Concept
Twombly’s Plausibility Standard
Sufficiency of the pleading
Twombly’s Plausibility Standard
In Bell Atlantic Corp. v. Twombly, decided in 2007, the Supreme Court read Rule 8(a)(2) to require enough facts to state a claim to relief that is plausible on its face, and it dismissed a complaint whose claim was merely conceivable.1 The case arose in antitrust, but the Court later confirmed that the holding governs every civil action,2 and many scholars read Twombly and Iqbal together as raising the pleading bar in all civil cases.3 Twombly therefore asks more of a complaint than notice of a theory: the court examines the facts alleged and asks what they support.
The case grew out of the breakup of the Bell system. The 1984 divestiture of AT&T’s local telephone business created seven regional monopolies, the incumbent local exchange carriers, and the Telecommunications Act of 1996 imposed duties on them to help competitive local exchange carriers enter their markets.4 Subscribers to local telephone and high-speed internet service then sued the incumbents as a putative class, alleging antitrust violations over a period of seven years.5
The complaint relied on two kinds of allegation. It alleged parallel conduct, meaning that the incumbents behaved alike in resisting the new entrants and in declining to compete in one another’s territories, and it alleged that they had agreed to behave that way.6 Parallel conduct carries an argument only so far, because it is consistent with conspiracy but just as much in line with rational business strategy that each firm adopts alone in response to common perceptions of the market.7
The district court dismissed the complaint under Rule 12(b)(6), and the Second Circuit reversed.8 The Supreme Court reversed in turn. It agreed with the district court that the facts did not suggest an antitrust conspiracy under either theory of the complaint, so the complaint failed to state a section 1 claim.9
Three ideas carry the Court’s reasoning, and the parts that follow take them in turn: what a court does with a legal conclusion, what plausibility demands, and why the Court retired the “no set of facts” language from Conley v. Gibson.
Labels and Conclusions Fall Short
Rule 8(a)(2) asks for a short and plain statement of the claim showing that the pleader is entitled to relief, and Twombly read that demand as setting both a ceiling and a floor. A complaint attacked under Rule 12(b)(6) needs no detailed factual allegations, but it must provide more than labels and conclusions, and a formulaic recitation of the elements of a cause of action will not do.10 The Court asked only that some pleaded fact stand behind each legal label. It did not ask for evidence or a full narrative.
The line follows from what a court does with a complaint on a motion to dismiss. With no evidence before it, the court takes the factual allegations as true, but it is “not bound to accept as true a legal conclusion couched as a factual allegation,” a limit Twombly drew from Papasan v. Allain.11 A conclusion is the very thing the facts exist to establish. If a court accepted it automatically, every plaintiff could survive by restating each element in a sentence, and Rule 12(b)(6) would screen out nothing. Later cases put the point as a division of labor: legal conclusions can provide the framework of a complaint, but factual allegations must support them.12
The Court also saw a practical reason to insist on facts. Supervising discovery is difficult, and costly discovery can push defendants to settle even weak claims.13 Twombly itself showed the scale: the plaintiffs represented a putative class of at least 90 percent of all subscribers to local telephone or high-speed internet service in the continental United States, and they sued America’s largest telecommunications firms over a period of seven years.14
The Twombly complaint shows a label at work. Section 1 reaches a “contract, combination or conspiracy,” and the complaint alleged one in nearly the statute’s words: the incumbents conspired to prevent competitive entry by, among other things, “agreeing not to compete with one another” and “allocating customers and markets to one another.”15 That sentence restates the element. The facts beneath it were the incumbents’ parallel conduct, and parallel conduct is ambiguous: it is consistent with conspiracy but just as much in line with a wide swath of rational and competitive business strategy unilaterally prompted by common perceptions of the market.16 An allegation of parallel conduct is thus much like a naked assertion of conspiracy. It gets the complaint close to stating a claim, but without some further factual enhancement it stops short of the line between possibility and plausibility.17
The Court noted that the parties themselves agreed on what such enhancement would look like: “complex and historically unprecedented changes in pricing structure made at the very same time by multiple competitors, and made for no other discernible reason” would support a plausible inference of conspiracy.18 It found nothing comparable in the Twombly complaint, whose facts did not suggest conspiracy under either of its theories.19
A Seventh Circuit antitrust case shows the same line from the pleader’s side. Those plaintiffs went beyond parallel conduct and alleged that all at once the defendants changed pricing structures that had been heterogeneous and complex to a uniform structure and then simultaneously raised their prices by a third, and that such detailed, simultaneous action was implausible without agreement. The court held that the plaintiffs should be allowed to proceed to discovery.20 The difference lies in the conduct. Parallel refusals to enter one another’s markets do not even suggest interdependence, since most firms do not enter every market an outsider might regard as profitable, whereas identical follow-the-leader pricing supports a far stronger inference.21
Two errors recur in student answers. The first reasons that because a court assumes the complaint’s allegations are true, the sentence “the defendants agreed not to compete” settles the agreement element. The assumption of truth attaches to facts, and a bare assertion of agreement is a conclusion about what the facts mean. The second reads Twombly to demand dates, meetings, and participants. The Court required no detailed factual allegations; it required facts that give the label content.22 How strongly those facts must point toward liability is the subject of the next part.
