eCases

29 L.Ed.2d 338

 (1971), and United Brotherhood of Carpenters & Joiners v. Scott,

 (1983). In Griffin, the Court glossed the statutory language with a caution against
reading § 1985(3) so broadly (literally) as to turn it into "a general federal tort law" that would
"apply to all tortious, conspiratorial interferences with the rights of others."

403 U.S. at 101-02

. In order to avoid such a result, the Griffin Court held that the statute prohibits only
those conspiracies motivated by "some racial, or perhaps otherwise class-based, invidiously
discriminatory animus...."

Id. at 102

 (emphasis added). The Court declined to
elaborate on the "perhaps" qualifier, however, stating in a footnote that it "need not decide, given
the facts of this case, whether a conspiracy motivated by invidiously discriminatory intent other than
racial bias would be actionable...."

Id.

 at 102 n. 9,

91 S.Ct. at

1798 n. 9.
Consistent with the gloss it placed on the statutory language in Griffin, the Court held in
Scott that § 1985(3) does not reach "conspiracies motivated by bias towards others on account of
their economic views, status, or activities,"

463 U.S. at 837

103 S.Ct. at 3361

, the "others" in that
case being nonunion members. As it had in Griffin, the Court in Scott withheld judgment on
whether § 1985(3) extends beyond "its central concern" of combating conspiracies driven by
race-based animus. See id. at 837,

103 S.Ct. at 3360

.
Thus, Griffin and Scott left open the issue we confront today, namely, whether a conspiracy
motivated by sex-based animus against women is actionable under § 1985(3). It is clear to us,
however, that if Griffin 's "perhaps otherwise class-based, invidiously discriminatory animus" means
anything at all—and we think it does—it includes sex-based animus against women. Sex-based
classifications receive heightened scrutiny under the Equal Protection Clause, see Craig,

429 U.S. at 197

97 S.Ct. at 457

, and discrimination based on sex frequently has been characterized as
"invidious." See, e.g., Frontiero v. Richardson,

411 U.S. 677

, 686-87,

93 S.Ct. 1764

, 1770,

36 L.Ed.2d 583

 (1973) ("[S]tatutory distinctions between the sexes often have the effect of invidiously
relegating the entire class of females to inferior legal status without regard to the actual capabilities
of its individual members."); United States v. Chandler,

996 F.2d 1073

, 1083 (11th Cir.1993)
("Invidious factors, such as race or sex, cannot influence a jury's recommendation of the death
penalty.").
Because the statutory language of § 1985(3) is unambiguous, ordinarily we would not
consult legislative history to discern its meaning. See, e.g., United States v. Gonzales,

520 U.S. 1

117 S.Ct. 1032

, 1035,

137 L.Ed.2d 132

 (1997) ("Given the straightforward statutory command, there
is no reason to resort to legislative history."); Ratzlaf v. United States,

510 U.S. 135

, 147-48,

114 S.Ct. 655

, 662,

126 L.Ed.2d 615

 (1994) ("we do not resort to legislative history to cloud a statutory
text that is clear"). In view of the Supreme Court's Griffin and Scott decisions, however, we think
it prudent to note that the legislative history of § 1985(3) is not manifestly inconsistent with our
holding. Although "[t]he legislative history of the Act confirms the conclusion that ... it was
primarily motivated by the lawless conduct directed at the recently emancipated citizens," Bray v.
Alexandria Women's Health Clinic,

, 319,

, 785,

 (1993)
(Stevens, J., dissenting), it also reveals that at least some members of Congress believed actionable
conspiracies would include those "against a person because he was a Democrat, ... or because he was
a Catholic, or because he was a Methodist, or because he was a Vermonter...."

Id. at 296

113 S.Ct. at 773

 (Souter, J., concurring in part and dissenting in part) (discussing remarks of Senator
Edmunds, who managed the bill on the Senate floor) (citations and quotations omitted). See also
Scott,

463 U.S. at 837

103 S.Ct. at 3360

 (although the operative provision originated in a House bill,
"Senator Edmunds' views, since he managed the bill on the floor of the Senate, are not without
weight.").
Given prevailing attitudes at the time § 1985(3) was enacted, it is certainly possible, if not
probable, that many legislators who voted for the statute were not concerned about affording legal
protection to women as a class. Nonetheless, we follow the plain meaning of the statute, because
"statutory prohibitions often go beyond the principal evil to cover reasonably comparable evils, and
it is ultimately the provisions of our laws rather than the principal concerns of our legislators by
which we are governed." Oncale v. Sundowner Offshore Servs., Inc.,

523 U.S. 75

118 S.Ct. 998

,
1002,

140 L.Ed.2d 201

 (1998).
Any conclusion that women are not a protected class under § 1985(3) would run into a solid
wall of contrary precedent. Relying primarily on the statutory language and the Supreme Court's
statement in Griffin, the seven other circuits that have actually decided the issue are unanimous in
their view that § 1985(3) applies to conspiracies motivated by sex-based animus against women.
See Libertad v. Welch,

53 F.3d 428

, 448-49 (1st Cir.1995) ("[I]t is logical that, at the very least, the
classes protected by § 1985(3) must encompass those classifications that merit heightened scrutiny
under Equal Protection Clause analysis, of which gender is one."); National Org. For Women v.
Operation Rescue,

914 F.2d 582

, 585 (4th Cir.1990) (collecting circuit cases), rev'd in part, vacated
in part on other grounds, Bray v. Alexandria Women's Health Clinic,

 (1993); New York State Nat'l Org. for Women v. Terry,

886 F.2d 1339

, 1359 (2d
Cir.1989) ("By its very language § 1985(3) is necessarily tied to evolving notions of equality and
citizenship. As conspiracies directed against women are inherently invidious, and repugnant to the
notion of equality of rights for all citizens, they are therefore encompassed under the Act."); Volk
v. Coler,

845 F.2d 1422

, 1434 (7th Cir.1988) ("[Section] 1985(3) extends beyond conspiracies to
discriminate against persons based on race to conspiracies to discriminate against persons based on
sex, religion, ethnicity or political loyalty."); Life Ins. Co. of N. Am. v. Reichardt,

591 F.2d 499

, 505
(9th Cir.1979) ("[W]e conclude that women purchasers of disability insurance are a sufficient class
[under § 1985(3) ]."); Novotny v. Great Am. Fed. Sav. & Loan Assoc.,

584 F.2d 1235

, 1243 (3d
Cir.1978) ("The fact that a person bears no responsibility for gender, combined with the pervasive
discrimination practiced against women, and the emerging rejection of sexual stereotyping as
incompatible with our ideals of equality convince us that whatever the outer boundaries of the
concept, an animus against women includes the elements of a "class-based invidiously
discriminatory' motivation."), vacated on other grounds,

442 U.S. 366

99 S.Ct. 2345

60 L.Ed.2d 957

 (1979); Conroy v. Conroy,

575 F.2d 175

, 177 (8th Cir.1978) (concluding that the district court
properly exercised jurisdiction where plaintiff had alleged a cause of action under § 1985(3) based
in part on sex discrimination).
In addition to the holdings of the circuits we have set out above, one other circuit has said
in dicta that sex-based animus is actionable under § 1985(3). See Haverstick Enters., Inc. v.
Financial Fed. Credit, Inc.,

32 F.3d 989

, 994 (6th Cir.1994). Only two circuits have indicated—and
then only in dicta—that they would hold women do not constitute a protected class under § 1985(3).
See Deubert v. Gulf Fed. Sav. Bank,

820 F.2d 754

, 757 (5th Cir.1987); Wilhelm v. Continental Title
Co.,

720 F.2d 1173

, 1176 (10th Cir.1983).1
The grand tally is this. Of the circuits that have squarely confronted and decided the issue,
seven have held that women are a protected class of persons under § 1985(3), and none have held
that they are not. If both holdings and dicta are counted, eight of the circuits that have taken a
position have said that women are a protected class under § 1985(3); only two have said that they
are not.
Even though the views of individual Justices are not binding on us, see, e.g., United States
v. Stewart,

65 F.3d 918

, 924 (11th Cir.1995), it is appropriate in a case of first impression in this
circuit to consider the views of those Justices who have expressed themselves on the issue. In Bray,
Justices Souter, Stevens, and O'Connor all stated independently and unequivocally that they believe
women are a protected class under § 1985(3). See Bray,

506 U.S. at 295-96

113 S.Ct. at

769-75
1
While the views of other circuits are not binding upon us in any event, we do accord them
respect. But we give views espoused as mere dicta less respect than we give those that are
forged as part of the holding in a case. See McDonald's Corp. v. Robertson,

147 F.3d 1301

,
1315 (11th Cir.1998) (Carnes, J., concurring) (explaining why conclusions stated as dicta are
less reliable than those that are part of a holding in a case).
(Souter, J., concurring in the judgment in part and dissenting in part);

id. at 322

113 S.Ct. at 787

(Stevens, J., dissenting);

id. at 350

113 S.Ct. at 802

 (O'Connor, J., dissenting). We add the views
of those three Justices to the decisional mix and give them weight.
In addition, although the majority opinion in Bray does not explicitly hold that women are
a protected class under § 1985(3), some of its analysis seems to suggest that they are. The issue in
Bray was whether attempts to blockade abortion clinics to keep people from going inside constituted
a conspiracy to deprive a person or a class of persons of equal protection of the laws in violation of
§ 1985(3). See Bray,

