Concept
Jurisdiction Versus the Merits
Federal question jurisdiction
Jurisdiction Versus the Merits
Everything so far has asked whether a complaint presents a federal question. A different question arises once the court is satisfied that it does: what happens when the federal claim the plaintiff has pleaded looks like a loser. A district judge who reads a complaint asserting a federal right and concludes that no such right exists is tempted to say that the case never belonged in federal court at all. That instinct conflates two inquiries that the arising-under cases keep separate, and the separation decides whether the plaintiff leaves with a judgment against him or with the freedom to sue again somewhere else. The subsections that follow take up three refinements of that separation: claims for which Congress supplied no express remedy, the two narrow situations in which a federal allegation is too weak even to confer power, and the motion that properly tests a claim's sufficiency.
The governing decision is Bell v. Hood, which reached the Supreme Court in an unusual posture. Plaintiffs sued agents of the Federal Bureau of Investigation for damages exceeding $3,000, alleging that the agents had imprisoned them without due process and had searched their premises and seized their possessions in violation of the Fourth and Fifth Amendments.1 The agents moved to dismiss for failure to state a cause of action and for summary judgment, but the district judge passed on neither motion; on his own motion he dismissed the suit for want of federal jurisdiction, reasoning that an action of this kind did not arise under the Constitution or laws of the United States, and the court of appeals affirmed on the same ground.2 What made the case hard in 1946 is that no statute then authorized a damages recovery against federal officers for a constitutional violation, so the lower courts had treated the absence of a remedy as an absence of power.
The Supreme Court reversed, and the holding has two halves that are worth holding apart. The first is that a case arises under federal law when the plaintiff alleges a right to relief directly under the Constitution or laws of the United States, whether or not the provision invoked expressly creates a civil cause of action.3 The second is remedial in the procedural sense: where a complaint is drawn to seek recovery under federal law but fails to state a proper cause of action, the correct disposition is a judgment on the merits rather than a dismissal for want of jurisdiction.4 A court must therefore assume jurisdiction in order to decide whether the allegations state a claim on which relief can be granted, and the probability that the plaintiff will ultimately lose does not subtract from the court's power to say so.
The reason for keeping the two apart begins with the structure of adjudication itself. Deciding whether the Fourth Amendment affords these plaintiffs damages is deciding a question of federal law, and deciding a question of federal law is an exercise of the judicial power that jurisdiction supplies; a court that answered the question in order to decide whether it could answer the question would be reasoning in a circle. The second reason is that the two dismissals do different things to the litigants. A merits judgment resolves the federal right and binds the parties, while a jurisdictional dismissal determines only that the suit was filed in the wrong court system, and saving statutes in every state give the plaintiff a grace period to refile in state court even after the limitations period has run.5 The third reason is about incentives, and it is the one that does the most work. Lack of subject-matter jurisdiction can be raised by any party at any stage and is never waived, so a rule that let a defendant defeat jurisdiction by denying that the plaintiff is entitled to win would hand every defendant a non-waivable objection to every contestable federal claim, and the federal courts would seldom have any arising-under jurisdiction at all.6 For that reason the pleading standard at the jurisdictional threshold is low: the plaintiff need make only a colorable, arguable assertion of a federal cause of action, not a viable one.7
The dissent of Chief Justice Stone and Justice Burton shows what the majority rejected. They would have held that where neither the constitutional provision nor any act of Congress affords a remedy to any person, a plaintiff's bare assertion that he is entitled to one cannot satisfy the jurisdictional requirement, so the lower courts had been right to dismiss for want of power.8 The disagreement is not about whether the plaintiffs could recover. It is about which judge decides that and under what label, and the majority's answer is that the question belongs to the merits docket of a court that already has jurisdiction.
The same line surfaced again in Powell v. McCormack, where a congressman excluded from his seat sued the House leadership. A suit arises under the Constitution, the Court repeated, if the claimant's claim will be sustained if the Constitution is given one construction and defeated if it is given another, and Powell's challenge to his exclusion satisfied that description.9 The district court had dismissed for want of jurisdiction of the subject matter on separation-of-powers grounds, and the Court corrected the label: any bar to reviewing the House's judgment came from the allocation of powers between the branches, which is a question of justiciability, and not from any failure to state a claim based on federal law.10 Jurisdiction existed under § 1331 because the complaint alleged that the House had violated Article I.11
A pair of employment examples shows the steps operating. An assembly-plant worker sues his employer under the federal Fair Labor Standards Act for unpaid overtime, and the employer answers that he is a manager rather than an "employee" within the Act and moves to dismiss for lack of subject-matter jurisdiction. The court should deny that motion, because federal law creates the cause of action the worker asserts and the employer's position is a denial that he can prove it; if the employer later wins on that point at summary judgment, the dismissal is on the merits.12 Change the pleading rather than the strength of the defense and the analysis flips. A driver injured by an interstate trucker on a federal highway who pleads only negligence has invoked no federal right, and the federal presence on the periphery of the dispute does not convert a state tort claim into one arising under § 1331, so dismissal there is jurisdictional.13 The variable is what the plaintiff claims entitles him to relief, not how likely he is to get it.
