Bell Atlantic Corp. v. Twombly

People and parties

Majority opinion author

Supreme Court2007

Plaintiff

v.

Defendant

Select a participant for details.

The agreement the complaint had to explain

The telephone companies had stayed out of one another’s territories. To their customers, that looked like a deal. To the Supreme Court’s majority, it also looked like something former monopolists might do without making one. The difference decided whether the customers could demand the companies’ internal records. [1, pp. 566–569]

Twombly made the factual basis of that inference a pleading question. A plaintiff could no longer rely on the possibility that discovery would uncover an agreement. The complaint itself needed enough facts to make agreement plausible. Yet the Court expressly rejected a probability requirement: a properly pleaded claim could proceed even when proof looked unlikely. [1, pp. 555–557]

That distinction is the case’s continuing difficulty. Courts must assess what the alleged facts support before the plaintiff has the evidence that discovery might supply. For a lawyer drafting or attacking a complaint, the work lies in explaining the inference between the facts and the required element.

A market opened by law, still divided in practice

The 1984 breakup of AT&T left regional local-telephone monopolies. The Telecommunications Act of 1996 sought to open those markets. Incumbents had to make their networks available to entrants through resale, the leasing of network elements, or interconnection. Congress also established conditions under which incumbents could enter long-distance service. [1, pp. 549–550]

William Twombly and Lawrence Marcus sued on behalf of a proposed class of local-telephone and high-speed-internet subscribers. The defendants were BellSouth, Qwest, SBC, and Verizon, Bell Atlantic’s successor. The subscribers alleged that the companies had preserved their territories through two forms of coordinated conduct and thereby inflated customer charges. They sought treble damages and declaratory and injunctive relief. No conspiracy had been proved, and the case was still a putative class action. [1, p. 550 & n.1]

Keeping entrants out

The complaint alleged that incumbents obstructed new providers through unfavorable network-access arrangements, inferior connections, overcharging, and billing practices that damaged entrants’ customer relationships. The alleged coordination mattered because successful entry in one region could expose the potential for competition elsewhere. The majority considered that proposed connection and found it insufficient: each incumbent already had its own incentive to resist entry. [1, pp. 550–551, 566]

Staying out of one another’s territories

The second theory concerned the incumbents’ failure to compete against one another despite allegedly attractive opportunities in neighboring markets. Subscribers treated that pattern as evidence of market allocation. But the majority read the same complaint as describing substantial barriers facing anyone who entered another incumbent’s territory. The facts supporting the exclusion story also helped explain why expansion might be unattractive without an agreement. [1, pp. 551, 567–569]

How the complaint reached the Court

1. 2003

   The Southern District of New York dismissed the complaint, concluding that the alleged parallel conduct did not adequately support an agreement. [1, p. 552]

2. 2005

   The Second Circuit reversed. It declined to require pleading additional “plus factors” and relied on the possibility that plaintiffs could later establish collusion. [1, p. 553]

3. May 21, 2007

   The Supreme Court reversed, held the complaint insufficient, and remanded for proceedings consistent with its opinion. [1, p. 570]

4. May 18, 2009

   Ashcroft v. Iqbal confirmed that the Rule 8 pleading standard applies across federal civil actions, including discrimination suits. [2, p. 684]

What makes an agreement plausible?

Section 1 of the Sherman Act requires concerted action. Businesses can independently reach similar decisions, including decisions made with awareness of their rivals’ likely responses. Parallel behavior may support an inference of agreement, but it does not by itself establish one. [1, pp. 553–554]

The question was whether this complaint supplied enough factual context to move from parallel behavior to a plausible agreement. It arose under Rule 12(b)(6), before proof of the claim. The Court was deciding the sufficiency of the pleading, not resolving a factual contest about whether executives had actually colluded. [1, pp. 548–549, 554–557] [7, Rule 12(b)(6)]

“a short and plain statement of the claim showing that the pleader is entitled to relief”

Federal Rule of Civil Procedure 8(a)(2). [6]

Both sides could invoke that sentence. The plaintiffs emphasized brevity and notice. The majority emphasized the required showing of entitlement. Repeating the elements of a cause of action did not supply the factual grounds for it. [1, p. 555 & n.3]

The argument over when to demand more

The subscribers: let discovery test the allegation

The subscribers relied on the companies’ parallel behavior, market structure, opportunities for expansion, and the complaint’s express allegation of agreement. They argued that notice pleading did not require the evidentiary showing needed at summary judgment. Their reliance on Conley v. Gibson also left room for facts that discovery might uncover. [1, pp. 550–553, 560–561, 569–570]

The companies: the missing agreement is the claim

The opposing position was that the pleaded behavior reflected independent business incentives. A conspiracy label could not supply the missing factual connection between the companies. Allowing such a complaint to proceed would expose defendants to broad, costly discovery without an adequate basis for expecting evidence of an agreement. The United States supported the companies as an amicus; it was not bringing an enforcement action. [1, pp. 547–548, 558–560] [8, pp. 10–18]

