Embedded Federal Issues

Embedded Federal Issues

The creation test sorts most arising-under cases, because § 1331 is invoked by and large by plaintiffs who plead a cause of action that federal law creates.¹ It leaves out a small class. In those cases state law supplies the cause of action, yet the claim turns on a substantial question of federal law, and the Supreme Court has held that a federal court may hear the case as one “arising under the Constitution, laws, or treaties of the United States.”² The reasoning is commonsense: a federal court ought to be able to hear a state-law claim that turns on a substantial federal question, because a federal forum offers experience with federal issues, solicitude for them, and a hope of uniform answers.³

Smith v. Kansas City Title & Trust Co. is the model. A shareholder sued to stop his corporation from buying bonds issued under the Federal Farm Loan Act. He argued that the purchase breached a fiduciary duty under state law, because the corporation could buy only bonds authorized by a valid law and the Act was unconstitutional.⁴ Missouri law supplied the cause of action, yet the Court found federal-question jurisdiction because the principal issue was the federal constitutionality of the bond issue.⁵ It reached that result over Justice Holmes’s vigorous dissent, on the ground that it appeared from the plaintiff’s statement that his right to relief depended on the construction or application of the Constitution.⁶ The federal issue also appeared on the face of a well-pleaded complaint and did not arise as a defense, which is what separates Smith from Mottley.⁷

Smith’s language was generous, and later cases trimmed it. The Court shied away from the view that a mere need to apply federal law in a state-law claim opens the arising-under door.⁸ Moore v. Chesapeake & Ohio Railway Co. marks one boundary. The plaintiff there alleged that the railroad had violated a federal statute governing the safety of railroad coupling devices, a violation that could be treated as negligence per se under state tort law and so as a federal ingredient of the state claim. Despite the lesson of Smith, the Court refused to treat the claim as arising under federal law, because virtually all of the determinative considerations were matters of state law.⁹

Merrell Dow Pharmaceuticals Inc. v. Thompson enforced the same boundary for negligence claims that borrow a federal standard. Grable later explained that jurisdiction in Merrell Dow would have federalized a tremendous number of state tort claims, because state negligence per se complaints incorporate federal standards with monotonous regularity.¹⁰ The recurring refrain in these cases is that federal jurisdiction demands not only a contested federal issue but a substantial one.¹¹

Grable & Sons Metal Products, Inc. v. Darue Engineering & Manufacturing gathered these strands into one question: does a state-law claim necessarily raise a stated federal issue, actually disputed and substantial, which a federal forum may entertain without disturbing any congressionally approved balance of federal and state judicial responsibilities?¹² Eight years later, Gunn v. Minton restated the question as a four-part test. Federal jurisdiction over a state-law claim lies if a federal issue is:¹³

1. necessarily raised,
2. actually disputed,
3. substantial, and
4. capable of resolution in federal court without disrupting the federal-state balance approved by Congress.

The test is narrow by design. The Smith approach does not operate as an exception to the well-pleaded complaint rule: the court still looks for the federal issue in the complaint, and Smith and Grable change only what it looks for, an embedded federal issue in a state-law cause of action in place of a federal cause of action.¹⁴ Grable also cabins Smith severely, so that under the four-prong test Smith is a limited exception to the primary rule of the creation test.¹⁵ The inquiry is contextual and not mechanical, and the Court has described the doctrine as a “special and small” exception, so successful invocations are statistically rare.¹⁶ The subsections below take the requirements in order.

Necessarily Raised

The first requirement asks whose proof the federal question belongs to. A federal issue is necessarily raised when the plaintiff cannot win the state-law claim without establishing a proposition of federal law, so that the issue forms a link in the plaintiff’s own chain of proof. The Court has put the point this way: Congress has given the lower federal courts jurisdiction only over cases in which a well-pleaded complaint establishes either that federal law creates the cause of action or that the plaintiff’s right to relief “necessarily depends on resolution of a substantial question of federal law.”¹⁷

Smith shows the chain. To win his injunction, the shareholder had to establish that investing in the bonds was beyond the corporation’s authority. To do that, he had to show the bonds were invalid, and to show that, he had to show that the federal act authorizing them was unconstitutional. Because he could win only by establishing that proposition of federal law, the Court held that the federal court had jurisdiction.¹⁸

