law of
this circuit [would] require[] . . . that we withhold Chevron
deference and remand to the agency so that it can fill in the gap.”
PDK Labs., Inc. v. DEA, 362 F.3d 786, 798 (D.C. Cir. 2004).
 10

At the first step of Chevron, we conclude that the statute is
silent on the issue of the Regional Director’s power to act
when the Board lacks a quorum. The relevant text of the
statute provides:

The Board is authorized to delegate to any group of three
or more members any or all of the powers which it may
itself exercise. The Board is also authorized to delegate to
its regional directors its powers . . . to direct an election or
take a secret ballot . . . and certify the results thereof,
except that upon the filing of a request therefor with the
Board by any interested person, the Board may review any
action of a regional director delegated to him under this
paragraph, but such a review shall not, unless specifically
ordered by the Board, operate as a stay of any action taken
by the regional director. A vacancy in the Board shall not
impair the right of the remaining members to exercise all
of the powers of the Board, and three members of the
Board shall, at all times, constitute a quorum of the Board
....

29 U.S.C. § 153(b). In its adjudication of the unfair labor
practice charge against UC Health, the Board explained that it
interpreted the NLRA to permit the delegation of authority to
the Regional Director and concluded that “[p]ursuant to this
delegation, NLRB Regional Directors remain vested with the
authority to conduct elections and certify their results,
regardless of the Board’s composition at any given moment.”
UC Health, 360 N.L.R.B. No. 71 at *1 n.2.

UC Health argues that the structure of the statute forbids
this interpretation. In its view, the three-member quorum
requirement applies to the activities of not just the Board but
also of the Regional Directors. UC Health points out that the
statute implements the quorum requirement in the sentence
 11

immediately succeeding the sentence that authorizes the
Board to delegate authority to the Regional Directors. This
ordering, UC Health argues, expressly limits actors wielding
delegated Board authority precisely as the Board itself is
limited: Neither may act unless the Board has a quorum.

We are not convinced that the statutory text and structure
unambiguously require this interpretation. The plain language
of the Act applies the quorum requirement to the Board’s
authority to act, not the Regional Directors’ ability to wield
delegated authority. To the contrary, the structure of the
statute supports the Board’s interpretation just as well as it
might support UC Health’s construction. The first sentence of
the provision empowers the Board to delegate its final,
plenary authority to panels of its own members. The second
sentence authorizes the Board to delegate to the Regional
Directors the authority to oversee elections, provided that the
Regional Directors’ decisions always remain subject to Board
review if the parties dispute them. And the third sentence
specifies that the Board can only exercise its plenary, final
authority—whether to adjudicate unfair labor practice charges
or to review the decisions of Regional Directors in
representation elections, whether wielded by the Board as a
whole or by three-member delegee panels—if the Board has
at least three validly appointed members. Thus, though the
statute cabins the Board’s own ability to function without a
quorum, it says nothing about what effect the loss of a
quorum has on pre-existing delegations of authority to the
Regional Directors.

2

Because the statute is ambiguous on this point, we owe
deference to the Board’s interpretation, City of Arlington, 133
S. Ct. at 1870-71, if it is reasonable and consistent with the
 12

statute’s purpose, see Indep. Ins. Agents of Am., Inc. v.
Hawke, 211 F.3d 638, 643 (D.C. Cir. 2000).

We think the Board’s interpretation easily meets this
requirement. The Board explained that the NLRA “expressly
authorize[s] the delegation” of power to the Regional
Director, that the Board acted under that authority and
“delegated decisional authority in representation cases to
Regional Directors,” and that “[p]ursuant to this delegation,
NLRB Regional Directors remain vested with the authority to
conduct elections and certify their results, regardless of the
Board’s composition at any given moment.” UC Health, 360
N.L.R.B. No. 71 at *1 n.2. This is a sensible interpretation
that is in no way contrary to the text, structure, or purpose of
the statute. Though the dissent suggests we have failed to
examine the language or structure of the statute, Dissent at 5,
we have already explained in our step-one analysis that the
Board’s interpretation gives effect to each part of the statutory
provision. Moreover, allowing the Regional Director to
continue to operate regardless of the Board’s quorum is fully
in line with the policy behind Congress’s decision to allow for
the delegation in the first place. Congress explained that the
amendment to the NLRA that permitted the Board to delegate
authority to the Regional Directors was “designed to expedite
final disposition of cases by the Board.” See 105 Cong. Rec.
19,770 (1959) (statement of Sen. Barry Goldwater).
Permitting Regional Directors to continue overseeing
elections and certifying the results while waiting for new
Board members to be confirmed allows representation
elections to proceed and tees up potential objections for the
Board, which can then exercise the power the NLRA
preserves for it to review the Regional Director’s decisions
once a quorum is restored. And at least those unions and
companies that have no objections to the conduct or result of
an election can agree to accept its outcome without any Board
 13

intervention at all. The Board’s interpretation thus avoids
unnecessarily halting representation elections any time a
quorum lapses due to gridlock elsewhere.

The Board’s interpretation of the statute reads every
clause of the statutory provision harmoniously, and, as a
policy matter, it ensures adequate protection for the rights of
employers and unions alike. It is eminently reasonable.
Therefore we defer to the Board’s interpretation under
Chevron step two and uphold the Regional Director’s
authority to direct and certify the union election even while
the Board itself had no quorum.

3

UC Health argues, however, that the Board’s
interpretation of the statute is foreclosed by our decision in
Laurel Baye Healthcare of Lake Lanier, Inc. v. NLRB, 564
F.3d 469 (D.C. Cir. 2009). We conclude that UC Health is
wrong. Laurel Baye has no direct application here because it
addressed different statutory questions involving different and
highly distinguishable statutory language and altogether
different facts. Neither the holding nor the reasoning of that
case forbids the Board’s interpretation of its authority here, to
which we owe deference.