506 U.S. at 266

113 S.Ct. at 757-58

. The Court rejected "the claim that
petitioners' opposition to abortion reflects an animus against women in general," and explained that
"[w]e do not think that the "animus' requirement can be met only by maliciously motivated, as
opposed to assertedly benign (though objectively invidious), discrimination against women. It does
demand, however, at least a purpose that focuses upon women by reason of their sex...." Bray, at
269-70,

113 S.Ct. at 759

. Although Bray did not hold § 1985(3) outlaws conspiracies driven by
sex-based animus against women, language from the opinion at least hints that it does. We take the
hint.
For all of the stated reasons, we conclude that women are a "class of persons" within the
meaning of § 1985(3), and therefore are protected by that provision from conspiracies against them
motivated by sex-based animus. We respond now to a point raised by Judge Tjoflat in his dissenting
opinion.
Judge Tjoflat concedes that the Thirteenth Amendment provides Congress with a valid basis
for enacting § 1985(3) insofar as that statutory provision involves race, but questions the
constitutional foundation for the provision insofar as it involves sex. We need not answer that
question as it applies to private actor conspiracies against women, because those are not the facts
of this case. This case involves alleged action under color of state law. All of the individual
defendants are state or local officials, and it is undisputed that any action they took involving the
plaintiff was action under color of state law. Thus, the only constitutional issue this case presents
is whether Congress had a valid basis for proscribing conspiratorial discrimination against women
by persons acting under color of state law.
Section 1985(3) does apply to conspiracies under color of state law, as well as private
conspiracies. See, e.g., Bray,

506 U.S. at 268

113 S.Ct. at 758

 ("In Griffin this Court held, reversing
a 20-year-old precedent, see Collins v. Hardyman,

341 U.S. 651

71 S.Ct. 937

95 L.Ed. 1253

(1951), that § 1985(3) reaches not only conspiracies under color of state law, but also purely private
conspiracies."). The Equal Protection Clause of the Fourteenth Amendment protects women from
sex discrimination by persons acting under color of state law. See Craig,

429 U.S. at 197

97 S.Ct. at 457

. Moreover, section 5 of the Fourteenth Amendment empowers Congress "to enforce, by
appropriate legislation, the provisions" of that amendment, including the Equal Protection Clause.
We believe that Congress had ample authority under Section 5 of the Fourteenth Amendment to
include women within the scope of § 1985(3)'s protection, at least where conspiracies to
discriminate against them through action under color of state law are involved.2
2
In his dissenting opinion, Judge Tjoflat flails away at two strawmen he has constructed. One
he labels "the broad reading of the majority's holding" and describes as the application of §
1985(3) to all conspiracies to commit torts arising out of sex-based animus. That broad reading
strawman goes beyond the facts of this case and our holding in it. This case does not involve all
such conspiracies; it involves only conspiracies by state actors utilizing state action. As Judge
Tjoflat himself concedes, § 1985(3) may be applied constitutionally to such a conspiracy. We
leave for another day and case the issue of whether it may be applied constitutionally to other
sex-based animus conspiracies.
Judge Tjoflat's dissenting opinion also attacks the "read more narrowly"
strawman, which he characterizes as inserting a state action requirement into § 1985(3).
We do not insert such a requirement into the statute. Instead, we simply recognize that
where a conspiracy that is covered by the statute does involve state action, the issue of
For all of the reasons stated, we hold that sex-based conspiracies against women are
actionable under § 1985(3), and that where the conspiracies involve state action, as the one alleged
in this case did, Congress clearly had the constitutional authority to make them actionable. We
reverse that portion of the district court's summary judgment that is to the contrary.
III. WHETHER THE COMMUNITY REDEVELOPMENT AGENCY AND THE CITY
SHOULD BE COUNTED TOGETHER AS AN "EMPLOYER" UNDER TITLE VII
A.
We turn now to the question whether the Community Redevelopment Agency (the "CRA")
and the City should be aggregated and considered as a single "employer" under Title VII's definition
of that term. Title VII defines "employer" as "a person engaged in an industry affecting commerce
who has fifteen or more employees for each working day in each of twenty or more calendar weeks
in the current or preceding calendar year, and any agent of such a person...." See 42 U.S.C. §
2000e(b). "Person" is defined as including "one or more individuals, governments, governmental
agencies, political subdivisions, labor unions, partnerships, associations, corporations, legal
representatives, mutual companies, joint-stock companies, trusts, unincorporated organizations,
trustees, trustees in cases under Title 11, or receivers." See 42 U.S.C. § 2000e(a). Because we have
treated the question of whether a defendant meets the statutory definition of "employer" as a
threshold jurisdictional matter under Title VII, see Virgo v. Riviera Beach Associates, Ltd.,

30 F.3d 1350

, 1359 (11th Cir.1994),3 a plaintiff must show that her "employer" had fifteen or more
whether Congress had authority to prohibit that conspiracy is an easy one. The
conspiracy alleged in this case was motivated by sex-based animus, and such
conspiracies fall within the scope of § 1985(3); the conspiracy also involved state action,
so Congress clearly had the authority to prohibit it. We need not and do not go beyond
those two holdings to decide this case.
3
Some judges in other circuits have doubted whether the number of employees is a
jurisdictional question. See Sharpe v. Jefferson Distrib. Co.,

148 F.3d 676

, 677-78 (7th
employees for the requisite period provided under the statute before her Title VII claims can be
reached.
It is undisputed that the CRA itself, during the relevant time, did not employ a sufficient
number of individuals to fall within the statutory definition of "employer." It is also undisputed that
if the CRA and the City are counted as one, collectively they employ enough people to meet that
definition. Lyes argues that the two entities should be aggregated and treated as a single "employer"
under Title VII, because the CRA and the City are interrelated in their operations and both entities
exercised substantial control over her conditions of employment. By contrast, the CRA and the City
contend that they are separate and distinct legal entities and should be treated as such, which would
remove the CRA from Title VII's definition of "employer" and effectively deprive us of jurisdiction
over that claim in this case. We must therefore decide what test applies in determining whether
separate state or local governmental entities should be counted as a single "employer" for purposes
of meeting the statutory definition under Title VII.
B.
We accord a liberal construction to the term "employer" under Title VII. See Virgo,

30 F.3d at 1359

; McKenzie v. Davenport-Harris Funeral Home,

834 F.2d 930

, 933 (11th Cir.1987);
Williams v. City of Montgomery,

742 F.2d 586

, 588 (11th Cir.1984). In keeping with this liberal
construction, we sometimes look beyond the nominal independence of an entity and ask whether two
or more ostensibly separate entities should be treated as a single, integrated enterprise when
determining whether a plaintiff's "employer" comes within the coverage of Title VII.
Cir.1998) (Easterbrook, J.) (dicta) (citing Steel Co. v. Citizens for a Better Env't,

523 U.S. 83

118 S.Ct. 1003

, 1009-12,

140 L.Ed.2d 210

 (1998)); EEOC v. St. Francis Xavier Parochial Sch.,

117 F.3d 621

, 624 (D.C.Cir.1997). The parties were not asked to brief this issue, and we will not
take this occasion to re-examine our circuit precedent holding that counting employees is a
jurisdictional matter.
We have identified three circumstances in which it is appropriate to aggregate multiple
entities for the purposes of counting employees. First, where two ostensibly separate entities are "
"highly integrated with respect to ownership and operations,' " we may count them together under
Title VII. McKenzie, 834 F.2d at 933 (quoting Fike v. Gold Kist, Inc.,

514 F.Supp. 722

, 726
(N.D.Ala.), aff'd,

664 F.2d 295

 (11th Cir.1981)). This is the "single employer" or "integrated
enterprise" test. Second, where two entities contract with each other for the performance of some
task, and one company retains sufficient control over the terms and conditions of employment of the
other company's employees, we may treat the entities as "joint employers" and aggregate them. See
Virgo,

30 F.3d at 1359-60

. This is the "joint employer" test. Third, where an employer delegates
sufficient control of some traditional rights over employees to a third party, we may treat the third
party as an agent of the employer and aggregate the two when counting employees. See Williams,
742 F.2d at 589. This is the "agency" test. See generally 2 Barbara Lindemann & Paul Grossman,
Employment Discrimination Law 1309-17 (3rd ed.1996).
The issue before us involves the "single employer" test.4 In determining whether two
non-governmental entities should be consolidated and counted as a single employer, we have applied
the standard promulgated in NLRA cases by the National Labor Relations Board. See, e.g.,
McKenzie, 834 F.2d at 933. This standard sets out four criteria for determining whether nominally
separate entities should be treated as an integrated enterprise. Under the so-called "NLRB test," we
look for "(1) interrelation of operations, (2) centralized control of labor relations, (3) common
4
Lyes argued before the district court and the panel that the "agency" test might also be
relevant in this case. Maybe, but some of the same factors discussed below that make the usual
"single employer" test (the NLRB test) difficult to apply in cases involving related governmental
entities may also make the usual agency test a poor fit where public entities are concerned. In
any event, we do not decide any agency issues, because the panel did not reach them, see