Claims Without an Express Remedy
The hardest version of this problem arises when the federal provision the plaintiff invokes says nothing at all about suing on it. The Fourth Amendment declares a right to be secure against unreasonable searches; it does not say that a person whose home is ransacked may collect money from the officers who did it. A judge facing that silence is tempted to reason that if federal law gives no remedy, there is no federal claim, and if there is no federal claim there is no federal question. The arising-under cases cut that chain at the first link. A case arises under federal law when the plaintiff alleges a right to relief directly under the Constitution or laws of the United States, whether or not the provision invoked expressly creates a civil cause of action, and the existence of jurisdiction does not depend on the presence of a remedial provision.15 Jurisdiction turns on what the plaintiff says federal law gives him, and whether federal law in fact gives it is the question the court takes jurisdiction in order to answer.
What makes that separation more than a formality is what happened to the reserved question afterward. Bell v. Hood was decided in 1946 against a legal background in which no statute authorized damages against federal officers for a constitutional violation, and the Court held only that the suit belonged in federal court, leaving the availability of damages to be decided there. Twenty-five years later the answer arrived. Narcotics agents entered Webster Bivens's apartment without a warrant, manacled him in front of his wife and children, searched the apartment, and took him to the federal courthouse to be interrogated and strip-searched; he sued each agent individually for $15,000, and the court of appeals held that no cause of action had been alleged.16 The Supreme Court reversed, holding that he was entitled to recover money damages for injuries suffered from the agents' violation of the Fourth Amendment, and that the federal question jurisdiction conferred by § 1331 over cases arising under the Constitution supports a cause of action for damages for violation of constitutionally protected interests.17 Notice what did not change between 1946 and 1971. The jurisdictional test was identical in both years; only the merits answer moved, and it moved because a court with jurisdiction was in a position to decide it.
The same relationship between right and remedy appears on the statutory side, where Congress frequently forbids conduct without saying who may sue over it. In J. I. Case Co. v. Borak a shareholder sued over a proxy solicitation that violated section 14(a) of the Securities Exchange Act, a provision that makes deceptive solicitation unlawful but says nothing about private suits.18 The Court treated the gap as an ordinary question of federal law rather than a hole in its power, reasoning that when a federal statute condemns an act as unlawful, the extent and nature of the legal consequences of that condemnation are federal questions left to judicial determination, and that "where legal rights have been invaded, and a federal statute provides for a general right to sue for such invasion, federal courts may use any available remedy to make good the wrong done."19 Whether a modern court would reach the same result on those facts is beside the present point. What matters here is the sequence: the court first had the case, and then decided what the statute supplied.
The reason the law is built this way becomes obvious once you try to run it the other way. If the availability of a remedy controlled jurisdiction, a district court would have to dismiss for want of power precisely those cases in which the existence of a new federal remedy is contested, which means no court could ever decide the question and federal law would be frozen wherever Congress had not spelled the answer out. That is the circularity the merits rule avoids: the court assumes power in order to decide the federal question, and deciding the federal question is what the power is for. The arrangement does not leave defendants exposed, because the ordinary consequence of a plaintiff who guesses wrong about the remedy is a judgment for the defendant on the merits, which ends the case and binds the parties. A defendant loses only the ability to characterize a weak claim as a defect in the court's authority.
The two poles differ in what the court is being asked to decide, not in how promising the claim looks. A plaintiff who seeks damages directly under a constitutional amendment invokes federal jurisdiction by that allegation alone, and the court must take the case before it can say whether such damages are available; that is the posture in which the Fourth Amendment damages action was eventually recognized.20 Now change the outcome of the merits question while holding the pleading constant. Suppose a consumer sues a manufacturer under a federal safety statute that imposes labeling duties and provides for enforcement only by an agency. The court should hold that the statute creates no private right of action and enter judgment against the plaintiff, because that conclusion is an interpretation of federal law and interpreting federal law is an exercise of jurisdiction rather than a confession that none existed. The point generalizes beyond implied rights: when the Court read the federal arson statute to reach only property actively used in commerce and therefore not to cover the arson of an owner-occupied home, it was narrowing the reach of federal law on the merits, not announcing that it had lacked authority to construe the statute.21
Insubstantial and Frivolous Claims
The first branch catches the pretextual pleading. Imagine a homeowner whose real quarrel with his neighbor is a boundary line, who pleads an ordinary state trespass claim and then adds a paragraph asserting that the neighbor's fence violates the Constitution, and who concedes at the hearing that the constitutional allegation has nothing to do with the relief he wants. That allegation is immaterial in the precise sense the exception uses: it is not the source of any right the plaintiff is actually pressing, and it was inserted to get through the door. A court faced with that concession dismisses for want of jurisdiction, because the federal element is not part of the claim at all.