Laurel Baye addressed the lawfulness of the Board’s
effort to evade the quorum requirement imposed on its own
activities, not the status of authority previously delegated to
the Regional Directors once the Board loses a quorum. In
Laurel Baye, we considered the consequence of events
beginning in December 2007, when the Board had four
members, two of whom had terms that expired at the end of
the month. With their departures, the Board would no longer
satisfy the three-member quorum requirement the NLRA
 14

imposes “at all times,” see 29 U.S.C. § 153(b). In an effort to
overcome this impediment, the four-member Board delegated
its plenary, final authority to a panel of three members made
up of the two members whose terms continued into 2008 and
one of the members whose term would soon expire. When the
outgoing members finished their service at the end of 2007,
the Board had a total of only two members and no longer met
the overall three-member quorum requirement. By the same
token, the panel to which the Board had delegated its
authority was composed only of the two remaining members
of the Board. The Board reasoned that because the NLRA
imposes a separate, lower quorum requirement for
three-member panels, those two remaining members made up
a quorum of the delegee panel. Thus, even though the Board
itself was paralyzed for want of a quorum, the Board believed
that those two remaining members could still act in the name
of the three-member panel to which they belonged and could
wield the plenary, final authority the Board had delegated to
that panel before the Board’s total membership fell below the
“at all times” three-member quorum requirement of 29 U.S.C.
§ 153(b).

In Laurel Baye, we rejected this tactic, concluding that
the plain language of the statute had a single, “unambiguous”
reading that foreclosed the Board’s interpretation. 2 564 F.3d

2
Though we did not indicate in so many words that this
conclusion constituted a Chevron step-one holding, we did without
question identify an “unambiguous” reading of the statute regarding
the Board’s authority to evade the quorum requirement on its own
activities that foreclosed any inconsistent Board interpretations.
Laurel Baye, 564 F.3d at 473; see also id. (explaining that the
statutory text “clearly require[d]” one result); id. (“Congress
provided unequivocally” that the Board’s interpretation was
 15

at 473. We held that the Board could never adjudicate unfair
labor practice charges with fewer than three active members
because the NLRA provides that the Board must satisfy the
three-member quorum requirement “at all times.” Id. at 472
(quoting 29 U.S.C. § 153(b)) (emphasis in Laurel Baye). “The
Board quorum requirement therefore must still be satisfied,
regardless of whether the Board’s authority is delegated to a
group of its members.” Id. And when the Board has delegated
its authority to a panel of three that meets its own
two-member quorum requirement, that panel has authority to
act on behalf of the Board only so long as the Board has
authority to act as well—that is, when the Board has at least
three members. “[T]he Board cannot by delegating its
authority circumvent the statutory Board quorum requirement,
because this requirement must always be satisfied.” Id. at 473.
“The delegee group’s delegated power to act [therefore]
ceases when the Board’s membership dips below the Board
quorum of three members.” Id. at 475. The Board’s
approach—allowing a delegee panel of two members to wield
plenary Board authority even if those two members comprised
the entire active Board membership—would “allow the Board
to reduce its operative quorum to two without further
congressional authorization.” Id. Thus we found that under
the plain meaning of the statute a three-member delegee panel
acting “on behalf of the Board” may only do so when the
Board satisfies its quorum requirement. Id. (emphasis added).

Separately, we noted that our conclusion was bolstered
by analogy to common-law agency principles. According to

foreclosed); id. at 475 (explaining that the delegated authority of a
Board panel was “necessarily limited” by the NLRA’s quorum
requirement); id. at 476 (noting that “[a]ny change to the statutory
structure must come from the Congress, not the courts”).
 16

the Restatement (Third) of Agency, we explained, “an agent’s
delegated authority terminates when the powers belonging to
the entity that bestowed the authority are suspended” and “is
also deemed to cease upon the resignation or termination of
the delegating authority.” Laurel Baye, 564 F.3d at 473. “The
statute confers no authority” on a delegee panel, but only
“permits” the Board to create such a committee. Id. “The only
authority by which the committee can act is that of the
Board,” and “[i]f the Board has no authority, it follows that
the committee” cannot exercise final authority in the place of
the Board, either. Id.

The Supreme Court took up the same question regarding
the Board’s authority to act without a quorum in New Process
Steel, L.P. v. NLRB, 560 U.S. 674 (2010), and reached a
similar conclusion based on different reasoning. The Supreme
Court’s interpretation of the plain terms of the statute led it to
conclude that the Board’s powers must “be vested at all times
in a group of at least three members.” Id. at 680. 3

3
The dissent suggests that we may never apply Chevron
deference to interpretations of section 3(b) because the Court did
not rely on Chevron at all in New Process Steel. Dissent at 6. There
are a number of cogent explanations that might explain why the
Court did not do so. Perhaps the Court simply concluded that the
Board was due no deference under any standard because the text of
the statute decided the question. Perhaps the Court viewed the
Board’s entitlement to Chevron deference as forfeited because the
Board had neglected to request deference at the court of appeals.
Perhaps the Court believed that the Board was not entitled to
deference because it relied on an OLC memorandum rather than
interpreting the statute itself. But the dissent offers no reason at all
justifying its contrary conclusion that New Process Steel prohibits
the application of Chevron to this precise section in any future case.
We will not take the extraordinary step of removing a provision
 17

The Court specifically declined to consider the discussion
of agency law in Laurel Baye, explaining that “failure to meet
a quorum requirement [does not] necessarily establish that an
entity’s power is suspended so that it can be exercised by no
delegee. . . . [The] conclusion that the delegee group ceases to
exist once there are no longer three Board members to
constitute the group does not cast doubt on the prior
delegations of authority to nongroup members, such as the
regional directors . . . .” Id. at 684 n.4. Such prior delegations
to nongroup members involved a separate question that the
Court expressly did not address.

Laurel Baye does not control this case because it
confronted an entirely different situation based on different
statutory language and policy considerations, and neither
could have nor did reach the question we face here. Therefore
nothing in that opinion alters our conclusion that we owe
deference to the Board’s reasonable interpretation of the
statute. As noted above, in New Process Steel the Supreme
Court highlighted the distinction between the two types of
authority the Board may delegate to different actors. The
Board may delegate its plenary, final authority to decide cases
to a subgroup of its own members. It may also delegate
nonfinal authority to supervise elections, subject to review
and approval by the Board itself, to “nongroup” actors like the

from an agency’s interpretive reach without any basis for such a
holding. Of course, even if New Process Steel did somehow
foreclose Chevron deference, that unusual result would apply only
to the issue presented in that case: the authority of Board delegee
panels, not the authority of the Regional Directors. And even if
New Process Steel did somehow apply to the issue before us, we
would nonetheless be left with ambiguous statutory text.
 18

Regional Directors. New Process Steel, 560 U.S. at 684 n.4.
This distinction between forms of delegated authority is
crucial. Laurel Baye considered only whether plenary, final
authority delegated to panels of the Board’s own members
could survive when the Board had no quorum; we concluded
that the delegation in question could not survive because it
was precluded by the plain meaning of the statutory provision
in question. Here, we must consider a different question,
arising from a different clause of the statute, involving
different analytical considerations: whether the statute vitiates
nonfinal authority already delegated to individuals outside the
Board’s membership when the Board loses its quorum. As the
Supreme Court expressly acknowledged, the two questions
are distinct, and the answer to one has no necessary logical
relationship to the answer to the other.