126 F.3d at

1385 n. 6, and we limited our briefing request to the "single employer" issue.
management, and (4) common ownership or financial control." Id.See also Radio and Television
Broad. Technicians Local Union 1264 v. Broadcast Serv. of Mobile, Inc.,

380 U.S. 255

, 256,

85 S.Ct. 876

, 877,

13 L.Ed.2d 789

 (1965).5
The four-factor NLRB "single employer" test was first applied in the Title VII context in
Baker v. Stuart Broad. Co.,

560 F.2d 389

 (8th Cir.1977). Since that time, most of the circuits
considering whether to integrate multiple entities under Title VII have seized upon that test. See,
e.g., Schweitzer v. Advanced Telemarketing Corp.,

104 F.3d 761

, 764 (5th Cir.1997); Cook v.
Arrowsmith Shelburne, Inc.,

69 F.3d 1235

, 1241 (2d Cir.1995); Childs v. Local 18, Int'l Bhd. of
Elec. Workers,

719 F.2d 1379

, 1382 (9th Cir.1983); Armbruster, 711 F.2d at 1337-38 (6th Cir.).
We, too, have applied the NLRB test. See McKenzie, 834 F.2d at 933. As one court employing the
test noted, the similarity in language between the NLRA and Title VII, and the fact that statute
served as a model for Title VII, makes decisions under the NLRA a useful resource for interpreting
the language of Title VII. See

29 U.S.C. § 152

(2); Armbruster, 711 F.2d at 1336 ("Since it is clear
that the framers of Title VII used the NLRA as its model, ... we find the similarity in language of
the Acts indicative of a willingness to allow the broad construction of the NLRA to provide
guidance in the determination of whether, under Title VII, two companies should be deemed to have
substantial identity and treated as a single employer.").
C.
However useful the four-factor NLRB "single employer" test may be in Title VII cases
where private entities are concerned, the question we face is the different one of whether that test
5
Courts applying the NLRB "single employer" test to private entities in Title VII cases have
held that not every factor need be present, and no single factor is controlling. See, e.g.,
Armbruster v. Quinn,

711 F.2d 1332

, 1337-38 (6th Cir.1983); Rivera v. Puerto Rican Home
Attendants Servs., Inc.,

922 F.Supp. 943

, 949 (S.D.N.Y.1996). See also Sakrete of N. Cal., Inc.
v. NLRB,

332 F.2d 902

, 905 n. 4 (9th Cir.1964) (stating the same rule in an NLRB case).
should be applied where state and local governmental entities are concerned. In answering that
question, we are aware of Congress's unmistakable intent that "Title VII principles be applied to
governmental and private employers alike." Dothard v. Rawlinson,

433 U.S. 321

, 331 n. 14,

97 S.Ct. 2720

, 2728 n. 14,

53 L.Ed.2d 786

 (1977). See also Owens v. Rush,

636 F.2d 283

, 287 (10th
Cir.1980) ("Governmental subdivisions were brought within the reach of Title VII" in the 1972
amendments to Title VII "so that "[a]ll state and local government employees would ... have access
to the remedies available under the Act.' ") (quoting 1972 U.S.C.C.A.N. 2137, 2152).
But we are also aware of the unmistakable reality that in certain key respects, governmental
entities are different from private ones. We will discuss some of those differences later. For now,
it is worth noting that equal treatment consists not only of treating like things alike, but also of
treating unlike things differently according to their differences. This is a proposition whose
provenance stretches back to antiquity, see, e.g., Aristotle, The Nicomachean Ethics, E.6.1131a-
1131b (M. Ostwald trans.1962), and it has as much logical force now as it had in Aristotle's day.
If we are to apply Title VII's principles equally to governmental and private employers in a fair and
intelligible manner, we must apply those principles in a way that respects the differences between
private and public entities. See also Riley v. County of Pike,

761 F.Supp. 74

, 76 (C.D.Ill.1991)
(applying the NLRB factors but noting that "[i]n addition to these factors, when the employers in
question are government entities, additional criteria come into play.").
The four NLRB factors are useful for counting employees of private entities under Title VII,
precisely because the test was "developed by the National Labor Relations Board to determine
whether consolidation of separate private corporations is proper in determining the relevant
employer" under the NLRA. Owens, 636 F.2d at 286 n. 2 (emphasis added). Of course, the NLRA
does not cover public employers, see

29 U.S.C. § 152

(2), and it is hardly surprising that a test
designed to operate in the context of private entities does not fit well cases involving governmental
entities. States are not the equivalent of corporations or companies, and local government bodies
are not the same as subsidiaries. With this in mind, we join those courts that have concluded that
"the [four-part NLRB] standard is not readily applicable to governmental subdivisions[.]" Trevino
v. Celanese Corp.,

701 F.2d 397

, 404 n. 10 (5th Cir.1983); Piper v. Junction City Hous. Auth.,

1995 WL 88232

, at *3 (D.Kan. Feb.1, 1995). See also Massey v. Emergency Assistance, Inc.,

724 F.2d 690

, 692 (8th Cir.1984) (Lay, C.J., dissenting) ("I think it clear such test is not applicable to a factual
situation concerning governmental entities.").6 But see Artis v. Francis Howell North Band Booster
Ass'n, Inc.,

161 F.3d 1178

 (8th Cir.1998) (applying, without discussion, the NLRB test to a private
entity and a government entity and deciding that they should not be combined); Massey, 724 F.2d
at 690-91 (affirming, without discussion, the district court's application of the NLRB test to a
government entity and the decision that it should not be aggregated with a private entity);
Vandermeer v. Douglas Cty.,

15 F.Supp.2d 970

, 974-75 (D.Nev.1998) (applying NLRB test to
county and local fire and paramedic districts, but noting that "[t]he fact that Nevada may consider
Douglas County and the Districts to be separate political subdivisions, while relevant, does not
control."); Rivera,

922 F.Supp. at 949

; County of Pike,

761 F.Supp. at 76-77

 (applying NLRB test
to Pike County and Pike County State's Attorney's office in ADEA case, but concluding that they
should not be treated as an integrated enterprise).
The most obvious way in which the NLRB "single employer" test is incompatible with cases
involving governmental entities involves the test's fourth factor—"common ownership or financial
6
In Dumas v. Town of Mt. Vernon, Ala.,

612 F.2d 974

, 980 n. 9 (5th Cir.1980), our
predecessor court declined without explanation to apply the NLRB test in a governmental
context. While we are not as sanguine as the panel about distinguishing Dumas, see

126 F.3d at 1385-86

, unlike the panel we are not bound by a prior panel decision. We consider the matter
anew.
control." McKenzie, 834 F.2d at 933. Governmental subdivisions such as counties or towns, or
smaller subdivisions such as local agencies, may share sources of ultimate political control or
funding, yet be wholly distinct with respect to their day-to-day operations or their control over
relationships with employees. Thus, the "common ownership or financial control" factor of the
NLRB test has no application to the usual case involving governmental subdivisions.
Nor is the NLRB test's third factor, "common management," McKenzie, 834 F.2d at 933,
readily applicable in the case of governmental entities. While it may be an appropriate yardstick in
some instances, in others two public entities may share managers or other employees while
remaining politically separate and distinct. In the present case, for example, each member of the
City Council also serves as a member of the CRA Board of Commissioners, but those city
councillors in their different capacity as commissioners comprise, by law, a distinct and independent
body. The Florida legislation that permits the members of a local governing body to declare
themselves a community redevelopment agency, explicitly provides that "such members constitute
the head of a legal entity, separate, distinct, and independent from the governing body of the county
or municipality."

Fla. Stat. Ann. § 163.357

(b) (West 1990). So the common management factor,
too, is not applicable in the context of governmental entities, at least not in a case like this one.7
In addition to the fact that these two prongs do not make sense in the context of state and
local governmental entities, there is another, perhaps more fundamental reason for not applying the
NLRB test. That reason involves federalism and comity concerns, which should play a significant
role in determining whether to treat as one body two governmental entities that are separate and
distinct under state law. When it comes to creating subordinate public bodies and defining their
7
See also Massey, 724 F.2d at 692 (Lay, C.J., dissenting) (arguing that the "centralized
control of labor relations" factor also is not relevant to an analysis involving local government
entities).
relationship to one another and to itself, " "the state is supreme and its legislative body, conforming
its action to the state Constitution, may do as it will.' " City of Trenton v. State of New Jersey,

262 U.S. 182

, 186-87,

43 S.Ct. 534

, 536,

67 L.Ed. 937

 (1923) (quoting Hunter v. Pittsburgh,

207 U.S. 161

, 178,

28 S.Ct. 40

, 46,

52 L.Ed. 151

 (1907)). Defining the nature and relationship of such bodies,
no less than determining their level of funding, is "uniquely an exercise of state sovereignty."
DeKalb Cty. Sch. Dist. v. Schrenko,

