Notice: This opinion is subject to formal revision before publication in the
Federal Reporter or U.S.App.D.C. Reports. Users are requested to notify
the Clerk of any formal errors in order that corrections may be made
before the bound volumes go to press.
United States Court of Appeals
FOR THE DISTRICT OF COLUMBIA CIRCUIT
Argued March 14, 2003 Decided June 20, 2003
No. 02-5069
AMERICAN FEDERATION OF LABOR AND
CONGRESS OF INDUSTRIAL ORGANIZATIONS AND
DNC SERVICES CORPORATION, DEMOCRATIC
NATIONAL COMMITTEE,
APPELLEES
v.
FEDERAL ELECTION COMMISSION,
APPELLANT
–————
Appeal from the United States District Court
for the District of Columbia
(No. 01cv01522)
–————
David B. Kolker, Attorney, Federal Election Commission,
argued the cause for appellant. With him on the briefs was
Richard B. Bader, Associate General Counsel.
Trevor Potter, Lisa J. Danetz and Lawrence M. Noble were
on the brief for amicus curiae Campaign and Media Legal
Center, et al. in support of appellant.
Bills of costs must be filed within 14 days after entry of judgment.
The court looks with disfavor upon motions to file bills of costs out
of time.
2
Laurence E. Gold argued the cause for appellees. With
him on the brief were Joseph E. Sandler and Michael B.
Trister.
James Bopp, Jr. and Raeanna S. Moore were on the brief
for amicus curiae James Madison Center for Free Speech in
support of appellees.
Before: SENTELLE, HENDERSON and TATEL, Circuit Judges.
Opinion for the Court filed by Circuit Judge TATEL.
Opinion concurring in the judgment filed by Circuit Judge
HENDERSON.
TATEL, Circuit Judge: Unique among federal administra-
tive agencies, the Federal Election Commission has as its sole
purpose the regulation of core constitutionally protected ac-
tivity—‘‘the behavior of individuals and groups only insofar as
they act, speak and associate for political purposes.’’ FEC v.
Machinists Non-Partisan Political League, 655 F.2d 380, 387
(D.C. Cir. 1981). As a result, Commission investigations into
alleged election law violations frequently involve subpoenaing
materials of a ‘‘delicate nature TTT represent[ing] the very
heart of the organism which the first amendment was intend-
ed to nurture and protect: political expression and association
concerning federal elections and officeholding.’’ Id. at 388.
At the close of such investigations, a Commission regulation
has long required public release of all investigatory file
materials not exempted by the Freedom of Information Act.
In this case, the subjects of a now-closed investigation chal-
lenge the regulation as inconsistent with both the Federal
Election Campaign Act and the First Amendment. We hold
that the regulation, though not contrary to the plain language
of the statute, is nevertheless impermissible because it fails to
account for the substantial First Amendment interests impli-
cated in releasing political groups’ strategic documents and
other internal materials.
I.
The Federal Election Commission’s administrative enforce-
ment procedures are governed by 2 U.S.C. § 437g(a) of the
3
Federal Election Campaign Act (FECA). When the Commis-
sion receives a sworn complaint alleging that an election law
violation has occurred, it must first notify the alleged violator
and give it an opportunity to respond to the accusation. If
four Commission members find ‘‘reason to believe’’ that the
respondent has committed or is about to commit a violation,
the Commission must proceed with an investigation. 2 U.S.C.
§ 437g(a)(1), (2). If the Commission then finds ‘‘probable
cause’’ to believe that a violation has occurred, it must
attempt to reach an informal conciliation agreement with the
respondent. Id. § 437g(a)(4)(A)(i). The Commission may
bring a civil enforcement proceeding in U.S. District Court if
conciliation negotiations fail. Id. § 437g(a)(6). Where the
Commission decides to dismiss a complaint at any stage of the
process, however, the statute allows ‘‘aggrieved’’ parties to
challenge the dismissal in U.S. District Court. Id.
§ 437g(a)(8).
Two parts of section 437g(a) directly address confidentiality
and disclosure of enforcement-related information. Subsec-
tion (a)(12)(A)—the provision at issue in this case—states that
‘‘[a]ny notification or investigation made under this section
shall not be made public by the Commission or by any person
without the written consent of the person receiving such
notification or the person with respect to whom such investi-
gation is made.’’ Id. § 437g(a)(12)(A). Subsection (a)(4)(B)
addresses disclosures in post-investigation proceedings:
(i) No action by the Commission or any person, and no
information derived, in connection with any conciliation
attempt by the Commission TTT may be made public by
the Commission without the written consent of the re-
spondent and the Commission.
(ii) If a conciliation agreement is agreed upon by the
Commission and the respondent, the Commission shall
make public any conciliation agreement signed by both
the Commission and the respondent. If the Commission
makes a determination that a person has not violated TTT
4
[election laws], the Commission shall make public such
determination.
Id. § 437g(a)(4)(B).
The Commission promulgated two regulations implement-
ing these provisions, 11 C.F.R. §§ 111.20, 111.21, plus a third
that reconciles FECA with the Freedom of Information Act, 5
U.S.C. § 552. The latter regulation, 11 C.F.R. § 5.4(a)(4),
requires the disclosure of investigatory file materials in closed
cases:
Opinions of Commissioners rendered in enforcement
cases and General Counsel’s Reports and non-exempt 2
U.S.C. 437g investigatory materials shall be placed on
the public record of the Agency no later than 30 days
from the date on which all respondents are notified that
the Commission has voted to close such an enforcement
file.
11 C.F.R. § 5.4(a)(4).
The roots of this case reach back to the mid-1990s when the
National Republican Senatorial Committee, the National Re-
publican Congressional Committee, and an independent politi-
cal action committee chaired by Oliver North filed eleven
complaints with the Commission alleging, among other things,
that the AFL-CIO and various individual unions had unlaw-
fully coordinated their ‘‘Labor ’96’’ campaign expenditures
with political candidates and party committees. Finding
‘‘reason to believe’’ that FECA violations had occurred, the
Commission embarked on a three-year investigation during
which it subpoenaed approximately 50,000 pages of docu-
ments from the AFL-CIO, the Democratic National Commit-
tee (DNC), and 150 other respondents, as well as third-party
witnesses. The AFL-CIO turned over documents containing
detailed descriptions of meetings with elected officials, train-
ing programs for union activists, convention and get-out-the-
vote activities, and polling data analyzing union members’
political attitudes and the effectiveness of particular political
messages. The DNC provided memoranda concerning inter-
nal deliberations between state and national party leaders, as
well as ‘‘Coordinated Campaign’’ plans describing state offi-
5
cials’ strategies, techniques, and timetables for winning up-
coming elections across the country.