UC Health insists, however, that the agency discussion in
Laurel Baye prohibits the Board from interpreting the statute
to authorize Regional Directors to continue acting when the
Board has no quorum. We disagree. In Laurel Baye, we
analyzed the plain language of the statute and found that it
prohibited the Board from employing the delegation at issue
there without giving any deference to the Board at all. The
discussion of agency principles in that case confirmed our
interpretation of the statute’s plain meaning. Here, because
the text is ambiguous, we must defer to the Board’s
reasonable interpretation. For that reason, Laurel Baye’s
agency analysis could only possibly be relevant here if it
rendered the Board’s interpretation unreasonable with the
same clarity that the plain language of the statute, reinforced
by the principles of agency law, foreclosed the Board’s effort
in Laurel Baye to delegate its plenary, final authority even
when it had no quorum. But there is a fundamental difference
in the nature of the authority delegated in these two cases.
Therefore we conclude that Laurel Baye’s agency discussion
 19

is simply off the mark in this case. A delegee panel, wielding
the Board’s plenary, final authority, speaks on the Board’s
behalf and in its place. The Regional Directors never similarly
occupy the Board’s role as a final decisionmaker. Indeed, the
statute and the Board’s own regulations expressly reserve for
the Board the power to review and reverse any determination
a Regional Director makes. 29 U.S.C. § 153(b); 29 C.F.R.
§ 102.67(c). 4 Therefore the statute makes clear that the
delegation at issue here does not implicate any of the concerns
that motivated us to draw on agency law in Laurel Baye.
Because the relationship between the Board and the Regional
Directors is so different from the relationship between the
Board and its delegee panels, we do not see how the agency
analysis in Laurel Baye sheds any light at all on the authority
the statute permits the Board to delegate to the Regional
Directors or when that authority expires. 5

4
The dissent claims that the right to appeal a Regional
Director’s decision is lost if the Board lacks a quorum. Dissent at 2
& n.1. This is incorrect. Even if the Board has no quorum when a
party appeals a Regional Director’s decision, nothing in the
regulation precludes the Board from taking up the objection once it
regains a quorum. See 29 C.F.R. § 102.67(c). The dissent suggests
that bad consequences might arise while appeals of Regional
Director determinations remain pending. This claim finds no
support in the record. In any event, considerations such as these
shed no light on the proper reading of the statute.
5
The dissent apparently takes issue with the fact that the
Board did not highlight this distinction in its brief. Dissent at 4. But
of course “[p]arties cannot waive the correct interpretation of the
law by failing to invoke it.” Zivotofsky ex rel. Zivotofsky v. Kerry,
135 S. Ct. 2076, 2101 n.2 (2015) (Thomas, J., concurring in the
judgment in part and dissenting in part) (citing EEOC v. FLRA, 476
U.S. 19, 23 (1986) (per curiam)).
 20

The important distinction between the final authority
delegated in Laurel Baye and the nonfinal authority delegated
here is all the more clear in light of the materials on which we
relied in our agency analysis in Laurel Baye. All the sources
we cited dealt with an agent who had authority to speak
finally on the principal’s behalf with permanent legal effect.
See, e.g., Restatement (Third) of Agency § 3.07(4) (2006)
(explaining that an agent’s actual authority to affect its
principal’s legal relations expires when the principal’s power
to act is suspended); 2 William Meade Fletcher, Fletcher
Cyclopedia of the Law of Corporations § 504 (2008) (same,
with respect to the resignation or termination of the delegating
authority); Id. § 421 (“If there are fewer than the minimum
number of directors required by statute, [the remaining
directors] cannot act as a board.”); Emerson v. Fisher, 246
F. 642, 648 (1st Cir. 1918) (holding that a corporate
treasurer’s delegee lacks authority to disburse corporate funds
after the treasurer himself resigns). In every case the cited rule
prohibits an agent from taking some final action on behalf of
its principal at a time when the principal could not act itself.
And indeed, the ability to stand in the principal’s place is
fundamental to the existence of an agency relationship at all.
See Restatement (Third) of Agency § 1.01 cmt. c (“[T]he
concept of agency posits a consensual relationship in which
one person . . . acts as a representative of or otherwise acts on
behalf of another person with power to affect the legal rights
and duties of the other person.”). For example, in Emerson, an
assistant corporate treasurer continued to sign checks drawn
on corporate funds in the treasurer’s name after the treasurer
had resigned. 246 F. at 648. The First Circuit found that the
assistant treasurer could not take final actions in the place of
the treasurer when no individual held that office; wielding the
principal’s authority and signing checks in his stead was
forbidden without a treasurer in place whose authority the
assistant could exercise by proxy.
 21

The outcome in Emerson and the issues identified in the
other authorities we cited in Laurel Baye were unmistakably
relevant to the issue we confronted in that case. There, the
Board had allowed a delegee panel to exercise the Board’s
plenary, final authority when the Board itself could not act at
all. The panel’s judgment was as final for the parties as the
full Board’s determination would have been. Allowing an
agent to act in those circumstances was at odds with basic
principles of agency law. But those principles have no bearing
here. No decision of the Regional Directors is ever final under
its own power. Only the acquiescence of the parties or the
Board’s ratification can give binding force to a Regional
Director’s determination. Therefore the Regional Directors
are not acting “on behalf of the Board” in the way that
doomed the Board’s tactic in Laurel Baye. 564 F.3d at 475
(emphasis added). And the agency analysis we expounded in
that case is not relevant here for the same reason.