109 F.3d 680

, 689 (11th Cir.) (quoting Stanley v. Darlington
Cty. Sch. Dist.,

84 F.3d 707

, 716 (4th Cir.1996)), cert. denied,

118 S.Ct. 601

 (1997). We owe such
determinations by the state legislature not only deference, but great deference.
In rejecting the four-factor NLRB test as a means of determining whether two or more
governmental entities should be aggregated for the purposes of determining if they are an
"employer" under Title VII, we do not mean to suggest that the two remaining prongs of the test,
interrelation of operations and centralized control of labor relations, are not relevant to our inquiry
in any way. But where state or local governmental entities are involved, any indicia of integration
must be considered in a framework that is sensitive to the differences between governmental
subdivisions and private entities. See Trevino, 701 F.2d at 404 n. 10.
D.
Given our conclusion that the traditional NLRB four-factor test is not readily applicable to
the question of whether to aggregate state and local governmental entities for Title VII purposes, we
must decide what standard or test should apply in its place.8 Our inquiry must be guided by a
combination of "respect [for] the way a state chooses to structure its government[,]" McMillian v.
Johnson,

88 F.3d 1573

, 1580-81 (11th Cir.1996), aff'd,

520 U.S. 781

117 S.Ct. 1734

138 L.Ed.2d 8

We are not required to decide in this case, and we express no view about, whether this or
some different test should apply to aggregation issues involving federal governmental entities.
1 (1997) (citing City of St. Louis v. Praprotnik,

485 U.S. 112

, 126,

108 S.Ct. 915

, 925,

99 L.Ed.2d 107

 (1988)), as well as recognition of our obligation to accord a liberal construction to the terms of
Title VII. See, e.g., Virgo,

30 F.3d at 1359

.
The strong comity and federalism concerns we have mentioned require that we accord
substantial deference to a state lawmaking body's determination of whether two or more
governmental entities are separate and distinct. We should not brush aside a state's own distinctions
between its governmental subdivisions, because even ostensibly formal distinctions are part of a
government's ability to shape its own institutions within constitutional bounds, and we are obligated
to respect a state's right to do so. See McMillian,

88 F.3d at 1581

 ("[W]e heed the Supreme Court's
admonition that federal courts respect the way a state chooses to structure its government."). There
are few things closer to the core of a state's political being and its sovereignty than the authority and
right to define itself and its institutions in relation to each other. Of course, states cannot abuse that
power to evade federal law, but it is unlikely that a state would structure its state and local entities
with that purpose in mind. Such an evasive purpose is especially unlikely in a state like Florida,
which has enacted its own anti-discrimination legislation and made it applicable to local
governmental agencies. See

Fla. Stat. Ann. § 760.02

(6) (West 1997) (defining "person" under the
Florida Civil Rights Act of 1992 to include "any governmental entity or agency.").
We think that where a state legislative body creates a public entity and declares it to be
separate and distinct, that declaration should be entitled to a significant degree of deference,
amounting to a presumption that the public entity is indeed separate and distinct for purposes of Title
VII. The presumption may be rebutted in some instances. In particular, if it is established that a
state's purpose in creating or maintaining nominally separate entities was to evade the reach of the
federal employment discrimination laws, that alone is enough for those entities to be aggregated
when counting employees.
Even absent an intent to evade the application of federal law, we will aggregate two or more
governmental entities and treat them as a single Title VII "employer" where other factors so plainly
indicate integration that they clearly outweigh the presumption that the entities are distinct. In order
to determine which factors should be considered in deciding whether the plaintiff has carried her
burden of showing that the presumption has been clearly outweighed, we look to the factors courts
have considered in Title VII cases involving private employers.
Despite the primacy of the NLRB test as applied to private employers under Title VII,
"[c]ourts have used numerous formulations in assessing whether a defendant is an "employer' within
the meaning of Title VII and other employment discrimination statutes." Rivera,

922 F.Supp. at 949

.
Our review of the different factors that have been considered convinces us that they all share a
common focus: all of them seek to determine who (or which entity) is in control of the fundamental
aspects of the employment relationship that gave rise to the claim. See, e.g.,

id. at 949

 ("[The
different tests] all have in common a focus on one factor at issue here: the amount of control or
supervision a defendant exerts."); Armbruster, 711 F.2d at 1337 ("[C]ontrol over the elements of
labor relations is a central concern" of the NLRB test); Trevino, 701 F.2d at 404 (centralized control
of labor relations test "has been further refined to the point that [t]he critical question to be answered
then is: What entity made the final decisions regarding employment matters related to the person
claiming discrimination?") (internal quotation and citation omitted); Chester v. Northwest Iowa
Youth Emergency Servs. Ctr.,

869 F.Supp. 700

, 717-18 & n. 10 (N.D.Iowa 1994) (collecting cases
and noting that "control of employment decisions and environment" is one of the central features
of the tests).
Thus, two or more state or local governmental entities will be treated as a single "employer"
under Title VII where one entity exerts or shares control over the fundamental aspects of the
employment relationships of another entity, to such a substantial extent that it clearly outweighs the
presumption that the entities are distinct. Several factors will guide our determination of whether
the presumption in favor of the distinctness of the public entities is clearly outweighed—or, at the
summary judgment stage, whether a finder of fact could reasonably conclude that it is clearly
outweighed. As we have already noted, the NLRB factors of "interrelation of operations" and
"centralized control of labor operations," McKenzie, 834 F.2d at 933, may continue to be helpful in
the inquiry. Useful "indicia of control" may be drawn from the agency context, including: " "the
authority to hire, transfer, promote, discipline or discharge; the authority to establish work
schedules or direct work assignments; [and] the obligation to pay or the duty to train the charging
party.' " Oaks v. City of Fairhope, Ala.,

515 F.Supp. 1004

, 1035 (S.D.Ala.1981) (quoting Barbara
Schlei and Paul Grossmann, Employment Discrimination Law 846 (1st ed.1976)). Our list of factors
is not intended to be all inclusive, and consideration must be given to the totality of the
circumstances.
To summarize, we hold that when assessing whether multiple governmental entities are a
single "employer" under Title VII, we begin with the presumption that governmental subdivisions
denominated as separate and distinct under state law should not be aggregated for purposes of Title
VII. That presumption may be rebutted by evidence establishing that a governmental entity was
structured with the purpose of evading the reach of federal employment discrimination law. Absent
an evasive purpose, the presumption against aggregating separate public entities will control the
inquiry, unless it is clearly outweighed by factors manifestly indicating that the public entities are
so closely interrelated with respect to control of the fundamental aspects of the employment
relationship that they should be counted together under Title VII.
The standard we adopt is not whether a factfinder reasonably could conclude the plaintiff has
overcome the presumption. Instead, the standard is whether the factfinder reasonably could
conclude the plaintiff has clearly overcome the presumption. The adverb "clearly," which derives
from the federalism concerns we have discussed, is meant to be limiting. It is a thumb on the scale,
and sometimes it will be decisive because federalism concerns should sometimes be decisive.
Absent evidence of evasive purpose, in order to survive a motion for summary judgment, a plaintiff
will have to show that a reasonable fact finder could conclude that the presumption of distinctness
is clearly outweighed.
E.
Applying this test to the present case, we conclude that Lyes has not adduced sufficient
evidence to allow a reasonable fact finder to conclude that the City and the CRA should be treated
as a single "employer," bringing them within the coverage of Title VII.9 See, e.g., Anderson v.
Liberty Lobby, Inc.,

477 U.S. 242

, 252,

106 S.Ct. 2505

, 2512,

91 L.Ed.2d 202

 (1986) (the court must
ask "whether reasonable jurors could find by a preponderance of the evidence that the plaintiff is
entitled to a verdict...."); Raney v. Vinson Guard Serv., Inc.,

120 F.3d 1192

, 1198 (11th Cir.1997)
("[T]he plaintiff is effectively required to put forth her entire case at summary judgment [to]
persuade the court that a reasonable fact finder could rule in the plaintiff's favor.") (internal
9
In treating the question of whether the City and the CRA are a single Title VII "employer" as
a question of fact, we follow the practice of this Court in analogous Title VII cases involving
private employers. See, e.g., McKenzie, 834 F.2d at 933 ("Our role is to decide whether
McKenzie presented sufficient evidence to create a genuine issue concerning whether
Davenport-Harris and Protective should be treated as a single entity."). See also Garcia v.
Copenhaver, Bell & Assocs.,

104 F.3d 1256

, 1262-65 (11th Cir.1997) (drawing on Title VII
cases to hold that "the question of whether or not a defendant is an "employer' is a substantive
element of an ADEA claim and [is] intertwined with the question of jurisdiction. That being the
case, the resolution of the question must be made by the fact finder deciding the merits of the
claim.").
quotation and citation omitted).
Florida law establishes a presumption that the CRA and the City are separate and distinct
bodies. As the district court noted, the State legislature created community redevelopment agencies
as independent legal bodies. See

Fla. Stat. Ann. § 163.356

 et seq. (West 1990 & Supp.1999). With
the exception of certain enumerated powers, they are granted "all the powers necessary or
convenient to carry out and effectuate the purposes and provisions of [the community redevelopment
agency legislation]...."