After the U.S. District Court for the District of Columbia
issued a decision narrowing the circumstances under which
the Commission could regulate coordination practices under
FECA, FEC v. Christian Coalition, 52 F. Supp. 2d 45
(D.D.C. 1999), the Commission dismissed the complaints in
this case. At that time, Commission investigators had yet to
review an estimated 10,000 to 20,000 pages of the materials
gathered during the course of the proceedings. FEC General
Counsel’s Report, In re AFL-CIO, et al., at 11 n.6 (June 12,
2000). None of the complainants sought judicial review of the
dismissal under 2 U.S.C. § 437g(a)(8).
Pursuant to 11 C.F.R. § 5.4(a)(4), the Commission then
redacted certain FOIA-exempt materials and made available
an initial 6000 pages of investigatory files in its public records
office. The AFL-CIO and DNC petitioned the Commission
to withdraw the released files from the records office and to
withhold virtually all remaining documents, arguing that the
materials were protected by certain FOIA exemptions and by
section 437g(a)(12)(A)’s prohibition against disclosing ‘‘[a]ny
notification or investigation made under this section’’ without
written consent of the respondent. According to the petition-
ers, disclosure of the materials would disadvantage them by
revealing political tactics and strategies to their opponents
and by needlessly releasing information identifying hundreds
of officials, employees, and volunteers named in the docu-
ments, making it harder for the two organizations to recruit
such personnel in the future. (Like the parties, we shall
refer to the two petitioners as the AFL-CIO.) The Commis-
sion denied both the section 437g and FOIA claims.
The AFL-CIO then filed suit against the Commission in the
U.S. District Court for the District of Columbia. Applying
Chevron U.S.A. Inc. v. Natural Resources Defense Council,
Inc., 467 U.S. 837, 842–43 (1984), the district court held that
section 437g(a)(12)(A)’s plain language protects investigatory
files from disclosure. AFL-CIO v. FEC, 177 F. Supp. 2d 48,
6
55–59 (D.D.C. 2001). The court also concluded that the
Commission’s position violated 11 C.F.R. § 111.21 and FOIA
Exemption 7(C), 5 U.S.C. § 552(b)(7)(C). AFL-CIO, 177 F.
Supp. 2d at 59–61.
The Commission appeals, arguing that its longstanding
disclosure policy warrants Chevron deference. In support of
the Commission, the Campaign and Media Legal Center, the
Center for Responsive Politics, and the National Voting
Rights Institute assert in their amicus brief that disclosure of
investigatory files is essential to public oversight of the
Commission. The AFL-CIO defends the district court’s
Chevron analysis and argues that even if FECA is ambiguous,
the Commission deserves no deference because its interpreta-
tion raises serious First Amendment problems. Amplifying
the latter point, the James Madison Center for Free Speech
argues in its amicus brief that the Commission’s policy cre-
ates an incentive for political groups to file complaints against
their opponents in order to gain access to their strategic
plans, as well as to chill the opponents’ activities. We review
the district court’s grant of summary judgment de novo.
Nat’l Mining Ass’n v. Fowler, 324 F.3d 752, 756 (D.C. Cir.
2003).
II.
Because the Commission is charged with administering
FECA, we analyze its regulations under the Chevron frame-
work. See, e.g., Republican Nat’l Comm. v. FEC, 76 F.3d
400, 404 (D.C. Cir. 1996). As usual, we begin by asking
whether Congress has spoken ‘‘directly TTT to the precise
question at issue,’’ since both we and the Commission must
give effect to Congress’s unambiguously expressed intent.
Chevron, 467 U.S. at 842–43. We consider the provisions at
issue in context, using traditional tools of statutory construc-
tion and legislative history. FDA v. Brown & Williamson
Tobacco Corp., 529 U.S. 120, 132–33 (2000). A statute is
considered ambiguous if it can be read more than one way.
United States v. Nofziger, 878 F.2d 442, 446–47 (D.C. Cir.
1989). We evaluate the statute’s clarity ourselves, giving no
deference to the agency’s interpretation. SBC Communica-
tions Inc. v. FCC, 138 F.3d 410, 418–19 (D.C. Cir. 1998).
7
Only if the statute is either silent or ambiguous on the
question at issue do we defer to the agency’s reasonable
interpretation. Chevron, 467 U.S. at 842–43.
Before considering the Commission’s plea for Chevron def-
erence, we must address the AFL-CIO’s argument that it
should prevail at Chevron step one. Echoing the district
court’s reasoning, it argues that subsection (a)(12)(A)—‘‘[a]ny
notification or investigation made under this section shall not
be made public by the Commission or by any person without
the written consent of the person receiving such notification
or the person with respect to whom such investigation is
made’’—speaks directly to the question at issue by prohibit-
ing the release of investigatory file materials in both open and
closed cases. For its part, the Commission asserts that the
term ‘‘investigation’’ refers only to the act of investigating a
complaint, not to the fruits of the inquiry. According to the
Commission, once a pending investigation has closed and later
stages of the enforcement process have implicitly revealed the
fact that the Commission engaged in the act of investigation,
the statute simply does not address the confidentiality of
related files.
As we read subsection (a)(12)(A), we cannot agree with the
AFL-CIO that the term ‘‘investigation’’ unambiguously en-
compasses files compiled during a Commission proceeding.
All other uses of ‘‘investigation’’ in section 437g(a) refer to a
fact-finding process, not to materials compiled during that
process. See 2 U.S.C § 437g(a)(1) (Commission ‘‘may not
conduct any investigation or take any other action’’ based
solely on anonymous complaints); id. § 437g(a)(2) (Commis-
sion ‘‘shall make an investigation of TTT [an] alleged violation,
which may include a field investigation or audit,’’ upon a
finding of ‘‘reason to believe’’); id. § 437g(a)(11) (Commission
may petition to hold a party in civil contempt if it ‘‘determines
after an investigation’’ that the party violated a court order).