The second branch is harder to satisfy, and the mistake students make is to read "insubstantial" as a synonym for "likely to lose." It is not. A claim is insubstantial in the jurisdictional sense only when it is so attenuated and unsubstantial as to be plainly devoid of merit, and the standard demands that the absence of merit be obvious rather than arguable.24 Novelty does not qualify, because every contested federal question was once novel. Uncertainty does not qualify, because a question on which reasonable judges divide is by definition one the court must answer. Even an eventual ruling against the plaintiff does not qualify, since that ruling is itself an exercise of the jurisdiction the court took in order to make it.
The reapportionment litigation shows how much room the standard leaves. Tennessee voters alleged that the 1901 apportionment statute, left in place through sixty years of population growth and redistribution, debased their votes and denied them equal protection of the laws under the Fourteenth Amendment.25 The district court dismissed on grounds of lack of subject-matter jurisdiction and lack of a justiciable cause of action, without attempting to distinguish between them, reasoning from a line of decisions in which federal courts had declined to intervene in apportionment disputes.26 In Baker v. Carr the Supreme Court separated the two grounds and held that the matter set forth in the complaint did arise under the Constitution and fell within the jurisdictional statute, because the complaint alleged that a state statute effected an apportionment depriving the plaintiffs of equal protection.27 The claim was novel, hotly contested, and backed by a long record of judicial reluctance, and it was still substantial. Whatever obstacle those earlier decisions presented belonged to justiciability, which asks whether the asserted duty can be judicially identified and the right judicially protected, and not to the court's power over the subject matter.
The same ordering appears in Powell v. McCormack, where the respondents did not challenge the substantiality of the congressman's constitutional claims at all, and the Court accordingly treated any bar to review as flowing from the allocation of powers between the branches rather than from a defect in jurisdiction.28 Substantiality is a threshold about whether a federal question is really being asked; justiciability is a separate question about whether this court may answer it.
Setting the filter at frivolousness rather than at probable failure is a deliberate choice, and two mechanisms explain it. The first is the circularity that any higher threshold produces: to decide whether a claim is strong enough to support jurisdiction, the judge would have to evaluate the federal claim, which is the very thing jurisdiction authorizes. Only an assessment that requires no evaluation, one that is very plain on the face of the pleading, escapes that trap. The second is an incentive problem, and it is the one that does the heavier work. Because lack of subject-matter jurisdiction can be raised at any time and is never waived, a defendant who could defeat jurisdiction merely by denying that the plaintiff is entitled to win on the federal claim would hold a non-waivable objection to every contestable federal case, and the federal courts would seldom have any arising-under jurisdiction left.29 Keeping the exceptions tiny also protects the defendant, since a jurisdictional dismissal resolves nothing and leaves the plaintiff free to sue elsewhere, while a judgment that the complaint fails to state a proper cause of action is a judgment on the merits that ends the dispute.30 Courts have never been fully comfortable calling these dismissals jurisdictional at all, precisely because labeling a claim frivolous requires at least a glance at federal law; the label survives because the category is kept so small that the glance is never a real adjudication.
The two poles differ by a single fact, which is whether the defect lies in the existence of a federal theory or in its strength. A plaintiff who concedes that his federal allegation is immaterial and was pleaded only to open the federal courthouse for his state-law claim has no federal case, and dismissal for want of jurisdiction is correct; so has a plaintiff whose asserted federal right no authority has ever recognized and whose own text contradicts it. Hold the pleading constant and change only the court's forecast, and the answer flips: a judge who believes a constitutional claim will probably fail after discovery, or who intends to reject it outright on the merits, must keep the case and decide it, because an unpromising federal claim is still a federal claim.