In other words, a Regional Director never has the last say
on anything unless a party fails to object. In that event, it is
the parties’ choice to leave the Regional Director’s decisions
unchallenged that effectively makes the election final.
Otherwise, a Regional Director’s decision becomes final only
when approved by the Board. Of course, Board review is
discretionary even when a party files an objection to a
Regional Director’s decision. Nonetheless any objection will
always be considered by the Board, and it is the Board’s
action—declining to grant review or granting review and
upholding or reversing the Regional Director’s decision—that
finally commits the Board’s imprimatur. Obviously the Board
could not consider any objection to a Regional Director’s
determination when it did not satisfy the NLRA’s
three-member quorum requirement. But when the Board acts
with a quorum to review and approve a Regional Director’s
 22

decision, the Board, not the Regional Director, has decided
the issue. Unlike in Laurel Baye, this delegation does not
allow some other actor to stand in the Board’s place and wield
its authority when it is otherwise statutorily immobilized.
Therefore the agency principles that bolstered our statutory
conclusion in Laurel Baye are as irrelevant in this case as is
our discussion there of a different clause of the statute. 6

6
The dissent suggests that, by declining to consider or adopt
the agency analysis in Laurel Baye, the Court in New Process Steel
understood Laurel Baye to have implicitly decided the issue this
case presents. Dissent at 3. It is far from clear whether it would
make any difference if the Court actually expressed a view, in dicta,
of the implicit scope of one of our decisions. Even so, we disagree
that the Court expressed any such opinion. The Court
acknowledged, as have we, that Laurel Baye relied on statutory and
agency grounds to foreclose the Board’s agents from acting when
the Board could not act. But the Court did not explain its view of
how far Laurel Baye’s reasoning necessarily extended any more
than Laurel Baye itself offered such an explanation. And after all,
New Process Steel, like Laurel Baye, decided a different question
from the issue before us here. Whether the agency analysis in
Laurel Baye necessarily also applies to “nongroup” actors like the
Regional Directors was, the Court explained, “a separate question.”
New Process Steel, 560 U.S. at 684 n.4.
Nor do we think the dissent is correct that other circuits’
pronouncements militate in favor of extending Laurel Baye to reach
delegations to the Regional Directors—an issue not raised in that
case. Dissent at 3 & n.3. It is true that several other courts since
New Process Steel have declined to apply the broadest possible
reading of our agency analysis in Laurel Baye to the distinct
question of whether the Board’s General Counsel may continue to
exercise authority when the Board has no quorum. But the
noteworthy point is that these circuits all agree with our
fundamental conclusion: The broad, general expressions of
common law agency principles stated in the different context of
 23

Indeed, we are all the more persuaded that the Board’s
interpretation of the statute is reasonable in light of the
structural distinction between the final character of its
authority to adjudicate unfair labor practice cases and the
nonfinal authority to oversee representation elections it may
delegate to the Regional Directors. Because any contested
decision a Regional Director makes is not final until the
Board acts, it is immaterial whether the Board had a quorum
at the time the Regional Director conducted the election. To
the contrary, when the Regional Directors exercise their
delegated authority to oversee elections, they further the
policy of the statute by increasing the efficiency with which
representation elections are held—irrespective of the Board’s
status at that time.

For all these reasons, there is nothing in Laurel Baye or its
broad discussion of principles of agency law that controls the
Board’s interpretation of its authority in this case. As shown
above, the Board’s interpretation of its authority was
reasonable, and we are bound to defer to the Board’s
reasonable interpretation of the statute it is charged to
administer. See City of Arlington, 133 S. Ct. at 1870-71.

Laurel Baye are not persuasive in the context of non-Board
delegees. We have explained why the fundamental differences
between the authority delegated in Laurel Baye and the authority
delegated to the Regional Directors here render Laurel Baye
irrelevant. See Restatement (Third) of Agency § 1.01 cmt. c (“[T]he
concept of agency posits a consensual relationship in which one
person . . . acts as a representative of or otherwise acts on behalf of
another person with power to affect the legal rights and duties of
the other person.”). Neither New Process Steel nor the circuit cases
cited in the dissent erase that structural distinction. In fact, they are
all best read to support it.
 24

III

For the foregoing reasons, we deny UC Health’s petition
for review and grant the NLRB’s cross-application for
enforcement of its order.
 EDWARDS, Senior Circuit Judge, concurring: The dissent
is mistaken in suggesting that if the rationale or logic
supporting a decision in one case is stated broadly enough to
cover future cases not at issue, the latter cases are necessarily
controlled by the earlier case. Were this the law, appellate
decisionmaking would be a mischievous enterprise.

****

It is well understood that “[t]he doctrine of stare decisis
is of fundamental importance to the rule of law.” Welch v.
Tex. Dep’t of Highways & Pub. Transp., 483 U.S. 468, 494
(1987). Stare decisis – to stand by things decided – embraces
the principle that each judicial decision is a statement of law
(or precedent) that may have binding force in future cases.
“Stare decisis . . . ‘promotes the evenhanded, predictable, and
consistent development of legal principles, fosters reliance on
judicial decisions, and contributes to the actual and perceived
integrity of the judicial process.’” Hohn v. United States, 524
U.S. 236, 251 (1998) (quoting Payne v. Tennessee, 501 U.S.
808, 827 (1991)). “For judges, the most basic principle of
jurisprudence is that we must act alike in all cases of like
nature” because “[i]nconsistency is the antithesis of the rule
of law.” LaShawn A. v. Barry, 87 F.3d 1389, 1393 (D.C. Cir.
1996) (en banc) (internal quotation marks omitted).

A judicial decision is viewed as “precedent” when it
controls the disposition of a pending case. Whenever a court
faces a situation in which a prior judicial decision has some
similarity to a pending case, the judges must initially
determine the rule established by the decision in the first case,
limited by the context in which the judgment was reached.
Then the judges must determine whether the rule of the prior
case controls the disposition of the pending case. It is easy to
subscribe to the goal of stare decisis. It is not always easy,
however, to determine when a prior case qualifies as
controlling precedent. See Jeremy Waldron, Stare Decisis and
 2
the Rule of Law: A Layered Approach, 111 MICH. L. REV. 1
(2012) (discussing applications of stare decisis); Ruggero J.
Aldisert, Precedent: What It Is and What It Isn’t; When Do
We Kiss It and When Do We Kill It?, 17 PEPPERDINE L. REV.
605 (1990) (same). Nonetheless, there are several important
principles that the courts routinely follow in determining the
applicability of precedent to the cases before them.