Plausible, Not Merely Possible
Twombly’s test asks whether the complaint pleads enough facts to state a claim to relief that is plausible on its face. A plaintiff whose facts leave the claim merely conceivable has not nudged it “across the line from conceivable to plausible,” and the complaint must be dismissed.24 Iqbal later restated the test: a claim has facial plausibility when the plaintiff pleads factual content that allows the court to draw the reasonable inference that the defendant is liable for the misconduct alleged.25 Three points keep that formula from being misread.
First, plausibility asks for less than probability. A well-pleaded complaint may proceed even if it strikes a savvy judge that proof of the alleged facts is improbable and recovery very remote and unlikely.26 The structure of the motion explains why. The judge has only the complaint, with no testimony or documents, so predicting that the plaintiff will lose would mean choosing whom to believe with nothing to base the choice on. Plausibility asks a smaller question: whether the pleaded facts raise a reasonable expectation that discovery will reveal evidence on the elements of the claim.27 The Court confirmed two years later, in a non-antitrust case, that plausibility “is not akin to a ‘probability requirement,’” though it “asks for more than a sheer possibility that the defendant has acted unlawfully.”28 A student who writes that a complaint fails because the plaintiff will probably not prove its facts has applied a test the Court disclaimed. A student who writes that any conceivable claim survives has ignored the demand for plausibility.
Second, the court assumes the pleaded facts are true even when it doubts them. The court must take the factual allegations as true,29 and its doubt about proof is no ground for dismissal.30 Two questions must stay apart. Whether the alleged events happened is a question of proof, reserved for discovery, summary judgment, and trial. What those events would suggest if they did happen is a question about the complaint, and the court can answer it on the pleadings.
Third, facts that fit liability no better than they fit innocent conduct leave the claim merely conceivable. That point decided Twombly. Parallel conduct gets a complaint close to a claim but stops short of the line between possibility and plausibility.31 The incumbents’ failure to enter one another’s territories shows why. “[F]irms do not expand without limit and none of them enters every market that an outside observer might regard as profitable,” so Congress may have expected some incumbents to compete in other incumbents’ territories, “but the disappointment does not make conspiracy plausible.”32 Parallel refusals to enter a market do not even suggest interdependent conduct, while identical follow-the-leader pricing creates a much stronger set of inferences.33
Matrixx Initiatives, Inc. v. Siracusano shows the standard met. Shareholders alleged that Matrixx told the market its revenues would rise 50 and then 80 percent while it held information indicating a significant risk to its leading product, and that it publicly dismissed reports linking Zicam to anosmia, the loss of smell, while holding evidence of a biological link between Zicam’s key ingredient and anosmia and having conducted no studies to disprove that link.34 Zicam allegedly accounted for 70 percent of Matrixx’s sales, and viewed as a whole the complaint alleged facts suggesting a significant risk to the commercial viability of Matrixx’s leading product.35 The two complaints differ in what their facts point toward. The incumbents’ conduct in Twombly fit ordinary self-interest as well as it fit conspiracy, so it pointed nowhere in particular. Each fact in Matrixx pointed the same way, toward undisclosed information that mattered to investors. Neither case turned on how likely the plaintiff was to win.
The Court drew the line at plausibility because of when discovery costs arrive. When the allegations in a complaint, however true, could not raise a claim of entitlement to relief, “this basic deficiency should . . . be exposed at the point of minimum expenditure of time and money by the parties and the court.”36 Careful case management and summary judgment come after the money is spent, and the threat of discovery expense will push cost-conscious defendants to settle even anemic cases before those stages arrive.37 The same reasoning explains why the line stops short of probability. Sorting likely claims from unlikely ones requires weighing evidence, and that task belongs to later stages, when evidence exists. Where the pleaded conduct fits lawful behavior as well as liability, the complaint gives no reason to open discovery. Where it points toward liability, a reasonable expectation of finding evidence justifies the expense.