Moreover, we think it telling that in contrast to subsection
(a)(12)(A), section 437g(a)’s other confidentiality provision ex-
pressly protects information revealed in the course of Com-
mission enforcement actions. Specifically, subsection
(a)(4)(B)(i) forbids the disclosure of both ‘‘action[s] by the
8
Commission or any person TTT in connection with any concili-
ation attempt’’ and ‘‘information derived’’ in connection with
such attempts. Id. § 437g(a)(4)(B)(i). Finally, the fact that
subsection (a)(12)(A) permits disclosure upon written consent
of ‘‘the person with respect to whom such investigation is
made’’ without addressing the interests of third party wit-
nesses who have provided confidential materials to investiga-
tors also implies that Congress was concerned about protect-
ing the targets of Commission investigations, not the agency’s
sources of information.
The AFL-CIO argues that its interpretation is supported
by viewing subsection (a)(12)(A) in the context of section
437g(a)’s larger structure, but we think that exercise further
demonstrates that the statute has two possible meanings.
According to the AFL-CIO, section 437g(a)’s express require-
ment to release Commission no-violation determinations and
signed conciliation agreements, id. § 437g(a)(4)(B)(ii), would
be superfluous if Congress had intended the Commission to
release all investigatory file materials in every closed case.
Thus, except for the limited disclosures required by subsec-
tion (a)(4)(B)(ii), the AFL-CIO asserts, subsection (a)(12)(A)’s
confidentiality mandate continues to protect all other investi-
gation-related documents even after an investigation has end-
ed. This argument makes sense, however, only if subsection
(a)(12)(A) in fact protects investigatory files. If, as the
Commission maintains, the statute prohibits only disclosure of
the fact that the Commission is engaged in the act of investi-
gating a complaint, then the provision the AFL-CIO relies on
merely serves to formally terminate that confidentiality re-
quirement at the time the Commission either determines that
no violation occurred or signs a conciliation agreement, since
the release of a no-violation determination or conciliation
agreement will implicitly reveal the fact that an investigation
occurred. For that matter, so will the filing of a civil
enforcement action or a dismissal challenge by an ‘‘aggrieved’’
party, neither of which is required to be litigated under seal.
Id. § 437g(a)(6), (8).
The AFL-CIO next argues that the statute’s legislative
history supports its interpretation of the statute. To the
extent the sparse legislative record provides any clue of
9
Congress’s intent, however, we think it supports the Commis-
sion’s interpretation. Congress provided no explanation
when it enacted the original version of subsection (a)(12)(A) in
1974. Pub. L. No. 93-443, § 208(a) (1974). Scattered refer-
ences in the legislative history of FECA’s 1976 amendments,
which added a civil penalty for subsection (a)(12)(A) violations
and the confidentiality and disclosure requirements in what is
now subsection (a)(4)(B), emphasize the importance of pro-
tecting the confidentiality of pending investigations and indi-
cate that some members believed that subsection (a)(12)(A)
helps to prevent political opponents from bringing baseless
charges against each other for purposes of generating nega-
tive publicity. See, e.g., H.R. CONF. REP. NO. 94-1057, at 49–50
(1976) (‘‘The conferees’ intent is that a violation within the
meaning of [subsection (a)(12)(A)] occurs when publicity is
given to a pending investigation, but [a violation] does not
occur when actions taken in carrying out an investigation lead
to public awareness of the investigation.’’); H.R. REP. NO. 94-
917, at 66 (1976) (indicating that the penalty for violating
subsection (a)(12)(A) applies to any party who reveals ‘‘the
identity of any person under investigation’’); 122 CONG. REC.
8566 (1976) (comments of Rep. Wayne Hays) (arguing when
presenting the bill on the House floor that subsection
(a)(12)(A) limits unfair publicity by prohibiting disclosure of
‘‘the fact that an individual is being investigated or is about to
be investigated’’ until after the Commission has determined
whether charges have merit). The legislative record, howev-
er, contains no references to protecting the confidentiality of
closed Commission files or to the interests of parties who
provide information during the course of an investigation.
Considering all of these factors together, we think the
Commission may well be correct that subsection (a)(12)(A) is
silent with regard to the confidentiality of investigatory files
in closed cases and that Congress merely intended to prevent
disclosure of the fact that an investigation is pending. But
even if the AFL-CIO could convince us that its alternate
construction represents the more natural reading of subsec-
tion (a)(12)(A), the fact that the provision can support two
10
plausible interpretations renders it ambiguous for purposes of
Chevron analysis. Nofziger, 878 F.2d at 446–47.
Contrary to the AFL-CIO’s contention, nothing in In re
Sealed Case, 237 F.3d 657 (D.C. Cir. 2001), requires a differ-
ent result. There we held that the Commission must file
motions to enforce its investigatory subpoenas under seal
because subsection (a)(12)(A) ‘‘plainly prohibit[s] the FEC
from disclosing information concerning ongoing investigations
under any circumstances without the written consent of the
subject of the investigation.’’ Id. at 666–67. In so holding,
we had no need to address the issue we face here, as the
Commission’s disclosures in that case revealed both that an
investigation was pending against a particular party and
several individual file documents compiled during the course
of the agency’s investigation. See id. at 662 (Commission
filed in open court the complaint, a staff analysis detailing the
alleged violations, the Commission’s ‘‘reason to believe’’ find-
ing, and information concerning an unrelated investigation).
Also, we specifically limited our analysis to the context of
ongoing investigations where, because secrecy is needed to
protect an innocent accused from damaging publicity, respon-
dents have a ‘‘strong confidentiality interest’’ analogous to the
interests of targets of grand jury investigations. Id. at 667.
But that analogy breaks down once a Commission investiga-
tion closes because FECA expressly requires disclosure of
‘‘no violation’’ findings, 2 U.S.C. § 437g(a)(4)(B)(ii), whereas
Federal Rule of Criminal Procedure 6(e)(6) continues to
protect suspects exonerated by a grand jury. Illinois v.
Abbott & Assocs., 460 U.S. 557, 566 n.11 (1983).
III.