First, “an issue of law must have been heard and decided”
in the same or a higher court for a decision to have
precedential value with respect to that issue. Gately v.
Massachusetts, 2 F.3d 1221, 1226 (1st Cir. 1993) (quoting
EEOC v. Trabucco, 791 F.2d 1, 4 (1st Cir. 1986)). Second, “if
an issue is not argued, or though argued is ignored by the
court, or is reserved, the decision does not constitute a
precedent to be followed” with respect to that issue. Id. Third,
a judicial decision “attaches a specific legal consequence to a
detailed set of facts.” Allegheny Gen. Hosp. v. NLRB, 608
F.2d 965, 969–70 (3d Cir. 1979). The decision “is then
considered as furnishing the rule for the determination of a
subsequent case involving identical or similar material facts.”
Id.; see also United States v. Holyfield, 703 F.3d 1173, 1177
& n.7 (10th Cir. 2013) (citing Allegheny’s definition of
precedent approvingly). Fourth, as Chief Justice Marshall
explained in his seminal opinion in Cohens v. Virginia, 19
U.S. (6 Wheat) 264 (1821):

It is a maxim, not to be disregarded, that general
expressions, in every opinion, are to be taken in
connection with the case in which those expressions are
used. If they go beyond the case, they may be respected,
but ought not to control the judgment in a subsequent
suit, when the very point is presented for decision. The
reason of this maxim is obvious. The question actually
before the court is investigated with care, and considered
 3
in its full extent. Other principles which may serve to
illustrate it, are considered in their relation to the case
decided, but their possible bearing on all other cases is
seldom completely investigated.

Id. at 399–400.

This last precept – that the force of a general expression
enunciated in a prior decision must be limited by reference to
its specific context – is so firmly embedded in stare decisis
jurisprudence that the Supreme Court has called it a “canon of
unquestionable vitality.” Landgraf v. USI Film Prods., 511
U.S. 244, 265 (1994). Indeed, the Court has said that it is the
“duty” of judges “to restrict general expressions in opinions in
earlier cases to their specific context.” Int’l Bhd. of Teamsters
Local 309 v. Hanke, 339 U.S. 470, 480 n.6 (1950) (plurality
opinion) (emphasis added). See also, e.g., King v. Morton,
520 F.2d 1140, 1147 (D.C. Cir. 1975) (“Chief Justice
Marshall warned against basing decisions on bare general
principles enunciated in other cases. . . . The simple words of
the opinions [cited by appellant] are not as important as the
contexts in which those cases were decided.”).

The obvious point is that the precedential value of a
decision is defined by the context of the case from which it
arose. If, in light of that context, the decided case is materially
or meaningfully different from a superficially similar later
case, the holding of the earlier case cannot control the latter.

Determining the proper scope of the rule of a prior
decision can be controversial. In his illuminating article on
“Stare Decisis and the Rule of Law,” Professor Waldron
points out that “[l]egal realists and critics are fond of”
accusing judicial panels of formulating the rule of prior
decisions as they see fit “in order to suit [their] own view[s]
 4
about how the case in front of [them] should be decided.”
Waldron, supra, at 26. He urges that, to avoid this pitfall in
decisionmaking, judicial panels should be ever mindful that
the rule of law commands them to view precedent “in a
responsible spirit of deference.” Id. I agree.

However, as Professor Waldron notes, the ascertainment
of the rule of a prior case and the determination whether the
prior case constitutes a binding precedent that controls the
disposition of a pending case are nuanced enterprises.

One case may seem superficially similar to another, but
the judge[s on the second panel] may be convinced that
there are differences that preclude simply subjecting a
subsequent case to the same rule that decided the
precedent case. . . . For example, a given statutory
provision may apply properly to one case but not another,
even though the second is superficially similar to the
first; therefore, we “distinguish” the second case. And
similarly, the rule that [the first panel] figured out as a
basis for [its] decision in the precedent case may not
apply to a subsequent case despite superficial similarities.
There may be things about the second case that pose a
distinct legal problem, which require a new and distinct
law-like solution to be figured out by [the second panel]
in the form of a rule . . . . To distinguish a case, then, is
not just to “come up with” some difference. It is to show
that the logic of what [the first panel] figured out does
not, despite appearances, apply. It means pointing to
some additional problematic feature of the subsequent
case that requires additional figuring.

Id. at 25–26. Courts routinely follow these principles of
precedent application. Looking to the context of the putative
controlling decision – the facts, the statutory or constitutional
 5
provisions at issue, the arguments made by the parties and
decided or reserved by the court, and the rationale underlying
the decision – they determine whether the holding of an
existing case controls the outcome of a pending case. See,
e.g., Humphrey’s Executor v. United States, 295 U.S. 602,
626–28 (1935) (distinguishing the pending case from the cited
precedent because the situations were “so essentially unlike”
each other); United States v. Thomas, 361 F.3d 653, 662–63
(D.C. Cir. 2004) (applying precedent selectively to two
defendants’ cases based on factual similarities and
dissimilarities to the prior case), reinstated following vacatur
sub nom. United States v. Cook, 161 Fed. App’x 7 (D.C. Cir.
2005).

****

Here, the considerations guiding the application of
precedent make clear that, although Laurel Baye Healthcare
of Lake Lanier, Inc. v. NLRB, 564 F.3d 469 (D.C. Cir. 2009),
is “superficially similar” in some respects to the case
presently before the court, it does not control the resolution of
the legal question presented in this case. First, as Judge
Griffith’s opinion explains, the facts of Laurel Baye are very
different from the facts of this case. And facts matter in
determining the precedential value of a prior case. See, e.g.,
Armour & Co. v. Wantock, 323 U.S. 126, 132–33 (1944)
(“[W]ords of our opinions are to be read in the light of the
facts of the case under discussion. . . . General expressions
transposed to other facts are often misleading.”). Second, just
as significantly, Laurel Baye did not involve the statutory and
regulatory provisions that are principally at issue in this case.
Finally, and most importantly, Laurel Baye did not in any way
address the question of the deference due the Board’s
construction of either the particular provisions of the statute at
issue here or any other provisions of the Act. This appears to
 6
be because the Board never offered an interpretation of the
Act for which deference was sought.

We face a very different situation in this case than the
situation faced by the court in Laurel Baye. Here, the Board
has offered a reasonable interpretation of statutory language
in the Act – statutory language different from the statutory
provisions at issue in Laurel Baye. And as Judge Griffith
explains, the Board’s interpretation is one to which we must
defer pursuant to the firmly established principles enunciated
by the Supreme Court in Chevron U.S.A. Inc. v. Natural
Resources Defense Council, Inc., 467 U.S. 837, 843 (1984).
Alternatively, as Judge Srinivasan contends, see SSC Mystic
Operating Co. v. NLRB, No. 14-1045 (D.C. Cir. September
__, 2015) (Srinivasan, J., concurring), we must defer to the
Board because a court’s prior judicial construction of a statute
does not trump an agency construction that is otherwise
entitled to deference under Chevron unless “the prior court
decision holds that its construction follows from the
unambiguous terms of the statute and thus leaves no room for
agency discretion.” Nat’l Cable & Telecomms. Ass’n v. Brand
X Internet Servs., 545 U.S. 967, 982 (2005). Under either
view, Laurel Baye is not binding precedent here because the
relevant terms of the statute at issue in this case are
unquestionably ambiguous.