Fla. Stat. Ann. § 163.358

 (West 1990 & Supp.1999). The five City Council
members also serve as the Board of Commissioners of the CRA. But

Fla. Stat. § 163.357

, which
permits this arrangement, also states that "[t]he members of the governing body shall be the
members of the agency, but such members constitute the head of a legal entity, separate, distinct,
and independent from the governing body of the county or municipality."

Fla. Stat. Ann. § 162.357

(1)(b) (West 1990). See also 1991 Fla. Op. Att'y. Gen. 148 (No. 91-49, 1991). The clear
distinction Florida law draws between the CRA and the City raises the presumption that they are
separate and distinct entities.10
Looking at the nature of Lyes's employment relationship itself, we conclude that she has not
produced evidence of interrelatedness with regard to control over employment sufficient to permit
a reasonable fact finder to conclude that the presumption that the governmental entities are distinct
is clearly outweighed. The CRA Board of Commissioners and its Executive Director control the
fundamental aspects of employment of the CRA's staff. The Executive Director hires and supervises
10
Moreover, these two entities are more than formally distinct. As the district court pointed
out, the CRA maintains separate bank accounts and records, and files its own tax returns. It
prepares its own budget and keeps its own offices. It is true that the City provides the majority
of the CRA's funding, but we have already concluded here that the source of a governmental
entity's funding is a poor indication of whether it should be aggregated with another; in any
event, the CRA still receives close to one-third of its funds from a source other than the City.
the staff, and the Board may hire, fire, and establish work schedules and assignments. The
Executive Director is employed by the CRA, and he serves at the pleasure of the CRA Board, not
at the pleasure of the City. Employees of the CRA receive their medical benefits, life insurance and
pension plans from the CRA, not from the City. Lyes was disciplined, suspended and eventually
terminated by the Executive Director of the CRA, and that decision was upheld by the CRA Board,
not by the City. All of these circumstances establish that the CRA retained and exercised control
over the fundamental aspects of its employee relations.
The panel below cited some evidence in support of its conclusion that the CRA and City
should be treated as a single employer. See

126 F.3d at 1386-87

. Of that evidence, two matters
deserve discussion. First, the panel pointed to a performance review of Lyes which was completed
by the Executive Director on a City form, listing the CRA as a department of the City. The
performance review was performed by Tony Smith, who was serving at the time as both City
Manager and interim Executive Director of the CRA. Given his dual duties, we do not think the fact
that he used City stationery has much, if any, significance. State law cannot be amended by
inferences drawn from printed forms.
Second, the panel noted that Neil Crilly, the CRA Executive Director, sought review from
the City's personnel director of his decision to discipline Lyes. We agree with the district court's
conclusion that there was no indication that this review had any binding effect. Instead, it appears
to have been an effort to seek the opinion of a third party as to whether Crilly had acted fairly in
disciplining Lyes. In this sense, Crilly's action is not unlike seeking advice from an expert in
employee relations.11
11
In reaching its conclusion, the panel also relied heavily on the NLRB factors of common
management and financial control. See

126 F.3d at 1387

. We have already concluded, however,
that these factors are not applicable to governmental entities.
Viewing the totality of the circumstances, we readily conclude that Lyes has not produced
evidence sufficient to allow a reasonable fact finder to conclude that the presumption that the CRA
and the City are separate entities is clearly outweighed and that they should therefore be treated as
a single employer under Title VII.
F.
Having decided the issues upon which we granted en banc review, we REMAND this case
to the panel for further proceedings not inconsistent with this decision.
TJOFLAT, Circuit Judge, concurring in part and dissenting in part, in which BIRCH, Circuit
Judge, joins:
This case requires us to interpret the scope of

42 U.S.C. § 1985

(3). Two possible
interpretations will sustain the majority's holding that the plaintiff in this case has a prima facie
claim. One is that Section 1985(3) applies to all conspiracies involving sex-based discriminatory
animus. The other is that it applies to conspiracies involving sex-based discriminatory animus in
which the conspirators are state actors.1 The former interpretation is unconstitutional; the latter
interpretation is ridiculous. I therefore dissent.
I.
The broad reading of the majority's holding is that any person injured by a conspiracy to
1
The majority avoids choosing between these two interpretations by holding that "sexbased
conspiracies against women are actionable under § 1985(3), and that where the conspiracies
involve state action ... Congress clearly had the constitutional authority to make them
actionable." Ante at 1371. The majority thus reserves the question of whether all sex-based
conspiracies against women are actionable under § 1985(3), or only those in which the
conspirators are state actors. As I will demonstrate, however, neither interpretation can be
sustained, and therefore the question reserved by the majority is nothing more than a choice
between two erroneous statements of law.
commit a tort arising out of sex-based animus has a cause of action under section 1985(3).2 Section
1985(3), as so interpreted, would be beyond Congress' power and thus unconstitutional.
The Supreme Court has warned us about "[t]he constitutional shoals that would lie in the
path of interpreting § 1985(3) as a general federal tort law." Griffin v. Breckenridge,

403 U.S. 88

,
102,

, 1798,

29 L.Ed.2d 338

 (1971). A broad reading of the majority's holding
founders on those shoals. Specifically, a rule that a conspiracy to commit a tort plus an injury plus
sex-based animus equals a section 1985(3) violation opens the door to substantial federal intrusion
into areas of law previously handled exclusively by state courts. For instance, every rape involving
more than one perpetrator would create a federal cause of action—the underlying tort would be
battery, and sex discrimination would be inherent in the conduct. Every domestic abuse case in
which the abuser was somehow aided by a friend would be actionable under the statute. A gang of
muggers who target women would also be amenable to suit under the statute. Indeed, even a group
of schoolboys who taunt a female classmate during recess would be subject to suit under section
1985(3).3
2
The statute creates a remedy for persons injured by a conspiracy to "depriv[e] ... any person
or class of persons of the equal protection of the laws."

42 U.S.C. § 1985

(3) (1994). The idea
that an individual can deprive another individual of the equal protection of the laws is somewhat
confusing. Our cases have made clear, however, that a person deprives another person of the
equal protection of the laws by committing a tort—any tort—against that person. See United
States v. Harris,

106 U.S. 629

, 643,

1 S.Ct. 601

, 612,

27 L.Ed. 290

 (1883) ("The only way,
therefore, in which one private person can deprive another of the equal protection of the laws is
by the commission of some offense against the laws which protect the rights of persons, as by
theft, burglary, arson, libel, assault, or murder."); McLellan v. Mississippi Power & Light Co.,

545 F.2d 919

, 925 (5th Cir.1977) (en banc).
3
I realize that if, under these factual scenarios, females were replaced with African
Americans, there would be no question that section 1985(3) would provide the victims with a
cause of action. As I demonstrate infra, however, there is a constitutional basis for applying
section 1985(3) in cases of racial animus (the Thirteenth Amendment), while no such basis exists
for extending the statute to cases of sex-based animus.
The greatest potential intrusion, however, would be in the area of workplace discrimination
and sexual harassment. The federal courts are already involved in these areas through Title VII. The
reach of Title VII, however, is limited in a variety of ways—for instance, it creates liability only for
"employers" as defined under the statute, see 42 U.S.C. § 2000e(b) (1994), and requires an aggrieved
party to file a complaint with the EEOC, see 42 U.S.C. § 2000e-5 (1994). The majority's holding
effectively casts aside these limitations.4 Every employee who is fired because of her sex, assuming
that more than one person was involved in the firing decision, would have a federal cause of action.
Male employees who touch women inappropriately—and any superiors who know about it and do
nothing—could be held jointly and severally liable if they acted in concert. Every group of men that
makes vulgar comments to women could be taken to federal court.5
This federal usurpation of state sovereignty implicates the core constitutional idea of
federalism. Under our constitutional system, the federal government is a government of limited
powers, and all powers not given to the federal government are reserved to the states. See U.S.
Const. amend. X. Federalism requires that any law enacted by the federal government must be
grounded in a specific grant of power under the Constitution.
It is clear that section 1985(3) was enacted, at least in part, under Congress' authority to
eliminate the "badges and incidents" of slavery pursuant to section 2 of the Thirteenth Amendment.
See U.S. Const. amend. XIII, § 2; Griffin,

403 U.S. at 104-05

91 S.Ct. at 1799

; see also Jones v.
Alfred H. Mayer Co.,

392 U.S. 409

, 437-44,

88 S.Ct. 2186

, 2202-05,

20 L.Ed.2d 1189

 (1968)
4
The potential use of section 1985(3) to get around the limitations of Title VII was explicitly
criticized in Great American Federal Savings & Loan Ass'n v. Novotny,

