Background on the original Erie case

1) Prior to  Erie, under the regime of Swift v. Tyson, a federal court sitting in diversity was free to engage in its own common-law making. It was obligated to follow state law only when a state statute was at issue or when the common law rule was a "local usage," that is a rule governing something having permanent location in the state. Under Erie, the Swift v. Tyson approach was held to be unconstitutional. Henceforth a federal court sitting in diversity (or taking a cause of action under supplemental jurisdiction) must apply the applicable state law as determined by that state's courts. What to do when the law of that state is uncertain? Generally the federal court will predict what the relevant state supreme court would say on the matter.

2) But Erie does not answer the following issues. Let us say that a federal court in New York, sitting in diversity, is entertaining a New York state cause of action. We know that the federal court must apply New York substantive law as determined by New York state courts. But to what extent should the federal court also apply New York procedure? It would seem that Erie is irrelevant here, since the federal court is not engaging in substantive law making if it used federal procedure. One might think that federal law should apply whenever it is classified as procedural. After all, that is how the matter is decided in a horizontal (state-state) choice of law context. A Connecticut court entertaining a New York cause of action will apply Connecticut procedure. Why shouldn't the same thing be true in a vertical (state-federal) choice of law context? Unfortunately, the matter is not so simple. In cases like York, Byrd, and Hanna, Erie is understood as also creating a policy that changes these choice of law inquiries, requiring some state procedure to be used by federal courts sitting in diversity.

York

1) How would York be decided if it were a horizontal choice of law case, e.g. if a Connecticut state court were choosing whether to use its own law on limitations of action or New York law when entertaining a New York cause of action?

2) Consider the Dunlap case on p. 707. How would that be answered in a horizontal context?

3) Consider the following rule: To get into federal court you must go through a metal detector. Assume that in state court the rule is you don't. Under York, should the state rule apply in federal court in a diversity case?

Byrd

1) It seems pretty clear that in the Byrd case the state rule is outcome determinative. Why then is the federal rule used? Can the result be squared with York?

2) Brennan argues that if a state procedural rule is bound up with state-created rights and obligations a federal court is constitutionally obligated, rather than being obligated under the "broader policy" in Erie, to apply the state procedural rule? What is an example?

Hanna

1. The Hanna court argues for a particular interpretation of the outcome determinative test, such that not every procedural rule will be understood as outcome determinative. Is that interpretation compatible with what was said in York? Look at the procedural rule at issue in York. Was York was wrongly decided according to the Hanna court's test?

2. The Hanna court claims that Erie is not the test for determining whether a Federal Rule of Civil Procedure can be used in a diversity case. Why? Why was Erie relevant in cases like Klaxon, York, Ragan, Woods, Cohen, and Byrd but not in Hanna? What's the difference?

3. What is the test for determining if a Federal Rule of Civil Procedure can be used in a diversity case? Can you think of an example where a Federal Rule would fail this test?

4. What about federal statutes regulating procedure? How does one determine whether they can be used in a diversity case? If Congress passed a statute spelling out choice-of-law rules or statutes of limitations to be used in diversity cases would it be OK for a federal court to use them? Or would that be contrary to the Supreme Court's decisions in Klaxon, Ragan, and York?

5. What is Harlan's solution to the Erie problem? Is it workable?