We turn, then, to the Chevron step two issue—whether the
Commission’s post-investigation disclosure policy reflects ‘‘a
permissible construction of the statute.’’ Chevron, 467 U.S.
at 843. The parties devote many pages of briefing to this
issue: The Commission argues that it deserves particular
deference because its regulation embodies a longstanding
procedural policy, while the AFL-CIO questions the policy’s
11
historical and legal origins. At this stage of our Chevron
analysis, we would normally accord ‘‘ ‘considerable defer-
ence’ ’’ to the Commission, United States v. Kanchanalak, 192
F.3d 1037, 1049 (D.C. Cir. 1999) (citation omitted), particular-
ly where, as here, Congress took no action to disapprove the
regulation when the agency submitted it for review pursuant
to 2 U.S.C. § 438(d). See, e.g., FEC v. Democratic Senatorial
Campaign Comm., 454 U.S. 27, 34 (1981). As the AFL-CIO
points out, however, because we are also ‘‘obliged to construe
the statute to avoid constitutional difficulties if such a con-
struction is not plainly contrary to the intent of Congress,’’
we do not accord the Commission deference when its regula-
tions create ‘‘serious constitutional difficulties.’’ Chamber of
Commerce v. FEC, 69 F.3d 600, 604–05 (D.C. Cir. 1995)
(citing Edward J. DeBartolo Corp. v. Fla. Gulf Coast Bldg. &
Constr. Trades Council, 438 U.S. 568, 575 (1988)). This is
just such a case because Congress’s failure to act obviously
cannot be viewed as a clear expression of intent and because,
as we shall show, the Commission failed to tailor its disclosure
policy to avoid unnecessarily infringing upon First Amend-
ment rights.
We begin with a little background. The Supreme Court
has long recognized that compelled disclosure of political
affiliations and activities can impose just as substantial a
burden on First Amendment rights as can direct regulation.
See, e.g., Buckley v. Valeo, 424 U.S. 1, 64–68 (1976) (disclosure
of campaign contributions); NAACP v. Alabama ex rel. Pat-
terson, 357 U.S. 449, 462–63 (1958) (disclosure of membership
lists). When facing a constitutional challenge to a disclosure
requirement, courts therefore balance the burdens imposed
on individuals and associations against the significance of the
government interest in disclosure and consider the degree to
which the government has tailored the disclosure requirement
to serve its interests. Buckley, 424 U.S. at 64–68; Block v.
Meese, 793 F.2d 1303, 1315–16 (D.C. Cir. 1986). Where a
political group demonstrates that the risk of retaliation and
harassment is ‘‘likely to affect adversely the ability of TTT [the
group] and its members to pursue their collective effort to
foster beliefs which they admittedly have the right to advo-
12
cate,’’ for instance, the government may justify the disclosure
requirement only by demonstrating that it directly serves a
compelling state interest. NAACP, 357 U.S. at 462–63. In
contrast, where the burden on associational rights is ‘‘insub-
stantial,’’ we have upheld a disclosure requirement that pro-
vided ‘‘the only sure means of achieving’’ a government
interest that was, though valid, ‘‘not TTT of the highest
importance.’’ Block, 793 F.2d at 1316–18. Furthermore,
even where requiring disclosure of political or speech activi-
ties to a government agency may be necessary to facilitate
law enforcement functions, we have held that ‘‘[c]ompelled
public disclosure presents a separate first amendment issue’’
that requires a separate justification. Id. at 1315 (emphasis
added).
In affidavits submitted in the district court, the AFL-CIO
and DNC assert that releasing the names of hundreds of
volunteers, members, and employees will make it more diffi-
cult for the organizations to recruit future personnel. The
Commission urges us to dismiss this assertion of a chilling
effect as speculative. Although we agree that the evidence in
this case is far less compelling than the evidence presented in
cases involving groups whose members had been subjected to
violence, economic reprisals, and police or private harass-
ment, see, e.g., Brown v. Socialist Workers ’74 Campaign
Comm. (Ohio), 459 U.S. 87, 99 (1982); NAACP, 357 U.S. at
462, that difference speaks to the strength of the First
Amendment interests asserted, not to their existence. In
Buckley v. Valeo, for example, the Supreme Court conclud-
ed—without considering either the popularity of the parties
involved or any specific evidence of retaliation—that disclo-
sure of campaign contributions would chill political activity
and therefore place ‘‘not insignificant burdens’’ on First
Amendment rights. 424 U.S. at 65–66, 68. Although it is
true, as the Commission points out, that a separate part of
the opinion held that minority parties could be exempted from
disclosure requirements by demonstrating ‘‘a reasonable
probability TTT [of] threats, harassment, or reprisals,’’ the
Commission fails to note that the Court required such an
evidentiary showing only after concluding that the disclosure
13
requirements at issue survived strict scrutiny as the least
intrusive means of achieving several compelling government
interests. Id. at 69–74; see also Cmty.–Serv. Broad. of Mid–
Am., Inc. v. FCC, 593 F.2d 1102, 1118 & n.38 (D.C. Cir. 1978)
(en banc) (Wright, C.J., joined by Wilkey, J.) (noting that
Buckley engaged in a full First Amendment analysis despite
the absence of concrete evidence of retaliation).
Moreover, in addition to arguing that disclosure will chill
future individual political activity, the AFL-CIO and DNC
affidavits charge that disclosing detailed descriptions of train-
ing programs, member mobilization campaigns, polling data,
and state-by-state strategies will directly frustrate the organi-
zations’ ability to pursue their political goals effectively by
revealing to their opponents ‘‘activities, strategies and tactics
[that] we have pursued in subsequent elections and will likely
follow in the future.’’ Rosenthal Aff. at 2–4; see also Stoltz
Aff. at 3. The Commission does not challenge the factual
basis for this claim. Although we have found no cases
presenting similar facts, the Supreme Court has concluded
that extensive interference with political groups’ internal
operations and with their effectiveness does implicate signifi-
cant First Amendment interests in associational autonomy.