442 U.S. 366

, 375-76,

99 S.Ct. 2345

, 2350-51,

60 L.Ed.2d 957

 (1979).
5
The underlying torts in these three examples are, respectively, wrongful termination, battery,
and intentional infliction of emotional distress.
(describing Congress' Thirteenth Amendment powers). Congress' authority under the Thirteenth
Amendment is limited, however, to the prevention of discrimination on the basis of race. Thus, if
we are to interpret section 1985(3) as preventing sex discrimination, we must find some other
constitutional grounding.
One possibility is to base 1985(3) on Congress' power under section 5 of the Fourteenth
Amendment to enforce that amendment's guarantee of equal protection.6 See U.S. Const. amend.
XIV, §§ 1, 5. The Equal Protection Clause, however, is a guarantee of protection against unjust state
action; it does not reach the conduct of private individuals. Section 1985(3), under a broad reading
of the majority's holding, would reach all individuals regardless of whether they are private or public
actors. Thus, section 1985(3), read broadly, cannot be grounded in Congress' power under section
5 of the Fourteenth Amendment.7
The other usual suspect for legislation of this sort is the Commerce Clause. See U.S. Const.
art. I, § 8. There are, however, at least four good reasons why section 1985(3) cannot be grounded
in Congress' power to regulate interstate commerce. First, there is nothing in the legislative history
of section 1985(3) to suggest that Congress was acting pursuant to that power. Second, there is
nothing in the statute itself that ties it to the Commerce Clause—it does not, for instance, create a
cause of action for only those conspiracies that affect interstate commerce. Third, it is hard to
6
In Griffin, the Supreme Court explicitly avoided the question of whether section 1985(3) was
within Congress' power under section 5 of the Fourteenth Amendment. See Griffin,

403 U.S. at 107

91 S.Ct. at 1801

.
7
The only way in which section 1985(3), read broadly, could be understood as being based on
Congress' section 5 powers is if Congress had concluded that state deprivations of equal
protection were so rampant that, as an extreme prophylactic measure, it needed to pass a statute
sweeping broadly enough to reach private individuals involved in any tortious behavior. The
majority makes no showing—and I seriously doubt that it could make a showing—that this was
the case.
imagine how private tortious behavior of the sort that section 1985(3) prohibits (under a broad
reading of the majority's holding) could—at least in most cases—reasonably be said to involve
"Commerce ... among the several States." U.S. Const. art. I, § 8; see also United States v. Lopez,

514 U.S. 549

115 S.Ct. 1624

131 L.Ed.2d 626

 (1995). Finally, the Supreme Court has already
unequivocally stated that section 1985(3) is not grounded in the Commerce Clause. See United
Brotherhood of Carpenters and Joiners of Am. v. Scott,

, 833,

, 3358,

 (1983).
No other constitutional provisions appear to be colorable candidates. Thus, I am forced to
conclude that Congress does not have the authority to create a remedy for all persons harmed by
conspiratorial tortious conduct arising from sex-based animus.8 Consequently, section 1985(3),
insofar as it is interpreted to create such a remedy, is unconstitutional.
II.
The majority's holding may be read more narrowly to mean that section 1985(3), at least
when applied to cases involving sex-based discriminatory animus, covers only those conspiracies
in which the conspirators are state actors. Such a reading of the statute brings it within Congress'
Fourteenth Amendment powers and thus avoids the constitutional problems outlined in the previous
section. However, it also reads a state action requirement into section 1985(3), thereby contravening
nearly every imaginable canon of statutory interpretation.9
8
As the majority points out, see ante at 1368 - 69, a number of other circuits have interpreted
section 1985(3) as providing a remedy in federal court for conspiracies arising out of a sex-based
animus. None of these circuits, however, has explained whence Congress derives the authority
to create such a remedy.
9
Indeed, such a reading of section 1985(3) would be so far removed from true statutory
interpretation as to be nothing more than pure judicial activism, and consequently a violation of
the constitutional doctrine of separation of powers.
The plain language of the statute says absolutely nothing about a state action requirement.10
The statute does not use the terms "state actors," "persons acting under color of state law," or any
other language that would suggest that section 1985(3) applies only to agents of the state. Instead,
the statute applies to conspiracies involving "persons," without qualification.11 Turning to the
legislative history, the statements of various representatives and senators regarding the enactment
of section 1985(3) clearly indicate that the statute's reach was in no way intended to be limited to
state action—on the contrary, their statements make clear that they deliberately chose to target
"individuals" (regardless of governmental affiliation) in contrast to states. See Griffin,

403 U.S. at 99-101

91 S.Ct. at 1797-98

.
Most importantly, the Supreme Court precedent interpreting section 1985(3) unequivocally
forecloses any attempt to read a state action requirement into the statute. In the words of that Court,
"all indicators—text, companion provisions, and legislative history—point unwaveringly to §
10
Of course, the plain language of the statute also says nothing about discriminatory animus;
nevertheless, the Supreme Court has read a discriminatory animus requirement into the statute.
See Griffin,

403 U.S. at 102

 (holding that proof of a violation of section
1985(3) requires a showing of "racial, or perhaps otherwise class-based, invidiously
discriminatory animus behind the conspirators' action"). This limitation on section
1985(3)—like the state action limitation that the majority suggests—was based on a need to keep
the statute within Congress' constitutional authority. See

id. at 101-02

. The
discriminatory animus requirement, however, was well grounded in section 1985(3)'s legislative
history. In contrast, as I will discuss in this section, the majority's state action requirement is in
direct conflict with the statute's legislative history and the dictates of the Supreme Court.
11
The statutory language argument is made even more powerful when § 1985(3) is viewed in
its original context. Section 1985(3) was enacted as § 2 of the Civil Rights Act of 1871,

17 Stat. 13

. Section 1 of that Act—now codified as

42 U.S.C. § 1983

—explicitly requires that the
alleged deprivation of constitutional rights be "under color of state law." Not only does this
indicate that Congress was perfectly capable of including a state action requirement if it so
desired, but it also suggests a statutory scheme in which § 1 (now § 1983) was aimed at state
action while § 2 (now § 1985(3)) was aimed primarily at private conduct. Under such a scheme,
reading a state action requirement into section 1985(3) would radically undermine congressional
intent.
1985(3)'s coverage of private conspiracies." Griffin,

403 U.S. at 101

; see also
Scott,

463 U.S. at 834

103 S.Ct. at 3359

 ("[T]he Griffin opinion emphatically declared that [section
1985(3) ] was intended to reach private conspiracies that in no way involved the state."). It would
be hard to imagine a more unequivocal statement that a plaintiff alleging a violation of section
1985(3) need not allege state action.12
In sum, imposing a state action requirement on section 1985(3) is absurd.13 It is abundantly
clear that the presence or absence of state action is completely irrelevant to the existence of a claim
under section 1985(3). It may be true that Congress has the power, pursuant to section 5 of the
Fourteenth Amendment, to pass a statute providing a remedy for women injured as a result of a
conspiracy to commit a tort arising out of sex-based animus where the conspirators are state actors.
Section 1985(3) is not that statute. Therefore, the fact that the defendants in this case are state actors
provides no basis for a section 1985(3) claim by the plaintiff.14
12
The majority opinion is notably bereft of precedent for the proposition that state action has
any relevance in a section 1985(3) claim—because, at least after Griffin, there is no such
precedent to be found.
13
I also note the patent inconsistency of reading a single statute as requiring state action in one
situation (sex discrimination) but not another (race discrimination). Such a reading in essence
converts section 1985(3) into two statutes—one for race-based animus that applies to all persons,
and one for sex-based animus that applies only to state actors.
14
The majority steadfastly resists the idea that it might be reading a state action requirement
into section 1985(3). See ante at 1370 n. 2. This resistance results from a failure by the majority
to consider the consequences of its reasoning. The majority's reasoning seems to operate as
follows: Section 1985(3) applies to all conspiracies motivated by sex-based animus; this may
be unconstitutional as applied to private conspiracies, but the defendants lack standing to
challenge the statute on that ground because they are state actors. Private defendants would of
course have standing to raise such a challenge; if that challenge were successful, we would have
created a scheme in which section 1985(3) could be used against state actors but not against
private conspirators. I cannot see the difference between creating such a scheme and imposing a
state action requirement on section 1985(3). The fact that the majority creates only half of the
scheme today does not alter the nature of the scheme it is creating.
III.
The majority says that it will "leave for another day and case the issue of whether [section
1985(3) ] may be applied constitutionally to other sex-based animus conspiracies [besides those
involving state actors]." Ante at 1370. The majority fails to consider what will happen on that day
and in that case. A plaintiff complaining of sex discrimination will bring a section 1985(3) suit
against private persons; the court will be forced to decide whether the plaintiff has stated a claim.
If the court determines that the plaintiff has a claim, then it ignores the Constitution by sustaining
a remedy that Congress did not have the power to create. If the court determines that the plaintiff
does not have a claim because the defendants are not state actors, then that holding—combined with
the holding in this case—imposes a state action requirement on section 1985(3), in blatant violation
of statutory language, legislative history, and Supreme Court precedent. There are no other options.
The majority boxes us into this corner by trying to decide this case without defining the
scope of section 1985(3). Section 1985(3) clearly covers conspiracies (involving either private or
public actors) motivated by race-based animus.           I leave open—as the Supreme Court has
consistently done—the possibility that some form of animus other than race-based animus may be
covered by the statute. See Bray,