Grable & Sons Metal Products, Inc. v. Darue Engineering & Manufacturing applied the same logic to the application of a federal statute instead of its validity. The IRS seized real property belonging to Grable & Sons to satisfy a federal tax deficiency and sold it to Darue Engineering. Five years later, Grable & Sons sued Darue in state court to quiet title, a state-law cause of action, asserting that Darue’s title was invalid because the IRS had given notice of the seizure in violation of provisions of the Internal Revenue Code.¹⁹ To decide whether Darue had acquired good title, a court would have to decide whether proper notice had been given under the federal tax code, so Grable & Sons had to prove a proposition of federal law just as the shareholder did in Smith.²⁰ Darue removed, and removal was proper only if Grable & Sons could have brought the action in federal court originally under § 1331.²¹

Gunn applied the requirement to a case within a case. Minton, convinced that his attorneys’ failure to raise an experimental-use argument earlier had cost him his patent suit and led to invalidation of his patent, sued them for malpractice in Texas state court.²² Texas law supplied the cause of action, yet Minton could recover only by showing that he would have won the earlier case, which meant showing that his patent was valid as a matter of federal patent law.²³ A malpractice plaintiff must prove a legally and factually valid underlying claim before collecting damages, so the patent question was necessarily presented on the face of the well-pleaded complaint.²⁴ The lawyers’ choice of defense does not change that result. They could argue either that their conduct was professionally reasonable or that the lost claim was worthless, and the client’s burden to prove a valid underlying claim is the same in both cases.²⁵

Necessity is only the first gate. Gunn reaffirmed the Grable framework while seemingly narrowing access to federal court, and Merrell Dow, in which the plaintiffs could prove negligence by showing a violation of a federal labeling standard, was later explained as a case where the federal issue was not substantial enough to justify a federal forum.²⁶

Requiring the federal issue to be part of the plaintiff’s own proof keeps the doctrine inside the well-pleaded complaint rule. Smith and Grable do not relax that rule; they change what the court looks for in the complaint, which is an embedded federal issue in a state-law claim in place of a federal cause of action.²⁷ The limit also guards against jurisdictional expansion. Federal defenses are easy to assert in the modern world, and a rule that let either side claim a federal forum by anticipating or asserting them could produce what one jurist called a relatively modest federal tail wagging an otherwise largely state-law dog.²⁸

The Mottleys show where the first requirement bites. They sued for breach of a contract to renew railroad passes, a state-law claim that required them to prove no proposition of federal law. They could establish their right to relief simply by proving that the railroad had agreed to renew the passes and then refused.²⁹ The railroad had given a federal statute as its reason for refusing, and everyone expected it to rely on the statute in defense. But the statute would arise as a defense and not as an element of the Mottleys’ case in chief, and it was up to the Mottleys to state their claim and leave it to the railroad to raise whatever defenses it chose.³⁰ Allegations that anticipate a defense and assert that a provision of the Constitution invalidates it show that a constitutional question will likely arise in the litigation. They do not show that the plaintiff’s own cause of action arises under the Constitution.³¹

Suppose the Mottleys tried again under Grable and argued that their case raises two substantial, sharply contested federal questions: whether the federal statute barred the railroad from honoring the passes, and whether the statute was unconstitutional if it did. Both questions are substantial, and the Mottleys can place them in their complaint. They still lose, because an issue that arrives as the railroad’s defense and the Mottleys’ reply to it is not necessarily raised, however important it is. Smith, Grable, and Gunn all ask what the plaintiff must prove.³²

An exam stem that describes a weighty, hotly contested federal question inside a state-law case invites the conclusion that Grable applies. Ask first whose proof the question belongs to, because the weight of a question cannot supply what the plaintiff’s own claim lacks. A defense of equal weight produces the same answer.

The shareholder in Smith could not state his claim without the federal proposition. The bank could state a complete claim without mentioning federal law at all, and that difference in where the question sits decides the result.

The fourth step does the work. The first requirement turns on the structure of the plaintiff’s own claim, which a court can read from the complaint. The contextual judgment that the Smith line demands comes later, with substantiality and the federal-state balance.³⁴

Actually Disputed

The second requirement asks whether the parties genuinely contest the federal question. A federal issue is actually disputed when the parties disagree about what federal law requires or how it applies, so that a court must resolve the disagreement to decide the case. The requirement follows from the reason for allowing a federal forum at all. The advantages the Court has named are the experience, solicitude, and hope of uniformity that a federal court offers on federal issues,³⁶ and each depends on there being an open question for the court to answer. Where both sides accept what a federal statute means, or controlling authority has already said, a federal judge adds little, and what remains is a dispute about state law.

Grable shows the requirement met. Grable & Sons said the IRS had given notice of the seizure in violation of the Internal Revenue Code, the parties contested the point, and the Court found the defective-notice issue actually in dispute.³⁷ Gunn shows the requirement inside a case within a case. Minton’s lawyers defended by arguing that the lease at issue was not, in fact, for an experimental use, so that his patent infringement claims would have failed even had the argument been timely raised.³⁸ Patent validity was therefore genuinely contested in the malpractice suit.³⁹ The two disputes differ in kind, since Grable turned on whether the IRS’s notice satisfied a statute and Gunn on whether a patent would have survived. Each was a live question that a court had to answer before it could rule for either side.