For example, in Eu v. San Francisco County Democratic
Central Committee, 489 U.S. 214 (1989), the Court held that
California statutes dictating the composition and leadership of
political parties’ governing bodies burdened significant associ-
ational rights by ‘‘limit[ing] a political party’s discretion in
how to organize itself, conduct its affairs, and select its
leaders,’’ thereby also potentially ‘‘color[ing] the parties’ mes-
sage and interfer[ing] with the parties’ decisions as to the
best means to promote that message.’’ Id. at 229–31 & n.21.
In the same case, the Court applied strict scrutiny to a
section of the statute that prohibited party leaders from
endorsing candidates in primaries because such limitations
‘‘directly hamper[ ] the ability of a party to spread its mes-
sage.’’ Id. at 223. Similarly, in Tashjian v. Republican
Party, 479 U.S. 208, 224 (1986), where the Court struck down
a closed primary system, it held, ‘‘[t]he Party’s determination
of the boundaries of its own association, and of the structure
14
which best allows it to pursue its political goals, is protected
by the Constitution.’’ Even the Court’s compelled disclosure
decisions emphasize the potential damage to political groups’
effectiveness, in addition to the risk of chilling individual
participation. In Buckley’s analysis of contribution disclosure
requirements, for example, the Court explained, ‘‘[t]he right
to join together ‘for the advancement of beliefs and ideas,’ is
diluted if it does not include the right to pool money through
contributions, for funds are often essential if ‘advocacy’ is to
be truly or optimally ‘effective.’ ’’ 424 U.S. at 65–66 (quoting
NAACP, 357 U.S. at 460) (citation omitted); see also Ripon
Soc’y, Inc. v. Nat’l Republican Party, 525 F.2d 567, 585 (D.C.
Cir. 1975) (en banc) (plurality opinion) (‘‘Speeches and assem-
blies are after all not ends in themselves but means to effect
change through the political process. If that is so, there must
be a right not only to form political associations but to
organize and direct them in the way that will make them most
effective.’’).
In this case, the AFL-CIO and DNC affidavits indicate that
compelled disclosure of internal planning materials, though
less direct than regulation of political group leadership or
structure, will similarly frustrate those groups’ decisions as to
‘‘how to organize TTT [themselves], conduct TTT [their] affairs,
and select TTT [their] leaders,’’ as well as their selection of a
‘‘message and TTT the best means to promote that message.’’
Eu, 489 U.S. at 230–31 & n.21. Although we do not suggest
that any Commission action that places a political association
at a disadvantage relative to its opponents violates the First
Amendment, where, as here, the Commission compels public
disclosure of an association’s confidential internal materials, it
intrudes on the ‘‘privacy of association and belief guaranteed
by the First Amendment,’’ Buckley, 424 U.S. at 64, as well as
seriously interferes with internal group operations and effec-
tiveness.
Having concluded that the AFL-CIO and DNC have assert-
ed substantial First Amendment interests in the disclosure of
their own internal materials and at least marginal interests in
preventing the chilling of political participation by their mem-
bers and officials, we proceed to assess the strength of the
15
government’s proffered interest in disclosure. The Commis-
sion offers two justifications for 11 C.F.R. § 5.4(a)(4): The
regulation deters FECA violations, and it promotes the agen-
cy’s own public accountability. Although we have no doubt
that these interests are valid, we need not engage in a
detailed balancing analysis, for the Commission made no
attempt to tailor its policy to avoid unnecessarily burdening
the First Amendment rights of the political organizations it
investigates. See, e.g., United States v. Popa, 187 F.3d 672,
676 (D.C. Cir. 1999) (declining to determine the precise level
of scrutiny applicable to a particular statute where it was
insufficiently tailored to meet even the least exacting stan-
dard). Indeed, the blanket nature of the Commission’s regu-
lation—requiring, as it does, the release of all information not
expressly exempted by FOIA—appears to result in the re-
lease of significant amounts of information that furthers
neither goal. For example, the Commission never explains
how releasing investigatory files will deter future violations in
cases where, as here, the respondents have been cleared of
wrongdoing. Nor does the Commission explain how a policy
requiring the release of materials that played no meaningful
role in its decisionmaking process will promote its own ac-
countability. The facts of this case are particularly disturb-
ing because the Commission proposes to release between
10,000 and 20,000 pages of documentation that it has never
examined. The materials therefore cannot shed light on the
Commission’s reasoning, and may not even relate to question-
able activities. The fact that the Commission redacts infor-
mation falling under one or more FOIA exemptions is no
answer, since the Freedom of Information Act does little to
protect the First Amendment interests at issue.
Adding to the First Amendment concerns in this case, the
AFL-CIO and James Madison Center argue—persuasively in
our view—that when combined with the Commission’s broad
subpoena practices, the automatic disclosure regulation ‘‘en-
courages political opponents to file charges against their
competitors to serve the dual purpose of ‘chilling’ the expres-
sive efforts of their competitor and learning their political
strategy so that it can be exploited to the complainant’s
16
advantage.’’ Madison Amicus Br. at 20. We have no doubt,
as agency counsel explained at oral argument, that the Com-
mission does its best to screen out frivolous complaints at the
‘‘reason to believe’’ stage, but such efforts do nothing to
reduce the incentive for political adversaries to attempt to
turn the Commission’s disclosure regulation to their own
advantage. As this case demonstrates, the release policy
gives parties a large potential ‘‘bonus’’ for filing a complaint
because even if their allegations of wrongdoing are rejected,
they may still obtain access to thousands of pages of their
opponents’ internal strategic information.
The Commission argues that no First Amendment problem
exists here because it is ‘‘merely disclosing its own agency
records to the public TTT, rather than coercing a private party
to produce information.’’ Commission Reply Br. at 24. This
position is quite remarkable. FECA authorizes the Commis-
sion to order any person to submit written reports and
answer its questions, to subpoena witnesses to testify or
present documentary evidence, and to seek judicial enforce-
ment of such orders and subpoenas. 2 U.S.C. § 437d(a)(1),
(a)(3), (b). Thus, agency investigators do use ‘‘some actual or
threatened imposition of governmental power or sanction’ ’’ to
obtain information from respondents and other witnesses.
Penthouse Int’l Ltd. v. Meese, 939 F.2d 1011, 1015 (D.C. Cir.