In sum, the decision in Laurel Baye does not control the
judgment in this case. This case poses what Professor
Waldron would call a “distinct legal problem” that was
neither raised by the parties nor addressed by the court in
Laurel Baye. Consequently, “[a]ny observations [from that
opinion] which could be regarded as having a bearing upon
the question now before us would be taken out of their proper
relation.” Wright v. United States, 302 U.S. 583, 593 (1938);
see also Whitacre v. Davey, 890 F.2d 1168, 1172 (D.C. Cir.
 7
1989) (“We cannot count as controlling a decision that never
touched upon the issue we confront” when that point “was
simply not considered” in the prior case). The general
expressions from Laurel Baye to which the dissent refers
cannot be confused with binding precedent. To rely on these
general expressions, taken from a case whose context is
materially different from the case before us, flies in the face
of the core principles of stare decisis. See Weyerhauser v.
Hoyt, 219 U.S. 380, 394 (1911) (“[G]eneral language” used in
a prior opinion should not be “separated from its context and
disassociated from the issues which the case involved” and
then given controlling weight.).

The dissent suggests that we have betrayed “the most
important characteristic of a collegial appellate court” by
failing to give “careful attention [and] respect” to the law of
the circuit. It is hard to take this claim seriously because it is
premised on misguided notions of stare decisis. I therefore
view the dissent’s unfortunate statement as nothing more than
a poignant example of hyperbole.
 SILBERMAN, Senior Circuit Judge, dissenting: The merits
of this case are not particularly important. I doubt whether we
will see many situations in which an NLRB regional director
certified an election during a period in which the NLRB lost its
quorum, and that certification is subsequently challenged in an
unfair labor practice proceeding. But, the case is nevertheless
of great significance because the most important characteristic
of a collegial appellate court is careful attention, respect, and
adherence to precedent. I am afraid the majority opinion is a
glaring example of a contrary approach.

I have previously authored an opinion in which I was faced
with conflicting lines of authority because of my colleagues’
failure to follow prior precedent, see Vietnam Veterans of
America v. Shinseki, 599 F.3d 654 (D.C. Cir. 2010) (Silberman,
J.), which I regarded then, and do now, as most unfortunate. Of
course, every case with which we are presented has some factual
difference, but not every case is legally distinguishable from
every other. In deciding whether a new case is covered by
previous precedent, it is the logic of the previous case’s holding
that is determinative. Petitioner’s brief elegantly argued,
succinctly, that we were bound by our prior decision in Laurel
Baye Healthcare of Lake Lanier, Inc. v. NLRB, 564 F.3d 469
(D.C. Cir. 2009). I think petitioner is exactly correct and that
should be the end of the matter.

I.

The majority concludes that section 3(b) of the National
Labor Relations Act, authorizing the Board to delegate to
regional directors – civil servants, not political appointees –
power to certify the results of an election, means that that power
remains indefinitely in the regional director, even if the Board
itself disappears (perhaps, because of a failure of the Senate to
confirm any appointee). This power supposedly resides
permanently in the regional director, even though the statute
specifically provides that any interested party can appeal a
 2

regional director’s determination to the Board. And if the Board
loses its quorum, or goes out of existence, that appeal right is
lost.1

But even if a Board were to read its regulation to permit an
appeal of a regional director’s decision out of time – perhaps
years later when a Board regained a quorum – the majority
ignores the impact of a regional director’s certification in the
interim. An employer who refuses to recognize that certification
may pay a price in labor relations. And the regional director’s
decision with respect to issues such as objectional conduct may
deviate from Board policy.

As I discuss below, I think this construction would be a
stretch, even if we were considering section 3(b) de novo, but
we are not. We are bound by our prior opinion in Laurel Baye.
In that case, we considered whether a Board delegation of all of
its power to a subgroup of only three survived the drop in
membership of the subgroup (and the Board) to only two.
Although the statutory language is convoluted, we relied on two
factors to conclude that the Board’s powers had lapsed. We
focused first on the most persuasive language – “[t]he quorum
provision clearly requires that a quorum of the Board is, ‘at all
times,’ three members,” id. at 473 (citing 29 U.S.C. § 153(b)) –
but we also relied on general agency principles reasoning
broadly that an agent – in that case, the two member Board
panel – lost its authority when the Board – the principal – lost its
quorum. We never mentioned Chevron, implicitly concluding
that the combination of the two factors ineluctably pointed to
only one interpretation. As Judge Sentelle writes in the

1
A party must seek Board review within 14 days of the regional
director’s issuance of a final disposition. 29 C.F.R. § 102.67(c) (as
amended Dec. 15, 2014). The majority’s reading of the regulation
appears incorrect.
 3

companion case, “Laurel Baye concluded that § 153(b)’s
quorum requirement provision unambiguously requires the
Board to have a quorum for a delegee to exercise its authority.”
SSC Mystic Operating Co., Slip op. at 2 (Sentelle, J.,
dissenting). The Supreme Court, reviewing the same issue in
New Process Steel v. NLRB, came to the same conclusion, that
the Board had lost its authority, but it relied primarily on
different language in the section.2 560 U.S. 674, 679-83 (2010).
It explicitly did not adopt the general agency principles so
important to our reasoning, recognizing that to do so would
decide the very issue now before us – whether a regional
director’s authority survives the loss of Board membership. Id.
at 684 n.4. The Supreme Court wanted that issue to remain open
in its court, but – and this is the crucial point – the Court
implicitly recognized that it did not remain open in our court.
Moreover, even though our sister circuits declined to adopt
Laurel Baye, they have uniformly read it as did the Supreme
Court: as having decided the validity of board delegations to
nonmembers in the absence of a quorum.3 Our agency

2
Although the Supreme Court’s footnote four is at least clear in not
adopting Laurel Baye’s agency rationale, its reasoning explaining why
it does not is impenetrable.
3
See Kreisberg v. HealthBridge Mgmt., LLC, 732 F.3d 131, 140 (2d
Cir. 2013); Frankl v. HTH Corp., 650 F.3d 1334, 1354 (9th Cir.
2011); Osthus v. Whitesell Corp., 639 F.3d 841, 844 (8th Cir. 2011);
Overstreet v. El Paso Disposal, L.P., 625 F.3d 844, 852-54 (5th Cir.
2010) (“[Laurel Baye] held that when the Board’s membership drops
to two, it loses its quorum and ‘[i]n the context of a board-like entity,
a delegee’s authority therefore ceases the moment the vacancies or
disqualifications on the board reduce the board’s membership below
a quorum.’”). These circuit cases deal with the General Counsel of the
Board’s authority to continue to seek 10(j) injunctions when the Board
loses its quorum. Board delegations of duties to the General Counsel
are governed by a separate statutory provision. See 29 U.S.C.
 4

reasoning was not, as the majority puts it, “separate,” it was
integral to Laurel Baye’s statutory interpretation, and therefore
applies equally to our case. Indeed, it is a fortiori because we
are dealing with a delegation, not to a subgroup of Board
members, but rather to a much lesser-ranked official, a regional
director.