An issue that controlling authority has already resolved may no longer qualify.

The two taxpayers share a claim and an embedded issue, and they differ in one fact: the first presented an open question of statutory meaning, while the second presents a question whose answer is on the books. The requirement asks whether a live question remains, and experience and uniformity have little to add once the answer is settled.

A federal issue that sits inside the plaintiff’s own proof is usually one the defendant has reason to contest, so this requirement seldom decides a case. The harder contests come at the next two requirements, which call for the contextual analysis that the Smith line demands.⁴⁰ An analysis that reaches this requirement should confirm that the parties genuinely disagree about the federal question, note whether controlling authority has already settled it, and then spend its effort on substantiality and the federal-state balance.

Substantial Federal Issue

The third requirement asks whether the federal issue is substantial. Federal jurisdiction over a state-law claim demands not only a contested federal issue but a substantial one, “indicating a serious federal interest in claiming the advantages thought to be inherent in a federal forum.”⁴² Those advantages are the experience and solicitude that a federal court brings to federal questions and its hope of uniform answers.⁴³ The first two requirements looked at the structure of the plaintiff’s proof and the positions the parties take. Substantiality looks at the federal question itself and asks whether it is worth a federal forum.

The interest that counts belongs to the federal system and not to the litigants. Every litigant cares intensely about every issue that decides the case, so a test that counted the parties’ stake would call every embedded federal issue substantial and screen out nothing. The inquiry asks instead whether anyone beyond the parties has a reason to want one authoritative answer. The federal government is the obvious candidate. Whether Congress had power to enact a statute, or whether an agency complied with the statute that governs it, should come out the same way in every state.

Smith is the plainest example of an issue that clears the line. The shareholder’s bill sought to enjoin the corporation from buying bonds issued by the federal land banks under the Federal Farm Loan Act, on the ground that the Act authorizing the bonds was unconstitutional.⁴⁴ The Court upheld jurisdiction because “the controversy concerns the constitutional validity of an act of Congress which is directly drawn in question,” and the decision depended upon the determination of that issue.⁴⁵ The cause of action came from state law, but the question that decided the case was whether Congress had power to create the bonds.

Grable carried the same reasoning from the validity of a statute to an agency’s compliance with one. The Court found the defective-notice issue substantial and found a fairly significant federal interest in issues relating to the IRS’s power to conduct tax sales of real property. It distinguished earlier decisions that had denied jurisdiction where federal law issues were merely in the background of the case.⁴⁶ Smith asked whether a federal act was valid, and Grable asked whether a federal agency had complied with a valid one. Each put the federal government’s own conduct at the center of a dispute between private litigants, and that placement is what made the issue matter beyond the case.

Merrell Dow marks the failing side of the line. A state-law negligence claim there rested in part on the manufacturer’s alleged violation of a federal labeling requirement. The manufacturer was an Ohio corporation with its principal place of business in Ohio, so removal was proper only if the action arose under federal law.⁴⁷ Grable later explained the decision as a judgment that the embedded federal issue was not substantial.⁴⁸ Treating a violation of a federal standard as enough would have federalized a tremendous number of state tort claims, since state negligence per se complaints incorporate federal standards with monotonous regularity.⁴⁹ Commentators add docket impact as a further ground for separating Smith from Moore: personal injury claims like the one in Moore might have brought many more cases to federal court than Smith’s shareholder action did.⁵⁰

Gunn shows the requirement failing even though the first two are met. Minton’s patent had been invalidated in earlier federal litigation, and he sued his lawyers in Texas state court for malpractice, alleging that they lost the patent case by failing to raise a crucial argument.⁵¹ Texas law supplied the cause of action, patent validity sat inside it as a case within a case, and the parties contested it, so the federal issue was necessarily raised and actually disputed.⁵² The Court nonetheless held that the issue was not substantial, because its importance was confined to the parties and did not reach the federal system as a whole.⁵³ A malpractice court’s answer on validity could not revive, invalidate, or change any patent, since the earlier federal case had already ended the patent’s life. The answer would decide only whether private lawyers owed a client damages. Grable differs because its answer concerned the federal government’s power to conduct tax sales, an interest the Court found fairly significant. Minton had argued on appeal that his malpractice claim arose under the patent laws, so that § 1338(a) stripped the Texas courts of jurisdiction, and the holding that the issue was not substantial left the case in state court.⁵⁴

A large stake and a hard-fought federal question are the usual reasons students call an issue substantial, and neither one is the test. The money and the intensity of the dispute belong to the parties, and the requirement asks what the federal system gains from a federal court’s answer. Had the manufacturers’ claim turned instead on whether the agency exceeded its statutory authority when it adopted the regulation, the issue would test federal action itself and would look more like Smith and Grable.

