The Erie doctrine holds that a federal court sitting in diversity jurisdiction must apply state substantive law and federal procedural law, so that the outcome of a case does not depend on whether it is heard in state or federal court. It comes from Erie Railroad Co. v. Tompkins (1938), which overruled Swift v. Tyson and declared that there is no federal general common law—federal courts must follow the law of the state, including state common law, on substantive questions.

Here's the thing about Erie: it looks simple stated as a one-liner ("apply state substantive law, federal procedural law") and then falls apart the second your professor asks you to draw the line between substance and procedure. That line is the entire game. Erie is one of the most heavily tested topics in 1L Civil Procedure precisely because the rule is easy and the application is a maze. This page walks you through the constitutional foundation, the line of cases that built the modern test, and the exact analytical framework you deploy on an exam.

What Is the Erie Doctrine?

The Erie doctrine is the rule that a federal court exercising diversity jurisdiction (or supplemental jurisdiction over state-law claims) must apply the substantive law of the state in which it sits, while applying federal law to procedural matters. It flows from Erie Railroad Co. v. Tompkins, 304 U.S. 64 (1938), where the Supreme Court held that "there is no federal general common law" and that Swift v. Tyson's 96-year experiment—letting federal courts craft their own general common law in diversity cases—was unconstitutional.

Why did the Court reach a constitutional conclusion? Because Swift had produced exactly the mischief that jurisdiction is supposed to prevent. Under Swift, the substantive law a litigant faced could differ depending on whether the case was in state or federal court in the very same state. That let out-of-state parties forum-shop for more favorable federal common law, and it created inequitable results between citizens and non-citizens of a state. In Erie itself, Harry Tompkins was walking along a railroad right-of-way in Pennsylvania when he was struck by something protruding from a passing Erie Railroad train and lost an arm. Under Pennsylvania common law, he was likely a trespasser owed only a duty to avoid wanton or willful injury; under the "general" federal common law the lower court applied, he could recover under an ordinary negligence standard. Same facts, different law, purely because of the courthouse. Justice Brandeis, writing for the Court, said that could not stand.

The Two Rules That Make Up Erie

Erie is not a single rule—it's a two-step allocation. Learn both as a numbered structure, because your exam answer lives inside it.

1. State substantive law governs. A federal court in diversity has no power to create its own substantive common law. It must apply the law of the relevant state, including state statutes and state judicial decisions. If the highest state court has not decided the question, the federal court makes an Erie guess about how that court would rule.
2. Federal procedural law governs. Federal courts apply their own procedural rules—most importantly the Federal Rules of Civil Procedure—even in diversity cases, because Congress and the Court have constitutional and statutory authority to govern the procedure of the federal courts.

The entire difficulty is that many rules are not cleanly "substantive" or "procedural." A statute of limitations, a burden of proof, a rule requiring an expert affidavit before filing a malpractice suit—each looks procedural on its face but can dramatically change who wins. Sorting these is what the post-Erie cases exist to do.

How Courts Apply Erie: The Cases That Built the Test

Do not treat Erie as one case. Your professor will test the evolution of the doctrine through four or five key decisions. Here is the through-line.

Guaranty Trust Co. v. York (1945): the outcome-determinative test

In Guaranty Trust Co. v. York, 326 U.S. 99 (1945), the question was whether a federal court in diversity had to apply the state statute of limitations. The Court said yes and articulated the outcome-determinative test: if applying a different (federal) rule would "significantly affect the result of a litigation," the state rule must be applied, because a difference in outcome based on the choice of forum is exactly what Erie forbids. A statute of limitations is outcome-determinative—if the state period has run, the plaintiff loses in state court, and Erie says she should lose in federal court too.

Byrd v. Blue Ridge Rural Electric Cooperative (1958): the balancing gloss

Byrd v. Blue Ridge Rural Electric Cooperative, 356 U.S. 525 (1958), complicated the raw outcome-determinative test. There the state rule assigned a certain factual question to the judge; the federal practice gave it to a jury. The Court held that the outcome-determinative test is not absolute—courts must also weigh countervailing federal interests. Here the federal interest in the jury as a fact-finder (reflecting the influence of the Seventh Amendment) was strong enough to justify applying the federal jury practice. Byrd introduced interest-balancing: state interest in the rule vs. federal interest in its own procedure, with an eye on whether the difference is truly likely to change the result.

Hanna v. Plumer (1965): the modern two-track framework

Hanna v. Plumer, 380 U.S. 460 (1965), is the case that reorganized everything and gave you the framework you actually use on an exam. In Hanna, the plaintiff served process by leaving copies at the defendant's home, which satisfied Federal Rule of Civil Procedure 4 but violated the stricter Massachusetts requirement of in-hand service on an executor. The Court held that Rule 4 controlled. Its reasoning split the world into two tracks:

• When a Federal Rule of Civil Procedure is directly on point, the court asks only whether the Rule is valid under the Rules Enabling Act (28 U.S.C. § 2072)—i.e., whether it is genuinely procedural and does not "abridge, enlarge, or modify any substantive right." A validly enacted Federal Rule that answers the question controls, even if it is outcome-determinative in the raw sense. This is the Hanna prong / Rules Enabling Act track.
• When there is no Federal Rule on point, the court applies the refined Erie analysis: the outcome-determinative test read in light of the twin aims of Erie—(1) discouraging forum-shopping and (2) avoiding the inequitable administration of the laws. This is the Erie/Hanna prong / relatively unguided track.