Plausibility is a judgment, and the judge makes it. The court draws on its judicial experience and common sense,38 so the answer can vary with the judge’s predispositions. One commentator calls plausibility a subjective assessment strongly influenced by the decision maker’s predispositions, and observes that a plaintiff’s ability to survive a motion to dismiss now turns much more on whether the trial judge is sympathetic to the underlying cause of action.39 Courts keep the inquiry anchored by working through the elements, as in Cavanaugh v. Bartelt, where the question was whether the facts alleged raised a reasonable expectation that discovery would reveal evidence substantiating the necessary elements of the claim.40
Retiring Conley
Twombly’s plausibility test seems to collide with the most often quoted sentence in federal pleading law. 44 Conley v. Gibson said that a complaint should not be dismissed for failure to state a claim “unless it appears beyond doubt that the plaintiff can prove no set of facts in support of his claim which would entitle him to relief.”45
Read in isolation, the Twombly Court said, that language suggests any statement revealing the theory of the claim will suffice unless its factual impossibility shows on the face of the pleadings, and the court of appeals appears to have read Conley that way when it formulated its pleading standard.46 Such an approach, the Court observed, would dispense with any showing of a “reasonably founded hope” that the plaintiff could make a case: “Mr. Micawber’s optimism would be enough.”47 A plaintiff can nearly always hope that discovery will turn something up. The literal test would therefore dismiss only a complaint that disproved itself, and it could not enforce the line, drawn in the preceding part, between a pleaded fact and a pleaded conclusion.
Justice Souter, writing for the majority, took care to say that Conley itself was not being overturned. It was the lower courts, he said, that had taken the “no set of facts” language out of context.48 After puzzling the profession for 50 years, the famous observation “has earned its retirement,” and the phrase is best forgotten as “an incomplete, negative gloss on an accepted pleading standard.”49 Sources that say Twombly abrogated Conley therefore mean the phrase,50 since the Court’s own account retires one sentence and leaves the decision in place.
What survives of the sentence is a statement about proof. “[O]nce a claim has been stated adequately, it may be supported by showing any set of facts consistent with the allegations in the complaint.”51 Conley’s language thus describes how freely a plaintiff may prove what an adequately stated claim asserts, and it no longer sets the minimum a complaint must plead. The two questions arise at different times: the first before any evidence exists and the second as the plaintiff assembles it.
Courts apply the sentence in just this way. Juliana v. United States recited both halves of the standard together: a claim is facially plausible when the pleaded facts allow a reasonable inference of liability, and once a claim has been stated adequately it may be supported by any set of facts consistent with the allegations.52
Retiring the phrase did not raise the floor to a prima facie case. In Swierkiewicz v. Sorema, N.A., the Court held that an employment discrimination plaintiff need not plead facts establishing a prima facie case, because the McDonnell Douglas framework does not apply in every employment discrimination case.53 Twombly distinguished that case rather than overruling it. Swierkiewicz reversed a court of appeals that had insisted on “specific facts” beyond those needed to state the claim, while Twombly requires “only enough facts to state a claim to relief that is plausible on its face.”54 The Rules demand more particularity only where they say so, as Rule 9(b) does for fraud or mistake: “a party must state with particularity the circumstances constituting fraud or mistake.”55
How far Swierkiewicz survives in discrimination cases remains unsettled. The plausibility standard has generated substantial confusion in employment cases and has called the viability of Swierkiewicz into question.56 Lower courts have divided: some require only the ultimate elements of the claim, others look for facts that would establish a prima facie case, and the Second Circuit has taken a middle course, requiring facts that “give plausible support to a minimal inference of discriminatory motivation.”57
The dissent attacked this account at each step. It argued that the Conley Court “was interested in what a complaint must contain, not what it may contain,” and that the Court had said without qualification that it was “appraising the sufficiency of the complaint.”58 On that view the majority’s reading of the phrase as a statement about permissible proof could not be what Conley meant.
The dissent read Conley in light of three earlier decisions, Leimer, Continental Collieries, and Dioguardi.59 Leimer, for example, held that a complaint should not be dismissed “except where it appears to a certainty that the plaintiff would be entitled to no relief under any state of facts which could be proved,” and it added that “[n]o matter how improbable it may be that she can prove her claim, she is entitled to an opportunity to make the attempt.”60 Seen that way, Conley’s phrase “reflects a philosophy that, unlike in the days of code pleading, separating the wheat from the chaff is a task assigned to the pretrial and trial process,” and it captures “the policy choice embodied in the Federal Rules and binding on the federal courts.”61
Two further objections completed the dissent’s case. In antitrust cases, it said, “the proof is largely in the hands of the alleged conspirators,” so “dismissals prior to giving the plaintiff ample opportunity for discovery should be granted very sparingly.”62 And it charged that the majority’s reading of Rule 12(b)(6) “seems to be driven by the majority’s appraisal of the plausibility of the ultimate factual allegation rather than its legal sufficiency.”63
The two opinions weigh opposing costs. The majority feared discovery that forces defendants to settle even anemic claims.64 The dissent feared dismissing claims before discovery when the proof lies with the defendants. A pleading standard decides which of those errors courts will tolerate more, and plausibility accepts the second to avoid the first.