1991). While it is true, as the Commission asserts, that
political groups could seek protective orders to ensure confi-
dentiality, turning every discovery request and subpoena into
a First Amendment court battle would burden both the
judiciary and the Commission. Furthermore, the considera-
ble time that transpires between an initial subpoena and the
Commission’s ultimate public disclosure of a closed case file
does not somehow convert the materials involved into the
Commission’s ‘‘own agency records.’’ Rather, where the
Commission compiles information relating to speech or politi-
cal activity for law enforcement purposes, our case law re-
quires that it provide a separate First Amendment justifica-
tion for publicly disclosing the materials. Block, 793 F.2d at
1315–16.
17
In sum, although we agree that deterring future violations
and promoting Commission accountability may well justify
releasing more information than the minimum disclosures
required by section 437g(a), the Commission must attempt to
avoid unnecessarily infringing on First Amendment interests
where it regularly subpoenas materials of a ‘‘delicate nature
TTT represent[ing] the very heart of the organism which the
first amendment was intended to nurture and protect.’’ Ma-
chinists Non–Partisan Political League, 655 F.2d at 388.
Because 11 C.F.R. § 5.4(a)(4) fails to undertake this tailoring,
it creates the ‘‘serious constitutional difficulties’’ outlined
above. Chamber of Commerce, 69 F.3d at 605. We therefore
conclude that the regulation is impermissible.
We end with a comment about the concurring opinion,
which reads Edward J. DeBartolo Corp. v. Florida Gulf
Coast Building & Construction Trades Council, 438 U.S. 568
(1988), as requiring that we account for the constitutional
concerns at step one of the Chevron analysis. Were there
only one way to read the statute that would avoid constitu-
tional problems, we might well agree, for ‘‘Congress TTT is
bound by and swears an oath to uphold the Constitution[,]
[and] [t]he courts will therefore not lightly assume that
Congress intended to infringe constitutionally protected liber-
ties or usurp power constitutionally forbidden it.’’ Id. at 575.
But here the statute is susceptible to more than one constitu-
tionally permissible interpretation: As we have indicated, the
Commission could tailor its disclosure policy to avoid unneces-
sary First Amendment infringements, or, as the concurrence
maintains, it could decline to release any materials other than
those expressly required by section 437g(a). Neither DeBar-
tolo nor Chevron suggest that the court—as opposed to the
agency—should choose between these permissible alterna-
tives. Rather, DeBartolo’s mandate that ‘‘ ‘every reasonable
construction must be resorted to, in order to save a statute
from unconstitutionality,’ ’’ id. (quoting Hooper v. California,
155 U.S. 648, 657 (1895)), suggests merely that an agency acts
unreasonably if, instead of choosing among constitutionally
permissible alternatives, it interprets ambiguous statutory
language as indicating that Congress intended to authorize
18
infringements on constitutional rights. Thus, we think the
constitutional issues raised by the Commission’s disclosure
policy are properly addressed at Chevron step two. See, e.g.,
Chamber of Commerce, 69 F.3d at 604–05 (addressing consti-
tutional problems created by the Commission’s interpretation
of an undefined statutory term at step two of the Chevron
analysis).
The judgment of the district court is affirmed.
So ordered.
1
KAREN LECRAFT HENDERSON, Circuit Judge, concurring in
the judgment:
We are asked in this case, as in so many others, to dance
the Chevron two-step, under which
the court must first exhaust the traditional tools of
statutory construction to determine whether Con-
gress has spoken to the precise question at is-
sueTTTT If the court can determine congressional
intent, then that interpretation must be given ef-
fectTTTT If, on the other hand, the statute is silent
or ambiguous with respect to the specific issue, then
the court will defer to a permissible agency con-
struction of the statute.
NRDC v. Browner, 57 F.3d 1122, 1125 (D.C. Cir. 1995)
(internal quotations omitted); see Chevron, U.S.A., Inc. v.
NRDC, 467 U.S. 837, 842–43 (1984) (‘‘If the intent of Con-
gress is clear, 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.’’). Unlike the majority, I
would stop the music at Chevron step one. In my view, a
proper employment of ‘‘the traditional tools of statutory
construction’’ yields a plain and affirmative congressional
response ‘‘to the precise question at issue’’—i.e., whether the
Federal Election Campaign Act (FECA or Act) prohibits the
Federal Election Commission (FEC or Commission) from
disclosing thousands of pages of politically sensitive docu-
ments it obtained while investigating the AFL–CIO and the
DNC (collectively, the appellees), now that the investigation
has been completed.
My inquiry into the Congress’s intent proceeds, as it must,
from ‘‘the fundamental canon that statutory interpretation
begins with the language of the statute itself.’’ Butler v.
West, 164 F.3d 634, 639 (D.C. Cir. 1999) (quotation omitted);
see HENRY J. FRIENDLY, BENCHMARKS 202 (1967) (‘‘(1) Read the
statute; (2) read the statute; (3) read the statute!’’ (quoting
Justice Frankfurter’s ‘‘threefold imperative to law stu-
dents’’)). The Act’s confidentiality provision states, in full,
that
2
[a]ny notification or investigation made under this
section shall not be made public by the Commission
or by any person without the written consent of the
person receiving such notification or the person with
respect to whom such investigation is made.
2 U.S.C. § 437g(a)(12)(A). Stretching the text in hopes of
reaching the deference it enjoys under Chevron step two, the
Commission contends that section 437g(a)(12)(A) ‘‘does not
specifically address closed cases, and its silence about wheth-
er and when the confidentiality requirement expires leaves
Congress’s intent on that question ambiguous.’’ Br. of Appel-
lant at 14; see id. at 16–17 (citing United States v. Kancha-
nalak, 192 F.3d 1037, 1049 (D.C. Cir. 1999) (FEC’s interpre-
tation of FECA deserves ‘‘considerable deference’’ where
statute ambiguous)). I disagree. While the provision does
not state in so many words that ‘‘no completed investigation
shall be made public,’’ that does not mean it is silent on the
matter; whatever the word ‘‘investigation’’ means, section
437g(a)(12)(A) plainly covers ‘‘[a]ny TTT investigation,’’ ongo-
ing or completed. 2 U.S.C. § 437g(a)(12)(A) (emphasis add-
ed). Indeed, even though the provision does not explicitly
state that ‘‘the FEC may not file information concerning an
ongoing investigation on the public record when it seeks to
enforce a subpoena,’’ we recently held that it ‘‘unambiguous-
ly,’’ ‘‘directly’’ and ‘‘unequivocally’’ prohibits precisely that.