At oral argument, the companies’ lawyer, Michael Kellogg, made the distinction unusually clear. Responding to Justice Breyer, he said the problem was not a lack of specificity: the complaint supplied detail, including color maps. His objection was that those facts did not support the inference the subscribers wanted. That was counsel’s argument, but it captures a useful drafting lesson: factual volume and inferential strength are different things. [9, transcript pp. 13–16]

The Court: factual support must come first

The majority accepted the need for facts suggesting agreement before discovery. It did not require the plaintiff to establish the agreement, produce direct evidence, or satisfy a probability threshold. The complaint’s defect was that the stated facts did not make its proposed inference plausible. [1, pp. 556–557, 569–570]

Seven votes to dismiss this complaint

Justice David Souter wrote for a seven-justice majority, joined by Chief Justice Roberts and Justices Scalia, Kennedy, Thomas, Breyer, and Alito. The Court reversed the Second Circuit and remanded. It held that the subscribers had failed to state a plausible section 1 claim. This was a defeat for the complaint, without a trial finding that the companies had acted independently. [1, pp. 547, 570]

Parallel facts still need a connecting explanation

For exclusion of entrants, the majority saw a strong independent incentive: each incumbent wanted to protect its own territory. The allegation that all incumbents benefited from preventing a successful entrant did not explain why they needed to coordinate. For mutual nonentry, the history of protected regional monopolies and the obstacles facing new providers supplied an alternative explanation. The complaint did not allege that expansion into another incumbent’s territory was more lucrative than opportunities the companies were already pursuing. [1, pp. 566–569]

One quotation, two readings

The complaint pointed to a reported remark by Qwest chief executive Richard Notebaert: entering another incumbent’s territory “might be a good way to turn a quick dollar but that doesn’t make it right.” To the subscribers, the remark supported their allegation of restraint among competitors. The majority looked at the surrounding published remarks cited in the complaint, which also described the entrant business model as unsustainable and the regulatory environment as unstable. [1, pp. 551, 568 n.13]

Stevens thought the ambiguity warranted asking Notebaert himself what he meant. He objected to resolving it through selected newspaper quotations before discovery. The disagreement gives the abstract pleading debate a concrete form: the same reported statement could be read as a clue to an agreement or as a business judgment about expansion. For a drafter, it also makes the full context of an incorporated source worth checking before relying on its most striking sentence. [1, pp. 591–593 (Stevens, J., dissenting)]

Conley’s familiar formula lost its force

Conley had supplied the often-quoted “no set of facts” language. Read literally, it could let a conclusory complaint survive whenever some unpleaded facts might later justify relief. Twombly retired that formulation as the minimum pleading test. It preserved the distinction between adequately stating a claim and later proving it with facts consistent with the pleading. Fair notice remained part of Rule 8. [1, pp. 555, 561–563]

Discovery costs shaped the timing

The majority considered the expense of searching the records of huge telecommunications companies for a conspiracy alleged over seven years. It worried that the cost itself could induce settlement of a weak claim. Waiting for summary judgment would not solve that problem: the expense would already have been incurred. Nor did the majority trust case management to cure a complaint that fell short of plausibility. [1, pp. 558–560]

Stevens’s objection: the evidence is on the other side

Justice John Paul Stevens dissented. Justice Ruth Bader Ginsburg joined him except for Part IV. Stevens agreed that parallel conduct alone was not unlawful. His disagreement concerned what the plaintiffs had alleged and what a court could demand before discovery. He treated the express allegation of agreement, supported by the surrounding circumstances, as an allegation of unlawful conduct that deserved a response. [1, pp. 570–573]

The majority treated the agreement language as a conclusion drawn from otherwise insufficient parallel conduct. Stevens regarded that treatment as an improper appraisal of factual plausibility. He favored controlling discovery and testing the evidence through later procedures, especially because evidence of a conspiracy would ordinarily be held by the alleged conspirators. He also argued that a procedural change of this scale belonged in the rulemaking process. [1, pp. 573, 579, 585–587]

The disagreement put the timing of factual scrutiny at the center of the case. Souter wanted a plausible basis for discovery before imposing its costs. Stevens wanted limited discovery before rejecting an allegation whose proof could be in the defendants’ files. Both required agreement; they differed over what plaintiffs had to supply at the outset. [1, pp. 558–560, 593]

What a lawyer should do with the rule

Start with the element the complaint has to establish. Here, that was agreement. More detail about high prices, market concentration, or aggressive conduct would not necessarily repair the absence of facts supporting coordination. The useful question is what each allegation adds to the inference the claim requires. [1, pp. 553–557, 566–569]