Notwithstanding the Supreme Court’s interpretation of
Laurel Baye, the majority would distinguish that case on the
theory that a delegation to an agent with lesser authority (a
regional director) somehow survives the disappearance of the
principal, even though a more senior agent’s authority would
not. It should be noted, the majority does not cite any case so
holding. To give the government its due, the Board never makes
this cockamamie argument (I thought we avoided relying on
arguments not presented by parties because to do so suggests
we are result-oriented). In any event, I do not understand why
the regional director’s authority is described as “non-final” when
the very issue in this case is whether the regional director’s
certification of an election is final.

II.

As I noted above, I think that if we were not bound by
Laurel Baye, I would still regard the Board’s contention that the
regional director’s authority to certify an election extends
indefinitely, even if the Board went out of existence for years,
as an impermissible construction of the statute.

§ 153(d). The General Counsel – a Presidential appointee – has
unreviewable discretion to seek a complaint whether or not the Board
is in existence. Therefore it is reasonable to conclude that the Board’s
permanent delegation to the General Counsel to seek an injunction is
simply an addition to his or her authority to file a complaint.
 5

The majority acknowledges the statute does not address the
question whether the regional director’s power lapses when the
Board loses its quorum or ceases to exist, but contends that that
is a statutory “silence” under the Chevron doctrine, which the
NLRB is authorized to fill. And it asserts that the Board’s
interpretation is reasonable under Chevron’s second step.
Actually, if Chevron applied, I think such an interpretation
would be flatly unreasonable. We must bear in mind that even
if we are following Chevron’s second step, we are construing a
Congressional act – the second step is not open sesame for the
Agency. After all, the rest of that statutory section deals
specifically, although in a rather convoluted fashion, with the
circumstances in which the Board loses authority because of a
lack of a quorum. It is quite incredible that Congress would
nevertheless have implicitly bestowed on regional directors
permanent authority ad infinitum, even in the total absence of a
supervising Board. It should be noted that neither the majority
opinion nor Judge Edwards’s and Judge Srinivasan’s
concurrences grapple with the statutory structure or language to
reach their conclusion that the interpretation they adopt is
actually “reasonable.”

Be that as it may, we cannot affirm the Board based on
Chevron deference in this case. The Board never purported to
interpret an ambiguity in the statute. Instead, it boldly asserted
that, pursuant to the statute, “NLRB Regional Directors remain
vested with the authority to conduct elections and certify their
results regardless of the Board’s composition at any given
moment.” UC Health & UC Health Pub. Safety Union, 360
N.L.R.B. No. 71, *1 n.2 (Mar. 31, 2014). We have held
repeatedly that “[d]eference to an agency’s statutory
interpretation is only appropriate when the agency has exercised
its own judgment, not when it believes that interpretation is
compelled by Congress.” Arizona v. Thompson, 281 F.3d 248,
254 (D.C. Cir. 2002) (internal citations and emphasis omitted).
 6

See also Peter Pan Bus Lines, Inc. v. Fed. Motor Carrier Safety
Admin., 471 F.3d 1350, 1354-55 (D.C. Cir. 2006); Transitional
Hosps. Corp. v. Shalala, 222 F.3d 1019, 1029 (D.C. Cir. 2000).
The majority argues that the Board never explicitly stated that
the statute compelled the construction, but what else can the
Board have intended when it flatly states what the statute
meant.4

Moreover, the Supreme Court has rejected Chevron’s
applicability to section 3(b). Even though the language the
Supreme Court relied on, as I noted, is rather convoluted, the
Supreme Court’s opinion in New Process Steel never mentioned
Chevron – despite the government’s reliance on Chevron
deference in its Supreme Court brief. Although the Court’s
opinion frankly acknowledged two possible interpretations of
what it called section 3(b)’s delegation clause, it simply picked
the one it thought preferable – leading to the same result we
chose in Laurel Baye. See New Process Steel, 560 U.S. at 679-
83. It is, therefore, decisive, for our purposes, that the Court
implicitly but necessarily concluded that, for whatever reason,
Chevron deference was inappropriate in construing section 3(b).
See SSC Mystic Operating Co., Slip op. at 2–3 (Sentelle, J.,
dissenting). In that regard, the other circuits, construing the
companion language dealing with the General Counsel’s
delegated authority to seek 10(j) injunctions, have followed the
Supreme Court’s lead and have ignored Chevron.5

***

4
If the majority was correct in concluding that Chevron applied, it
certainly should remand and allow the agency to exercise its judgment.
See PDK Labs. Inc. v. U.S. DEA, 362 F. 3d 786, 798 (D.C. Cir. 2004).
5
With the exception of one concurring Eighth Circuit Judge. See
Osthus, 639 F.3d at 845-48 (Colloton, J. concurring).
 7

Judge Srinivasan’s concurrence in the companion case
disagrees with the majority’s conclusion that the panel’s
decision in Laurel Baye was based on an unambiguous reading
of the plain language, and chooses to read Laurel Baye as only
adopting the “best” reading of an ambiguous statute.6 By so
doing, Judge Srinivasan essentially accuses the Laurel Baye
panel of disregarding governing law applying to judicial review
of agency statutory interpretations in formal adjudication. That
governing law, Chevron – with certain specific exceptions, such
as avoidance of serious constitutional issues7 – for over thirty
years has banned courts of appeal from doing exactly what
Judge Srinivasan accuses the Laurel Baye panel of doing;
rejecting an agency statutory interpretation of supposedly
ambiguous language in favor of what a reviewing court believes
is a better or best reading.