See In re Sealed Case, 237 F.3d 657, 667 (D.C. Cir. 2001).
Taken to its logical conclusion, the FEC’s argument would
render every prohibition in the United States Code suscepti-
ble of ambiguity. ‘‘Thou shall not kill’’ is a mandate neither
silent nor ambiguous about whether murder is permissible if
committed after 5.00 p.m.—or, for that matter, if committed
in the billiard room with the candlestick—but the FEC’s
reasoning would lead one to conclude otherwise.
Moreover, the fact that the provision does not specify
‘‘when the confidentiality requirement expires’’ suggests to
me that it never expires. Resisting this logic, the Commis-
sion contends that the ‘‘disclosures TTT required by other
provisions of section 437g(a) when the administrative proceed-
ing concludes TTT make section 437g(a)(12)(A) inapplicable to
3
closed cases.’’ Br. of Appellant at 20–21. In doing so, the
FEC again neglects the plain language of the statute; section
437g(a)(12)(A)’s prohibition against disclosure of an ‘‘investi-
gation’’ or a ‘‘notification’’ admits of no textual exceptions.
See Sealed Case, 237 F.3d at 667. True, section 437g(a)
elsewhere states that ‘‘the Commission shall make public any
conciliation agreement signed by both the Commission and
the respondent.’’ 2 U.S.C. § 437g(a)(4)(B)(ii) (emphasis add-
ed). True, it also provides that ‘‘[i]f the Commission makes a
determination that a person has not violated this Act TTT [it]
shall make public such determination.’’ Id. (emphasis add-
ed). But section 437g(a) nowhere requires or permits the
FEC to disclose an ‘‘investigation’’ or a ‘‘notification.’’ Nor
should we be willing, in the face of well-settled principles of
statutory construction, to equate the distinct terms—‘‘investi-
gation’’/‘‘notification’’ on the one hand and ‘‘conciliation agree-
ment’’/‘‘determination’’ on the other—such that disclosure of
an ‘‘investigation’’ is permitted under the circumstances enu-
merated in section 437g(a)(4)(B)(ii). See Barnhart v. Sigmon
Coal Co., 534 U.S. 438, 452 (2002) (‘‘[W]hen Congress includes
particular language in one section of a statute but omits it in
another section TTT , it is generally presumed that Congress
acts intentionally and purposely in the disparate inclusion or
exclusion.’’ (quotations omitted)); Russello v. United States,
464 U.S. 16, 23 (1983) (‘‘We refrain from concluding TTT that
the differing language in the two subsections has the same
meaning in each.’’); see also 2A NORMAN J. SINGER, SUTHER-
LAND’S STATUTES AND STATUTORY CONSTRUCTION § 46.06, at 194
(6th ed. 2000) (‘‘[W]hen the legislature uses certain language
in one part of the statute and different language in another,
the court assumes different meanings were intended.’’). If
the Congress had meant to exclude the majority1 of ‘‘investi-
gation[s]’’ from the otherwise comprehensive shelter of sec-
tion 437g(a)(12)(A), it would have done so more obviously than
the FEC suggests—either by including ‘‘investigation’’ in
section 437g(a)(4)(B)(ii), as it did in section 437g(a)(12)(A), or
1 There are far fewer open cases at any given time than there
are closed ones—a disparity that will inevitably grow over time as
ongoing investigations are completed.
4
by removing completed ‘‘investigation[s]’’ from the latter’s
coverage explicitly. See Russello, 464 U.S. at 23 (‘‘We would
not presume to ascribe this difference [in the subsections’
language] to a simple mistake in draftsmanship.’’); cf. Cal.
Med. Ass’n v. FEC, 453 U.S. 182, 191 (1981) (‘‘If Congress
had intended to remove a whole category of constitutional
challenges from the purview of § 437h, thereby significantly
limiting the usefulness of that provision, it surely would have
made such a limitation explicit.’’). In my view, therefore,
section 437g(a)(12)(A) plainly prohibits publication of any
‘‘investigation,’’ whether it is ongoing or completed.
In a belated nod to ordinary meaning, and in the absence of
a statutory definition, the Commission claims that ‘‘investiga-
tion’’ refers to ‘‘a process for discovering facts, not a file of
documents.’’2 Br. of Appellant at 22 (citing BLACK’S LAW
DICTIONARY 825 (6th ed. 1990); 8 OXFORD ENGLISH DICTIONARY
47 (2d ed. 1989); RANDOM HOUSE DICTIONARY OF THE ENGLISH
LANGUAGE, UNABRIDGED 1004 (2d ed. 1983)) (emphasis in origi-
nal). It argues that section 437g(a)(12)(A) prohibits disclo-
sure of the fact that an investigative process is occurring or
has occurred but does not cover the documents generated
during the process. Once again, I disagree.
Although we observed in Sealed Case that the main pur-
pose of section 437g(a)(12)(A) is ‘‘to protect [an] TTT accused
who is exonerated from disclosure of the fact that he has been
under investigation,’’ Sealed Case, 237 F.3d at 667 (quoting
United States v. Proctor & Gamble Co., 356 U.S. 677, 682 n.6
(1958)) (emphasis added), we suggested as well that the
‘‘investigation[s]’’ to be kept confidential include all documen-
2 As the appellees point out, the FEC did not articulate this
proposition ‘‘in its administrative dispositions of [the appellees’]
objections to public release of the investigative file, where the
Commission otherwise extensively analyzed FECA, its legislative
history, the FEC’s regulations and FOIA.’’ Br. of Appellees at 19
n.11. ‘‘Rather,’’ they note, ‘‘this argument first surfaced in the
FEC’s briefs on the cross-motions for summary judgment.’’ Id.