• Identify the factual connection. Explain how the alleged events support the required element, rather than relying on a legal label to connect them.
• Keep pleading and proof separate. Plausibility requires more than compatibility with wrongdoing. It does not demand that success be probable or that the complaint supply the evidence needed at trial. [1, p. 556]
• Distinguish Rule 8 from particularity. The Court expressly disclaimed a heightened requirement of detailed facts under Rule 9. Its observation that this complaint lacked a particular time, place, or participant did not establish a universal checklist for all claims. [1, p. 565 n.10; p. 569 n.14]
• Apply the rule to the actual claim. Later decisions separate insufficient facts from a missing legal citation and from matters allocated to an affirmative defense. [4] [5]

Direct evidence is not the only way across the threshold. In footnote 4, the Court noted the parties’ agreement that simultaneous, complex, historically unprecedented pricing changes, with no other discernible reason, could support a plausible conspiracy inference. That was an illustrative example, not a finding about these telephone companies. The point is that circumstances can do the work if their context makes coordination plausible. [1, p. 556 n.4]

Change one fact

Hypothetical: suppose a complaint described a particular communication in which competing executives agreed to divide territories, followed by conduct implementing that arrangement. That allegation would address the connection missing from this complaint and could materially strengthen plausibility. Merely adding that executives attended the same trade-association meeting would do much less. The content and context of the alleged communication would matter; this example does not predict the result of a different case. [1, pp. 556–557, 565 n.10, 567 n.12]

The rule after Twombly

Erickson: short allegations can be enough

Two weeks later, Erickson v. Pardus rejected a lower court’s treatment of a prisoner’s allegations of medical harm as too conclusory. Citing Twombly, the Court stressed fair notice, the acceptance of factual allegations, and liberal construction of a pro se complaint. It left other objections open. Brevity alone does not make an allegation a legal conclusion. [3, pp. 93–94]

Iqbal: the standard reaches beyond antitrust

In 2009, Ashcroft v. Iqbal expressly rejected an antitrust-only reading. It described two working principles: conclusions do not receive the assumption of truth, and the remaining factual allegations must plausibly support relief in context. That inquiry requires judicial experience and common sense. It still does not impose a probability requirement. [2, pp. 678–680, 684]

Johnson: adequate facts need no perfect legal label

In Johnson v. City of Shelby, the Court rejected the loss of a sufficiently pleaded constitutional claim merely because the complaint omitted an express reference to 42 U.S.C. § 1983. Twombly and Iqbal concerned factual sufficiency; the officers had supplied the factual basis for their claim. An imperfect statement of legal theory was a different problem. [4, pp. 11–12]

Cunningham: identify who must plead what

In 2025, Cunningham v. Cornell University held that plaintiffs asserting an ERISA prohibited-transaction claim under 29 U.S.C. § 1106(a)(1)(C) need plausibly allege its statutory elements, without negating the separate section 1108 exemptions. Those exemptions were affirmative defenses. Concerns about meritless suits and discovery costs did not justify changing that allocation. The decision did not abandon plausibility; it identified what had to be made plausible. [5, slip op. pp. 6–8, 13–15]

These decisions explain the federal Rule 8 framework. The treatment discussed here was checked on September 30, 2026 and is selective; it does not purport to survey every circuit’s application or state pleading rules.

Sources and citations

1. Bell Atlantic Corp. v. Twombly, 550 U.S. 544 (2007)

   Supreme Court · May 21, 2007 · Majority pp. 548–570; Stevens dissent pp. 570–597. Citations refer to printed U.S. Reports pages.

2. Ashcroft v. Iqbal, 556 U.S. 662 (2009)

   Supreme Court · May 18, 2009 · pp. 678–680, 684–687.

3. Erickson v. Pardus, 551 U.S. 89 (2007)

   Supreme Court · June 4, 2007 · pp. 93–94. Opinion reproduced by Justia.

4. Johnson v. City of Shelby, 574 U.S. 10 (2014)

   Supreme Court · November 10, 2014 · pp. 11–12. Opinion reproduced by Cornell LII.

5. Cunningham v. Cornell University, No. 23-1007

   Supreme Court · April 17, 2025 · Majority slip opinion pp. 6–8, 13–15; pagination excludes the syllabus.

6. Federal Rule of Civil Procedure 8

   Rule 8(a)(2), (d), (e) · Rule text via Cornell LII; accessed September 30, 2026.

7. Federal Rule of Civil Procedure 12

   Rule 12(b)(6), (d) · Rule text via Cornell LII; accessed September 30, 2026.

8. Brief for the United States as amicus curiae supporting petitioners

   Department of Justice · August 25, 2006 · pp. 10–18. Advocacy, not judicial findings.

9. Bell Atlantic Corp. v. Twombly, oral argument transcript

   Supreme Court · November 27, 2006 · pp. 13–16. Michael Kellogg’s argument for petitioners; questions and advocacy are not holdings.

Image credits

David Souter’s portrait is a generated monochrome illustration based on his official Supreme Court photograph. The Bell Atlantic identifier is a generated rendition of a historical mark, using the Freebie Supply reference. Neither is historical photographic evidence. No verified portrait of William Twombly was located; his entry uses the shared unavailable-portrait symbol.