Judge Srinivasan relies on Brand X as support for his
analysis – suggesting that it allows a court reviewing agency
interpretation of ambiguous language to choose the better
interpretation. But that is a flagrant misreading of the case.
Brand X applies in situations quite apart from Laurel Baye. The
first is the question of how do reviewing courts deal with a pre-

6
Judge Srinivasan ignores the fact that the Supreme Court’s decision -
both the majority and the dissent as Judge Sentelle points out, see SSC
Mystic Operating Co., Slip op. at 2 (Sentelle, J., dissenting) – ignored
Chevron, or its doctrine, in deciding New Process Steel – which
certainly indicates, as I have argued above, that Chevron is not to be
used in interpreting section 3(b), as Judge Srinivasan does. Granted,
the Supreme Court in New Process Steel did ignore Chevron without
explanation, but it is not subject to the same restraints which bind
lower federal courts.
7
See Edward J. DeBartolo Corp. v. Florida Gulf Coast Building &
Construction Trades Council, 485 U.S. 568 (1988).
 8

Chevron judicial decision if the agency subsequently disagrees.
The Supreme Court explained that if a prior judicial decision
announced the only acceptable interpretation of a statute that
opinion governed, but if the earlier judicial opinion – properly
read – only relied on a better interpretation, the agency was free
to adopt a different reasonable construction. The second
situation involves a judicial affirmance of a prior agency
interpretation. Again, the agency can change its interpretation
to another reasonable one if the prior judicial opinion was
limited to accepting the agency interpretation as reasonable. The
third situation, as occurred in Brand X itself, is when an agency
departs from a prior judicial decision that had reviewed the
interpretation of an entity not entitled to Chevron deference.8
Because the previous court needn’t have deferred under
Chevron, it may have opted for the “best,” but not the only
permissible, interpretation. The agency may thus later adopt
another reasonable construction. None of those situations apply
here. Brand X hardly suggested that in the future courts may
reject a federal agency statutory interpretation as not the “best
one,” which would be inconsistent with Chevron.

Judge Srinivasan suggests, although he does not quite
assert, that significance should be placed on the fact that the
NLRB did not seek Chevron deference in Laurel Baye. But that
is often true when an agency believes its interpretation is
compelled by the statute. Indeed, in Laurel Baye, as in many
other cases, the agency flatly asserted that its reading resulted

8
Brand X reviewed an agency interpretation that departed from a
previous judicial construction in AT&T Corporation v. City of
Portland, 216 F.3d 871 (9th Cir. 2000). AT&T itself did not apply
Chevron in its analysis – and opted for the best, not the only,
interpretation – because it was reviewing the construction of a local,
municipal franchising board (an entity to which Chevron deference
was not owed).
 9

from the plain language of the statute, rather than an exercise in
discretion. That is not a hypothesis, as Judge Srinivasan
suggests; it is the obvious reading of the Board’s decision. And
as Judge Sentelle points out, because the statute was thought by
the panel to be unambiguous, Chevron was irrelevant. The
important principle of administrative law is that federal courts
of appeals for almost 30 years have followed Chevron’s
command (with certain recognized exceptions) that agency
interpretations of ambiguous language in a formal adjudication,
see United States v. Mead Corp., 533 U.S. 218 (2001), are
entitled to deference and affirmance if reasonable. That is so –
and there are no deviations in our cases – notwithstanding
whether Chevron’s familiar two-step analysis is explicitly
“walk[ed] through,” see SSC Mystic Operating Co., Slip op. at
3 (Srinivasan, J., concurring). If we think, notwithstanding the
agency’s claims, that a statute is actually ambiguous, we are not
free to disregard Chevron and opt for our own “best” reading, as
Judge Srinivasan seems to suggest we can. We must instead
remand to an agency for its subsequent resolution of the
ambiguity. See PDK Labs. Inc. v. U.S. DEA, 362 F. 3d 786, 798
(D.C. Cir. 2004). Chevron is as much a principle of judicial
review of agency action as is Chenery9 or Vermont Yankee.10

Judge Srinivasan searches for another reason Chevron may
not have been applied. He suggests that the Laurel Baye panel
may have secretly denied deference based on an argument made
by the company in Laurel Baye – that Chevron deference was
inappropriate because the issue went to the Board’s jurisdiction.
To be sure, this was a question which once troubled panels of
this court. See New York Shipping Ass’n, Inc. v. Federal

9
See SEC v. Chenery Corp.,332 U.S. 194 (1947).
10
See Vermont Yankee Nuclear Power Corp. v. Natural Resources
Defense Council, 435 U.S. 519 (1978).
 10

Maritime Comm’n, 854 F.2d 1338, 1363 (D.C. Cir. 1988); Nat’l
Wildlife Fed’n v. ICC, 850 F.2d 694, 699 (D.C. Cir. 1988);
American Civil Liberties Union v. FCC, 823 F.2d 1554, 1567
n.32 (D.C. Cir. 1987). But since 1990 we have consistently
rejected that concept as an exception to Chevron.11 See
Connecticut Dep’t of Pub. Util. Control v. FERC, 569 F.3d 477,
481 (D.C. Circ. 2009); Nat’l Ass’n of Regulatory Util. Comm’rs
v. FERC, 475 F.3d 1277, 1279 (D.C. Cir. 2007); Detroit Edison
Co. v. FERC, 334 F.3d 48, 53 (D.C. Cir. 2003); Transmission
Access Policy Study Group v. FERC, 225 F.3d 667, 694 (D.C.
Cir. 2000); Oklahoma Natural Gas Co. v. FERC, 28 F.3d 1281,
1284 (D.C. Cir. 1994); Bus. Roundtable v. SEC, 905 F.2d 406,
408 (D.C. Cir. 1990). The cases cited by Judge Srinivasan, with
all due respect for the University of Illinois Law Review, do not
stand for the contrary proposition; they don’t even mention
agency “jurisdiction.” It is flatly inconceivable that any panel
would accept Judge Srinivasan’s hypothetical reasoning sub
silentio.12

In sum, since Laurel Baye failed to accept the agency’s
interpretation, it must have been because it determined that the
language was susceptible of only one meaning. It is the only
explanation of Laurel Baye consistent with both Supreme Court
commands and our own precedent.

***

11
Of course, the Supreme Court subsequently put the issue to rest in
City of Arlington v. FCC, 133 S. Ct. 1863 (2013).
12
It is worth noting that the Laurel Baye panel members Judges
Williams, Sentelle, and Tatel had, collectively, nearly 60 years of
experience reviewing agency statutory interpretations.
 11

Three of my colleagues have explored reasons not to
follow Laurel Baye. I think they are all unpersuasive; Laurel
Baye is binding precedent.

Regretfully I dissent.