506 U.S. at 268-69

113 S.Ct. at 759

; Scott,

463 U.S. at 836-37

103 S.Ct. at 3360

; Griffin,

403 U.S. at

102 n. 9,

91 S.Ct. at

1798 n. 9. I cannot, however, discern
any reasoning that is true to both the statute and the Constitution under which I may hold that section
1985(3) covers conspiracies motivated by sex-based animus. I am therefore persuaded that the
plaintiff in this case does not have a prima facie case under section 1985(3). Insofar as the majority
holds that she does, I dissent. As for the remaining portions of the majority opinion (Parts I and III),
I wholeheartedly concur.
ANDERSON, Circuit Judge, concurring in part and dissenting in part:
I concur in all of Judge Carnes' opinion for the court, except Part III.E. In applying the new
test for determining whether two state or local governmental entities will be treated as a "single
employer" under Title VII, I agree with Judge Kravitch (and for the reasons she articulates) that
there remain genuine issues of material fact, thus making summary judgment inappropriate.
EDMONDSON, Circuit Judge, concurring in part and dissenting in part, in which COX,
BIRCH and DUBINA, Circuit Judges, join:
The Civil War came close to destroying the United States forever. The War was a brutal
physical contest, in which valor was common on both sides. Our great nation was preserved, but
at a terrible price. For example, in the four years of the War, the federal army of about 1,500,000
troops suffered 634,703 casualties (359,528 dead and 275,175 wounded).1 The War also cost the
federal government, at least, $15,000,000,000. I doubt that we can today appreciate the depth of
feeling about sacrifice and about loss (tinged with bitterness) that loomed over the country and
Congress in the decade following the War.
One issue over which the War had been fought was slavery. The fate of slaves, recently
freed at such severe cost to the Union, was a great concern during the Reconstruction era. Radical
Reconstruction was rough for southern whites. Some strongly resented the political influence of
freed slaves and of those persons working with freed slaves. There was resistance, and organizations
such as the Ku Klux Klan arose in the South. These organizations engaged in terroristic conduct.
This conduct was aimed at intimidating the freed slaves and their white allies. The allies were often
former union soldiers and mostly Republicans. These hard realities made up the situation Congress
(dominated by northern Republicans) faced in 1871 when it enacted the Ku Klux Klan Act,
including what is now

42 U.S.C. § 1985

(3).
1
By way of comparison, a much more populous United States, with a much larger military,
fighting worldwide, suffered a loss of 292,131 dead in World War II.
Our job in construing a statute is to determine congressional intent. In my view,

42 U.S.C. § 1985

(3) is the product of this concrete historical situation. The statute speaks to the concerns of
the situation. The text of the statute must be interpreted in the context of this history. Most
important, it is the qualities of the text when it was written—and not our response to it as modern
readers—that must be our guide. In addition, what we personally might like this statute to mean (or
the law, in general, to be) in the light of current circumstances and changed ideas has no rightful
place in our work: reinvention of the statute cannot properly be what we are after.
The question today is whether

42 U.S.C. § 1985

(3) extends to cover persons conspired
against because of their sex.2     Sexual discrimination was no issue in the Civil War or in
Reconstruction.3 And looking at the circumstances existing in 1871, I am unpersuaded that sexual
discrimination was in the mind of Congress when section 1985(3) was passed. The language of the
pertinent statute is pretty broad. But the Supreme Court has already indicated—in 1983—that
neither the language of section 1985(3) nor its legislative history readily answers questions about
the section's scope. Despite the statute's language, "it is a close question whether section 1985(3)
was intended to reach any class-based animus other than animus against Negroes and those who
championed their cause, most notably Republicans." United Bhd. of Carpenters & Joiners v. Scott,

, 3360,

 (1983) (construing "intent of the 1871
2
This case concerns the scope of the cause of action made available by section 1985(3) to
those injured by conspiracies formed "for the purpose of depriving, either directly or indirectly,
any person or class of persons of the equal protection of the laws, or of equal privileges and
immunities under the laws."
3
Plaintiff is a white woman. At its core, the section was intended to protect blacks and those
who supported blacks; thus, white women, along with everyone else, are doubtlessly protected
by the statute as long as they are being conspired against on account of helping a racial group.
The question presented in this case then is not whether women are protected by the statute (they
are), but whether the statute extends protection to women (or to anyone else) in this kind of case.
Congress" and holding that section does not reach conspirators motivated by bias against a group
on account of their economic views).4
Since 1983, no new information has come to light that would make this court better informed
about the intent of the 1871 Congress than the Supreme Court was informed in 1983.5 The
4
The Supreme Court in Scott was unpersuaded by a "plainlanguage" argument that
conspiracies not based on race are included in 1985(3). So am I.
The plain-language doctrine of statutory interpretation is usually best to determine
the true intention of Congress, but judges usually are looking at statutes enacted by—at
least roughly speaking—their own contemporaries. As contemporaries, we share with
the pertinent Congress basically the same environment of circumstances and facts that
surrounded the enactment of the statute; and when we see in the statute's words an
unambiguous plain meaning, we see that plain meaning in the light of our shared
contemporary conditions. But as I have observed, "[w]hen ... much older statutes are
being construed by modern courts, our response as modern readers to the words of the
statute may not be what the words meant to the Congress speaking at a very different
time." United States v. Steele,

147 F.3d 1316

, 1320-21 (11th Cir.1998) (en banc)
(Edmondson, J., concurring).
When we are looking at a statute that is almost 130 years-old, we know that, in
the meantime, conditions and ideas have changed; the passage of time is enough to inject
considerable uncertainty in my mind about the meaning of what the 1871 Congress said
with the statute. Just reading the words today is not enough to know what the words used
in 1871 meant to that Congress. Congress's intent at the time of the statute's enactment
cannot be reliably ascertained without taking into account the historical context of the
statute.
5
I do not say that the Supreme Court's words in Scott can be read as a holding which is
binding on us, but the Court's words have consequence. Furthermore, I agree with the Supreme
Court. The arguments made to support a construction of 1985(3) to extend beyond race are not
compelling.
Some argue that the section's protections track the protections of the Equal
Protection Clause of the Fourteenth Amendment, which the Supreme Court has construed
to go beyond race. But, section 1985(3) does not actually say (although it could have
done so à la

42 U.S.C. § 1983

) that it incorporates the Fourteenth Amendment or the
Constitution. And the Fourteenth Amendment was enacted by a different Congress (the
39th) than the Congress (the 42d) that enacted section 1985(3): words can mean different
things to different Congresses. In addition, the Supreme Court has already said that
"Equal Protection Clause jurisprudence is [not] automatically incorporated into §
1985(3)." Bray v. Alexandria Women's Health Clinic,

, 761
legislative-history equivalent of the Dead Sea Scrolls has not been discovered or called to our
n. 4,

 (1993). Even if one accepts that the Congress in 1871 intended the
words in section 1985(3) to mean the same thing as the Equal Protection Clause, it would
seem that the intention would be that the section's words mean the same thing as the
Equal Protection Clause's words were understood to mean in 1871, which may well be
different than what we understand the Clause to mean today. Furthermore, section
1985(3) was enacted—at least in substantial part—on the basis of the Thirteenth
Amendment, which deals directly with the badges and incidents of "slavery."
Others point out bits of legislative history that they say show a broader intent.
The legislative history—which is filled with references to freed slaves and their
supporters and to the Klan—is long. The bits that might be said to point beyond race are
very few and short. And, the statements of individual legislators are not ordinarily given
controlling effect. See, e.g., Brock v. Pierce County,

476 U.S. 253

106 S.Ct. 1834

, 1840-
41,

90 L.Ed.2d 248

 (1986).
Some others say that, even if it is true that the 1871 Congress did not have in
mind discrimination against persons other than black people when Congress enacted
section 1985(3), it is proper to extend the statute to cover reasonably comparable evils.
They cite Oncale v. Sundowner Offshore Services, Inc.,

523 U.S. 75

118 S.Ct. 998

140 L.Ed.2d 201

(1998). Oncale decided that a statute that undoubtedly was enacted to
prohibit sexual discrimination reaches all kinds of sexual discrimination, including
male-on-male discrimination. I say the Oncale reasoning would allow section 1985(3) to
reach all kinds of racial discrimination and not just discrimination against black persons,
although former slaves were precisely who Congress had in mind. But Oncale 's
reasoning does not license federal courts to start including a whole new category of
discrimination within a statute's coverage if Congress did not have that kind of
discrimination in mind at all.
Never has the Supreme Court decided whether section 1985(3) extends beyond
race to sex. Therefore, stare decisis and similar doctrines—which require courts to stand
by previously decided matters—can dictate nothing to us. Still, some people take what
they might call a realist view, observing that several of the present members of the
Supreme Court have individually said—at one time or another—that 1985(3) does cover
conspiracies other than those based on race. These "realists" see the handwriting on the
wall. They predict the Supreme Court ultimately will extend the statute to conspiracies
based on sex and say, therefore, that lower courts should take the step now. I think that
predicting what the Supreme Court will do in the future is hard. Justices are not bound
by their statements in concurrences or dissents or in decisions while serving on lower
courts. More important, our judicial duty, as I understand our promises, is to study
section 1985(3) using the usual legal methods and to decide the case accordingly and
independently. The job of a United States Circuit Judge is not just to bet on which way
the Supreme Court will come down later. I think the nonbinding observations of
individual Supreme Court justices are not so much to be counted by us, as they are to be
weighed.
attention. When we are construing statutes and, therefore, congressional intent, we must never allow
guessing to masquerade as interpretation. Also we, as judges, must always know what we do not
know. In my view, we cannot know with reasonable certainty that the 1871 Congress intended to
reach conspiracies other than those based on race.
To extend section 1985(3) to cover persons conspired against because of their sex would
represent an important decision on public policy. This kind of public-policy decision is beyond our
legitimate power when the congressional intent is so unclear. An ancient legal maxim says that
"where you doubt, do nothing."6 I think we should do nothing here. We should not err on the side
of an innovative expansion of an old statute. To do so is not just a mistake of principle, it is
dangerous. "The federal balance is a fragile one, and a false step in interpreting § 1985(3) risks
making a whole catalog of ordinary state crimes a concurrent violation of a single congressional
statute passed more than a century ago." Bray v. Alexandria Women's Health Clinic,