An issue that passes these steps has cleared the third requirement and nothing more. Even a contested and substantial federal question can be vetoed if exercising jurisdiction would upset the division of labor between state and federal courts that Congress has approved.⁵⁵ The next subsection takes up that last requirement.

Federal-State Balance

The fourth requirement works as a possible veto. Even when a state-law action discloses a contested and substantial federal question, federal jurisdiction is proper only if it is consistent with “congressional judgment about the sound division of labor between state and federal courts governing the application of § 1331.”⁵⁷ The first three requirements examine the federal issue itself. The fourth examines what happens if a federal forum opens to every claim like this one.

The Court explained why a separate check is needed. Because arising-under jurisdiction over a state-law claim always raises the possibility of upsetting the state-federal line drawn, or at least assumed, by Congress, the presence of a disputed federal issue and the ostensible importance of a federal forum are never necessarily dispositive. A court must always assess any disruptive portent in exercising jurisdiction.⁵⁸ The assessment concerns a category of cases. The court asks what becomes of the state and federal dockets if every case of this type comes through the same door.

The two leading cases show the question at work. In Merrell Dow the state-law theory was negligence per se, under which violating a safety standard counts as negligence itself, and the standard was a federal one.⁵⁹ Jurisdiction over that complaint would have federalized a tremendous number of state tort claims, because negligence per se complaints incorporate federal standards with monotonous regularity.⁶⁰ That is the disruption the fourth requirement exists to prevent, since a rule that reached the Merrell Dow plaintiffs would reach every tort suit that borrows a federal standard.

Grable sits at the other end. Grable & Sons attacked the title of a buyer at an IRS tax sale on the ground that the IRS’s notice violated the Internal Revenue Code. A quiet title action arises under state property law,⁶¹ and the Court found that a federal forum could hear this one without disturbing any congressionally approved balance, noting a fairly significant federal interest in issues relating to the IRS’s power to conduct tax sales of real property.⁶² The two cases do not differ in whether a federal rule lay inside the state claim, because both had one. They differ in how many cases would follow the first case into federal court.

Congress’s silence plays a part, and the Court’s treatment of it shifted between the two cases. Federal law can supply a cause of action expressly, as Title VII does, or by implication, and a claim on either footing arises under federal law directly.⁶³ In Merrell Dow, by contrast, both parties agreed that the statute created no right of action, express or implied. The Court treated that absence as an argument against jurisdiction: if Congress declined to create a federal action, it must also have preferred to avoid federal adjudication of state claims enforcing the federal standard.⁶⁴ Grable corrected what it saw as Merrell Dow’s over-emphasis on that absence. Congress’s failure to create a federal action may have some continuing relevance in assessing congressional intent, but it does not decide the question.⁶⁵ A rule that made the omission fatal would also prove too much. In virtually every case in which a state claim incorporates a federal standard, including Smith, Congress will have created no explicit federal right of action, and strict reliance on the omission could foreclose embedded-issue jurisdiction altogether.⁶⁶

Congress’s affirmative acts carry more information than its silence. In City of Chicago v. International College of Surgeons, the Court upheld federal jurisdiction over a state-law claim that embedded allegations of federal due process and equal protection violations. Those allegations were proper subjects of an action under 42 U.S.C. § 1983, so Congress had already authorized the federal courts to hear such constitutional claims, and jurisdiction over similar claims embedded in a state right of action did not upset the balance.⁶⁷ A federal forum for the embedded version added nothing that Congress had not already placed in the district courts. Silence can point either way depending on context, whereas an existing federal authorization for the same kind of claim points toward jurisdiction.

The first two requirements ask about the plaintiff’s proof and the parties’ positions, and a court can read much of that from the pleadings. The last two require judgment about the federal system, which is why the Smith test rides on a more contextual analysis than the creation test does.⁶⁸ The four requirements are cumulative, and the doctrine stays a special and small exception.⁶⁹

The whole inquiry also stays tied to the well-pleaded complaint rule. Smith and Grable go beyond the creation test, but the court still looks only at what the plaintiff must establish as part of the plaintiff’s own case.⁷⁰ A problem on this doctrine therefore starts with the complaint, finds the federal proposition inside the plaintiff’s own proof, and only then asks whether a federal forum for that kind of claim fits the balance Congress approved.