5
tary materials gathered during the ‘‘process’’ to which the
FEC refers:
When the FEC issues a subpoena as part of an
investigation, § 437g mandates those subpoenas, like
other components of the investigation, ‘‘shall not be
made public.’’TTTT Even if we assume that the
FEC’s argument is correct (which it is not) and the
Commission could disclose the subpoenas themselves
(which it cannot), the Commission would still lack
the authority to divulge information pertaining to
the underlying investigationTTTT
Sealed Case, 237 F.3d at 667–68 (emphases altered); see id.
at 668 (‘‘We cannot fathom why the FEC’s issuance of a
subpoena in furtherance of an ongoing investigation would not
be considered part of that ‘investigation’ within the meaning
of § 437g.’’). Indeed, we stated without qualification that the
FEC cannot ‘‘under any circumstances TTT introduce evi-
dence concerning an ongoing investigation on the public rec-
ord’’ without the written consent of the subject of the investi-
gation.3 Id. at 667, 669 (emphasis added).
Even if Sealed Case does not give an all-encompassing
interpretation to ‘‘investigation’’—and I acknowledge that it
(properly) does not—the thoroughgoing First Amendment
analysis in today’s majority opinion, in my view, removes any
doubt that the term covers the documents at issue here. The
majority quite justifiably echoes the concern of amicus curi-
3 Even if the Commission’s reading of ‘‘investigation’’ were
correct, section 437g(a)(12)(A) would nonetheless prohibit the FEC
from disclosing several of the documents it planned to (and initially
did) publicize in April 2001. The written correspondence among
and between the appellees, their counsel and the Commission, for
example, would inevitably reveal upon publication the fact that the
appellees had been investigated. Because the correspondence is
neither a ‘‘conciliation agreement signed by both the Commission
and the [appellees],’’ 2 U.S.C. § 437g(a)(4)(B)(ii), nor a ‘‘determina-
tion that [the appellees have] not violated this Act,’’ id., the Act
would plainly foreclose publication of the correspondence even
under the Commission’s ‘‘process’’ interpretation of ‘‘investigation.’’
6
ae, the James Madison Center for Free Speech, that the
Commission’s automatic disclosure policy ‘‘encourages politi-
cal opponents to file charges against their competitors to
serve the dual purpose of ‘chilling’ the expressive efforts of
their competitor and learning their political strategy so that it
can be exploited to the complainant’s advantage.’’ Maj. op. at
15–16 (quoting Br. of Amicus Curiae James Madison Center
at 20). I agree that the Commission has failed to show that
the speech-chilling disclosure regulation set forth in 11 C.F.R.
§ 5.4 bears a ‘‘relevant correlation or substantial relation’’ to
a ‘‘substantial governmental interest[ ].’’ Buckley v. Valeo,
424 U.S. 1, 64, 68 (1976) (per curiam) (quotations omitted);
see maj. op. at 14–15. But I believe the majority commits an
error—if only a minor one—in holding that the ‘‘serious
constitutional difficulties’’ raised by the regulation preclude
the court from deferring to the Commission’s interpretation
at Chevron step two. Chamber of Commerce v. FEC, 69 F.3d
600, 605 (D.C. Cir. 1995); see maj. op. at 10–11. In my view,
those very same difficulties vindicate, at Chevron step one,
the appellees’ contention that the Act ‘‘unambiguously’’ pro-
hibits the FEC from publicizing the documents at issue.
Chevron, 467 U.S. at 843; see Br. of Appellees at 12–27; see
also AFL–CIO v. FEC, 177 F. Supp. 2d 48, 59 (D.D.C. 2001)
(‘‘[T]he plain meaning of § 437g(a)(12)(A) prohibits the FEC
from disclosing the investigative file.’’).
Among the ‘‘traditional tools of statutory construction’’ ‘‘the
court must first exhaust’’ under Chevron and its progeny are
the linguistic and substantive canons of interpretation, one of
which—the canon of ‘‘constitutional avoidance’’—is particular-
ly useful in resolving the dispute before us. Invoking the
canon in Edward J. DeBartolo Corp. v. Florida Gulf Coast
Building & Construction Trades Council, 485 U.S. 568
(1988), the United States Supreme Court held that ‘‘where an
otherwise acceptable [agency] construction of a statute would
raise serious constitutional problems, the Court will construe
the statute to avoid such problems unless such construction is
plainly contrary to the intent of Congress.’’ Id. at 575.
Language like this might suggest that the canon operates at
7
Chevron step two or even displaces Chevron altogether. But
the Court in DeBartolo made clear that the avoidance canon
not only reflects the prudential concern that consti-
tutional issues [should] not be needlessly confronted,
but also recognizes that Congress, like this Court, is
bound by and swears an oath to uphold the Constitu-
tion. The courts will therefore not lightly assume
that Congress intended to infringe constitutionally
protected liberties or usurp power constitutionally
forbidden it.
DeBartolo, 485 U.S. at 575 (emphasis added); see Grenada
County Supervisors v. Brogden, 112 U.S. 261, 269 (1884) (‘‘It
ought never to be assumed that the law-making department
of the government intended to usurp or assume power prohib-
ited to it.’’ (quotation omitted)). In other words, the canon
assists us in determining the Congress’s intent and, accord-
ingly, it operates at Chevron step one. Circuit precedent
supports this proposition; we recently reaffirmed that ‘‘[i]f
employment of an accepted canon of construction illustrates
that Congress had a specific intent on the issue in question,
then the case can be disposed of under the first prong of
Chevron.’’ Halverson v. Slater, 129 F.3d 180, 184 (D.C. Cir.
1997) (emphasis and quotations omitted); see Mich. Citizens
for an Indep. Press v. Thornburgh, 868 F.2d 1285, 1292–93
(D.C. Cir.), aff’d by equally divided Court, 493 U.S. 38 (1989).
In sum, I do not believe the Congress intended section
437g(a)(12)(A) to apply so narrowly as to permit the Commis-
sion to publicize the documents at issue, in light of the
‘‘serious constitutional difficulties’’ attending such publication.
* * *
For the foregoing reasons, I would hold that section
437g(a)(12)(A) plainly prohibits the FEC from disclosing in-
vestigative records pertaining to a completed investigation if
the investigated party does not consent to disclosure. While
I disagree somewhat with the reasoning of my colleagues, I
do agree that the district court’s December 19, 2001 judgment
8
should be affirmed because the Commission’s interpretation
of the Act is ‘‘arbitrary, capricious, an abuse of discretion, or
otherwise not in accordance with law,’’ 5 U.S.C. § 706(2)(A).