, 768,

 (1993) (Kennedy, J., concurring).
I believe that we do our duty when we admit that we do not know whether certain conduct
was intended by Congress to be covered by a statute or not. We cannot properly do more in this
case. I accept the principle that Congress and the President—political officers elected by the
American people—should decide public policy questions, such as the one underlying this case. The
political branches (and not the judicial branch) should decide whether a federal statute—along the
lines of

42 U.S.C. § 1985

(3): that is, a statute directed largely towards private conduct and without
expressed limits tied to state action, interstate commerce, federal funding and so on—is fitting and
6
Quod dubitas, ne feceris. S.S. Peloubet, Legal Maxims 253 (1985); Black's Law Dictionary,
1127 (5th Ed.1979).
needed to cover persons conspired against on account of their sex.7
When I disagree with the court, I commonly file no dissent. Dissenting is rarely the best use
of the judge's time and energy. Dissents can also be counterproductive in other ways. If the legal
principles involved seem to me to be important enough, I may write something, however. Even
then, because my dissent will govern no one (not even me) while the court's opinion will govern
everyone in three states, I sincerely hope that the court's view of the case is the correct one—setting
out the true law—and that my view is the incorrect view. I know that my colleagues' motives are
good. But for the reasons I have tried to explain briefly, I cannot today go along completely.
I concur in this court's opinion and judgment, except the part about

42 U.S.C. § 1985

(3).
HULL, Circuit Judge, concurring specially:
I concur in all of the majority opinion including the entire section addressing the § 1985(3)
issue. I write separately only to emphasize my view that, on the § 1985(3) issue, we do not write
on a clean slate. It is true, as the dissents point out, that the Supreme Court's decisions in Griffin v.
Breckenridge,

403 U.S. 88

29 L.Ed.2d 338

 (1971), United Brotherhood of
Carpenters and Joiners of America v. Scott,

 (1983),
and Bray v. Alexandria Women's Health Clinic,

 (1993)
do not expressly hold that § 1985(3) protects against conspiracies motivated by a gender-based
animus. However, the Supreme Court has strongly indicated that the broad language of § 1985(3)
prohibits gender-based conspiracies. In my view, the Supreme Court's direction on this issue is
unmistakable to the point that no principled way exists for this Court to hold, as suggested by the
7
By the way, although the Supreme Court signaled in 1983 that it was by no means clear that
section 1985(3) reached beyond race, Congress in 1991—when it amended many civil rights
statutes—did not expand the section to make it plain that nowadays the section's coverage ought
to go beyond race. But Congress has enacted other statutes that prohibit discrimination on the
basis of sex in a number of specific contexts.
dissents, that § 1985(3) prohibits only conspiracies motivated by racial animus.
In Griffin, the Supreme Court recited its interpretation of the broad language in § 1985(3).
The Court held that to prove a § 1985(3) violation, a plaintiff must show "some racial, or perhaps
otherwise class-based, invidiously discriminatory animus" behind the conspirators' actions. Griffin,

403 U.S. at 102

. In Griffin, the Court expressly reserved the question of whether §
1985(3) applied to conspiracies motivated by invidious intent other than race. Griffin,

403 U.S. at

102 n. 9,

. Nonetheless, as the majority in this case cogently observes, for the Court's
language "otherwise class-based, invidiously discriminatory animus" to have any meaning,
"otherwise class-based ... animus" must include animus directed against women—or men for that
matter.
Quite to the contrary of the inferences drawn by the dissents, in my view, Griffin, Scott, and
Bray, provide strong guidance that the Supreme Court views § 1985(3) as encompassing
gender-based conspiracies. Each of these cases presented the Court with the opportunity to limit the
reach of § 1985(3) to only conspiracies motivated by racial animus. That the Court declined all
three of these opportunities speaks loudly.
Griffin involved a racial conspiracy. As a race case, Griffin presented the perfect opportunity
for the Court to suggest that § 1985(3) was intended solely to protect African Americans. Indeed,
Griffin did not require the Court to discuss the scope of § 1985(3) beyond holding that it prohibited
conspiracies motivated by a racial animus. Nevertheless, the Court made its pronouncement that
§ 1985(3) required a showing of "some racial, or perhaps otherwise classbased, invidiously
discriminatory animus." Griffin,

403 U.S. at 102

. The Court's reasoning behind this
statement is apparent. Obviously, § 1985(3) was intended, first and foremost, to address the plight
of African Americans during Reconstruction, but the language used in the statute suggests that it
encompasses a broader range of conspiracies—thus, the phrase "otherwise class-based, invidiously
discriminatory animus."
Likewise, in Scott, the Supreme Court declined again to interpret § 1985(3) as prohibiting
only race-based conspiracies. As noted by the majority in the instant case, the Court in Scott held
that § 1985(3) does not prohibit "conspiracies motivated by economic or commercial animus."

463 U.S. at 836

. In doing so, the Court recited legislative history supporting the notion
that "[t]he predominate purpose of § 1985(3) was to combat the prevalent animus against Negroes
and their supporters." Id. Despite the Court's emphasis of this legislative history, the Court declined
the opportunity to limit the applicability of § 1985(3) to race-based conspiracies.
Moreover, Bray presents the clearest illustration of the Supreme Court's implicit recognition
that § 1985(3) prohibits conspiracies motivated by a gender-based animus. In Bray, the Court
addressed the issue of whether opposition to abortion constitutes "class based, invidiously
discriminatory animus."

506 U.S. at 269

. The district court in Bray had concluded
that such a conspiracy amounted to discrimination against the " "class' of women seeking abortions,"
and the plaintiffs argued that opposition to abortion included an animus against "women in general."

Id.

 Obviously, the Court's easiest and most direct path to reversing the district court, and rejecting
the plaintiffs' argument, would have been to hold simply that § 1985(3) does not prohibit
gender-based conspiracies. However, the Court did not follow this path.1 Instead, Justice Scalia's
majority opinion undertook a protracted, detailed analysis to reject the conclusion that opposition
to abortion equates with discrimination against women. The Court's opinion in Bray reflects the
unavoidable facts that (1) § 1985(3)'s broad language does not itself suggest that it is limited only
1
Importantly, in none of the five separate opinions in Bray did any of the Justices opine that
gender-based conspiracies were beyond the reach of § 1985(3).
to race-based conspiracies and (2) the Supreme Court's statement in Griffin that § 1985(3) requires
proof of a "class-based animus" recognizes the broad sweep of the protection afforded by § 1985(3).
In the instant case, the dissents do not sufficiently heed the Supreme Court's guidance on this
issue. One dissent admonishes that "[t]he job of a United States Circuit Judge is not just to bet on
which way the Supreme Court will come down later." Similarly, the other dissent describes this
issue simply as "open."
In my view, interpreting Griffin, Scott, and Bray as providing strong guidance on this issue
does not involve betting "on which way the Supreme Court will come down later," nor is this issue
simply "open." Instead, the majority's conclusion that gender is a class protected under § 1985(3)
is the only holding that follows in a principled way the Court's statement in Griffin and its
reaffirmation of this statement in Scott and Bray. The lack of a direct holding leaves this Court at
best "wiggle room" to write around the Court's statement in Griffin. However, Griffin, Scott, and
Bray, in my opinion, direct us to the path to follow in § 1985(3) cases, and stare decisis dictates that
we should follow it—until the Supreme Court tells us otherwise. In this regard, I note Justice
Souter's observation in his opinion in Bray that "I know of no reason that would exempt us from the
counsel of stare decisis in adhering to this settled statutory construction ... which Congress is free
to change if it should think our prior reading unsound."

506 U.S. at 289

 (Souter J.,
concurring in part and dissenting in part).
Griffin, Scott, and Bray have guided seven circuits to conclude that § 1985(3) embraces suits
premised on gender-based conspiracies. Likewise, this Supreme Court precedent is why four
Supreme Court Justices have stated recently their views that § 1985(3) encompasses gender-based
conspiracies. Bray,

506 U.S. at 295

 (Souter, J., concurring in part and dissenting in
part) ("To be sure, there is some resonance between Griffin 's animus requirement and those
constitutional equal protection cases that deal with classifications calling for strict or heightened
scrutiny ... [such as] race, national origin, alienage, gender, or illegitimacy."); Bray,

506 U.S. at 319

 (Stevens, J., dissenting) ("The text of the statute provides no basis for excluding from
its coverage any cognizable class of persons who are entitled to the equal protection of the laws.");
Bray,

506 U.S. at 348