Dismissal on the Merits Compared
The two exceptions just described are narrow because the ordinary way to defeat a bad federal claim is the motion that tests the claim itself, not the motion that tests the court's power. Both defenses live in the same rule and can be raised in the same document, but they ask different questions. A motion asserting lack of subject-matter jurisdiction asks whether this court may decide the dispute at all, and it is directed at the court's authority rather than at anything the plaintiff did wrong in stating the claim. A motion asserting failure to state a claim upon which relief can be granted asks whether the facts alleged, assumed to be true, entitle the plaintiff to anything as a matter of law.32 Treatise writers group pleading challenges accordingly: one family of objections attacks the court's power to entertain the action, and a separate family attacks the complaint itself, pointing to defects on the face of the pleading even though the plaintiff may in fact possess a good claim and simply have failed to plead one.33
What the claim-testing motion does, and does not, examine is worth stating precisely, because students routinely imagine it as a preview of the trial. The court takes every factual allegation in the complaint as true, however doubtful those allegations may look, and asks only whether they add up to a legal entitlement.34 The pleader therefore cannot survive on labels and conclusions or on a formulaic recitation of the elements; the factual allegations must raise a right to relief above the speculative level, which means they must make the asserted wrong plausible rather than merely imaginable.35 In the antitrust setting that produced that formulation, allegations that the regional telephone companies had behaved in parallel did not suffice, because parallel behavior is as consistent with independent competitive judgment as with an agreement, and only an agreement violates the statute.36 Notice that none of this inquiry has anything to do with whether a federal court may hear the case. The court reads the complaint as a lawyer reads a statute, and the answer it reaches is an answer about federal law.
Because that answer is about federal law, granting the motion produces a judgment the defendant can carry into any later litigation. A dismissal for failure to state a claim is treated as a judgment on the merits for preclusion purposes, so the plaintiff cannot take the same claim to another court and start over.37 That label is not automatic, and two qualifications keep it honest. A court may designate the dismissal "without prejudice," which ordinarily strips it of preclusive effect and leaves the plaintiff free to replead, and in practice dismissals for pleading defects often come with leave to amend precisely because the defect is curable.38 The deeper qualification is that the phrase "on the merits" has drifted from its original sense of a judgment that passes directly on the substance of the claim, so what controls preclusion is the ground the court actually rested on rather than the words in the order.39 Where the ruling rests on the substance of the claim, a later court will ordinarily treat it as conclusive; where it rests on a procedural ground, it is unlikely to bar anything beyond a return to the same federal court.40
The jurisdictional dismissal sits on the other side of that line, and the rules say so expressly. The default provision governing involuntary dismissals makes a dismissal an adjudication on the merits unless the court specifies otherwise, but it carves out dismissals for lack of jurisdiction, for improper venue, and for failure to join a required party.41 The carve-out follows from what the court actually decided. A judge who dismisses because the complaint pleads only a state-law cause of action has formed no view about whether the plaintiff was injured, whether the defendant breached anything, or whether the plaintiff should win; the ruling locates the wrong courthouse and nothing more. That is why a plaintiff whose federal suit is dismissed for want of subject-matter jurisdiction after the limitations period has run can generally refile in state court under a state saving statute, which gives a grace period for exactly that situation.42
Keeping the two motions distinct is not fussiness about labels, and two mechanisms explain why the separation is load-bearing. The first concerns finality: a merits dismissal ends the controversy and can be pleaded in bar of a later suit, while a jurisdictional dismissal resolves nothing and sends the plaintiff down the hall to the state system, so collapsing the two would either give conclusive effect to rulings no court ever made or deny it to rulings that were fully litigated. The second concerns timing, and it is the subtler of the two. The jurisdictional objection is never waived and may be raised at any point, even for the first time on appeal, whereas the objection that the complaint fails to state a claim is subject to the ordinary limits on successive motions.43 A defendant permitted to dress a merits argument as a jurisdictional one would thereby convert a limited objection into a permanent one, and could spring it after years of discovery and a trial. That is also why the direction runs only one way for the judge: a court may not dismiss for want of jurisdiction on the ground that the complaint fails to state a claim on which relief can be granted, because saying the plaintiff should lose is an exercise of the power the court was supposed to be disclaiming.
The two poles differ in a single respect, which is what the defendant is actually contesting. Suppose a plaintiff sues under a federal consumer-protection statute and the defendant responds that even on the plaintiff's own account of the facts the defendant's disclosures satisfied the statute's requirements. That is an argument about what the statute demands, so it belongs in the failure-to-state-a-claim motion, and a judgment granting it decides the dispute and ordinarily bars a second suit on the same claim.44 Now hold everything constant except the ground of the attack. If the defendant instead argues that the complaint, read fairly, pleads only a state-law misrepresentation claim with no federal element on its face, the objection is jurisdictional; the dismissal says nothing about whether the disclosures were adequate, and the plaintiff may bring the identical claim in state court the next morning.45