The genius (and the trap) of Hanna is the reframed outcome-determinative test. The question is no longer just "would the result change?"—almost any rule can change a result. The question is whether the difference between the state and federal rule is the kind of difference that, viewed at the outset of litigation, would cause a rational litigant to choose federal court to escape the state rule, or that would inequitably favor federal litigants. Service-of-process mechanics don't do that; a statute of limitations does.

Shady Grove v. Allstate (2010): the Rule-on-point question sharpened

In Shady Grove Orthopedic Associates v. Allstate Insurance Co., 559 U.S. 393 (2010), the Court split badly over whether Federal Rule of Civil Procedure 23 (class actions) controlled over a New York statute barring class actions for certain penalty claims. The plurality held Rule 23 was on point and valid, so it controlled. The importance of Shady Grove for a 1L is that it shows the first, hardest step of the Hanna analysis is often whether the Federal Rule actually answers the same question the state law answers—if it does, you're on the Rules Enabling Act track; if it merely brushes up against the state law, you may still be on the twin-aims track.

The Exam Framework: Your Erie Attack Plan

Here is the sequence to write out on any Erie issue-spotter. Do it in this exact order—professors reward the structure.

1. Confirm you're in an Erie situation. Federal court, jurisdiction resting (at least in part) on diversity or supplemental jurisdiction over state claims, and a conflict between a state rule and a federal practice. If federal-question jurisdiction supplies the substantive law, Erie's substance rule usually isn't the issue.
2. Ask: is there a valid Federal Rule of Civil Procedure (or federal statute) directly on point? If a Federal Rule answers the same question as the state law (the Shady Grove question), go to step 3. If not, skip to step 4.
3. Apply the Rules Enabling Act / Hanna prong. The Federal Rule controls if it is valid—arguably procedural and not abridging a substantive right. Federal Rules are presumptively valid; the Court has never struck one down under the REA. Conclude the federal rule applies.
4. No Rule on point? Apply the refined Erie analysis. Use the modified outcome-determinative test through the twin aims of Erie: would applying the federal practice (a) encourage forum-shopping into federal court, or (b) lead to inequitable administration of the laws? If yes, apply state law. If the difference is trivial (like service mechanics), federal practice can apply. Fold in Byrd balancing where a strong federal interest (e.g., the jury right) is present.
5. State your conclusion and name the winner. State law or federal law—and say which specific rule governs the issue.

Worked Examples

Example 1 — Statute of limitations. Plaintiff sues in federal court in diversity two years after injury. The state limitations period is one year; there is no federal statute of limitations for this state-law claim. No Federal Rule is on point. Apply the twin aims: a shorter state period is outcome-determinative and litigants would absolutely forum-shop to escape it, and letting the federal plaintiff win where a state plaintiff would lose is inequitable. State statute of limitations applies; suit is time-barred. (This is Guaranty Trust v. York.)

Example 2 — Method of service. Plaintiff serves the defendant in a manner allowed by Federal Rule 4 but not by state law. A valid Federal Rule is directly on point. Apply the Rules Enabling Act track: Rule 4 is procedural and valid, so it controls even though it changes whether service was proper. Federal Rule 4 applies. (This is Hanna v. Plumer.)

Example 3 — Judge vs. jury on a fact question. State law assigns a preliminary factual issue to the judge; federal practice would give it to the jury. No controlling Federal Rule squarely dictates the allocation. Balance under Byrd: the federal interest in the jury as fact-finder, reflecting the Seventh Amendment's influence, is strong, and the difference is not clearly outcome-determinative. Federal jury practice applies. (This is Byrd.)

The Exam Trap: The Two Biggest Erie Mistakes

Trap #1 — Skipping the "Rule on point" step. Weak answers jump straight to "is it outcome-determinative?" for everything. That's wrong. If a valid Federal Rule of Civil Procedure directly answers the question, you never reach the outcome-determinative/twin-aims test—you're on the Rules Enabling Act track. Failing to bifurcate is the single most common Erie error. Always ask first whether a Federal Rule is on point.

Trap #2 — Using the naive outcome-determinative test. Post-Hanna, "would the result change?" is not enough, because almost anything can change a result. Frame the question the way Hanna did: would this difference, viewed at the outset, drive forum-shopping or produce inequitable administration of the laws? If you write "the state rule is outcome-determinative because the plaintiff loses under it, so state law applies" without the twin-aims gloss, you're writing 1945 law, not 1965 law.

Related Doctrines and How They Interact

Klaxon Co. v. Stentor (1941) extends Erie to choice-of-law: a federal court in diversity must apply the choice-of-law rules of the state in which it sits, not craft its own. This prevents forum-shopping across the choice-of-law layer. Supplemental jurisdiction (28 U.S.C. § 1367) matters because Erie's command to apply state substantive law reaches state-law claims heard in federal court even outside pure diversity. Finally, distinguish Erie from the federal common law that legitimately survives it—in genuinely federal enclaves like interstate disputes, admiralty, and matters implicating uniquely federal interests, federal courts still make common law. Erie killed "general" federal common law, not all federal common law.

Why Erie Matters

Erie sits at the intersection of federalism, the separation of powers, and the practical mechanics of litigation. It's tested constantly because it forces you to reason at two levels at once: the doctrinal test (Rule on point? twin aims?) and the policy engine underneath (forum-shopping and equal treatment). Nail the framework in the order above, and you can handle any Erie fact pattern your professor throws at you.