The Erie Doctrine and Choice of Law
I. Pre-Erie Choice of law:
A. Only a problem with diversity jurisdiction. With Erie, the 'horizontal' uniformity of laws (between states) created by "federal common law" gave way to 'vertical' uniformity (between state and federal courts).
B. Swift v. Tyson (1842) -> Tyson wants I.O.U. voided for fraud, and under N.Y. law, he wins. But court applied the "federal common law" -> There is a uniform set of commercial rules in effect in every state. The federal courts were sympathetic to an expanding economy, and thus were a "uniform national law merchant".
1. Court first looked to Article III which sets limits on the jurisdiction of the federal court system, but it did not establish any lower federal courts.
2. So they looked to the Rules of Decision Act, 28 U.S.C. § 1652, which reads: "The laws of the several states ... shall be regarded as rules of decision in civil actions in the courts of the United States ..."
3. Swift read the "laws of the several states" to mean statutory only, not court decisions.
a. Judges don't create the law -> it was always there, judges are just 'discovering' it through reasoning. The law is some transcendental notion of what's right - there's A right answer.
b. So common law was universal, state and federal judges both use the same thought processes to decide issues.
II. Erie R.R. v. Tompkins -> Federal Courts must apply the law of the state in which it sits, including state court decisions. A federal court must look to highest state court decisions to see what it would do. If they have not ruled on the issue, you can ask for certification by the court, or you might try and predict how they would rule.
A. Brandeis overruled Swift on three grounds:
1. Philosophy -> law is not some transcendental notion of justice, it is man- made, and therefore it differs from place to place. So law is a question of POWER; there is no one right answer.
a. But then how can we say there are any common notions of justice, like international human rights?
b. So Erie is a modern statement about what judges do. It took away their autonomy to create federal common law, except in constitutional cases. This forced judges to "constitutionalize" many issues to allow creativity over them.
2. Equality/Discrimination -> Erie attempted to equalize things, so that in-state and out-of-state parties had the same choice of law. a. Black & White Taxicab -> B&W (KY) seeks to sue in federal court to get federal common law, so it re-incorporates across the river in TN. P gets choice of two laws; home-state D can't remove. Do you need to overrule Swift to prevent this?
i. use § 1359 (parties improperly or collusively evoking diversity).
ii. principal place of business is in KY.
iii. change removal so an in-state defendant can remove.
b. Plaintiff often has choice of law, even without federal courts, so where's the discrimination -> in the removal rules.
c. So what Brandeis had in mind was not discrimination between actual parties (plaintiff over defendant), but rather groups of litigants: P suing D from different state has choice of law; P suing same-state D does not.
d. Brandeis said Erie puts an end to forum shopping by making law uniform. It can't be that broad; but more importantly, is forum shopping bad? What's wrong with seeking to get your client the advantage? So some forum shopping is inevitable.
3. Federalism -> states should be able to decide their own contract, tort, etc. laws. For federal courts to determine state law violates federalism.
a. But if the federal legislatures can trump a state statute or judge under the commerce clause, why can't a federal judge trump a state judge?
b. This is due to the fact that legislature acts with a popular mandate. To allow judges to trump takes it out of localities' hands.
4. Practical -> State standards may be less strict than federal, so you could get cacophonous judgments. Law affects behavior, and so Erie prevented parties from being told two conflicting norms to follow. The law could differ depending on who got hurt - in-state or out-of-state resident.
III. Substance or Procedure? Erie says state substantive law has to be followed, procedural rules do not. What is the difference, and how can you tell one from the other?
A. Klaxon Co. v. Stentor Electric -> Conflicts of Laws Rules are Substantive. To keep outcomes constant, federal courts should apply the conflicts laws of the state in which it sits. Federal courts cannot make up substantive state law, and can't decide which state's laws apply, either.
1. But if a case involving P's from several states is consolidated into one federal case, you could have two P's sitting next to each other, and one wins, one loses because they're governed by different laws.
2. So there is pressure in this case to allow judges to break Klaxon and apply the law that should govern the case.
3. Why are conflicts rules substantive?
a. They are outcome determinative, if they weren't no one would argue them.
b. They could create a forum shopping "frenzy". But does just one more court (fed.) really lead to more forum shopping?
c. Federalism concerns -> you don't want to trump state's sovereignty
B. Guaranty Trust v. York -> Statutes of Limitations are Substantive. Federal equity courts are allowed to make statute of limitation decisions on a case-by-case basis to "do justice". But statutes of limitations are about staleness of claims, as well as a state's desire for repose (finality). A state makes it statute of limitations for the efficiency of their own courts, do we want the federal courts to just ignore that decision (federalism concern?)
1. Frankfurter's Outcome Determinative Test: Erie is about getting the same outcome, so any state rule that effects the outcome of a case must be applied.
a. But just about every procedural rule can be said to effect the outcome of a case, or else we wouldn't fight about it.
b. Hypo: State court closes 2 hours before federal. Even this procedural rule could effect an outcome if you want to file at 4pm on the last day.
i. This is cautionary example against applying the outcome- determinative test too blindly.
ii. By hypothesis, every issue that comes up after Erie is outcome determinative, so every state procedure would apply.
c. Instead, maybe Frankfurter meant it as a wholesale test: Not is it outcome determinative in this case, but whether it would effect the outcome in a large number of cases.
2. Also, Injunctions are Procedural. Whether an injunction can be ordered (or other remedy available) is not outcome determinative. Frankfurter says outcome determinative is a bottom-line decision.
3. How can statute of limitations be substantive, while injunctions are procedural? Better way to look at it -> how does it affect behavior? (Harlan)
a. Post-event behavior: A defendant can never relax, because you never know when a judge can re-open your case (no sense of finality). Laches -> plaintiff can't sleep on his claim.
b. Pre-event behavior: With shorter statute of limitations, people will be riskier because it is tougher to get you claims in on time. States have carefully calculated the level of risk they want (vs. the need for repose), so federal judges should not alter it, because it would have an effect on the behavior within the state.
4. Also concerned about forum shopping (with statute of limitations). What about remedies, don't they encourage forum shopping?
C. Ragan v. Merchants Transfer & Warehouse Co. -> Followed Guaranty. Service is not given until 1 day after statute of limitations runs out. State rule says statute runs until service is made (illustrates state's interest in repose); Rule 3 says filing complaint stops statute of limitations from running.
1. What do you do if late service was the marshal's fault?
2. Shouldn't we apply the federal rule in federal cases, because it could take longer to serve someone outside the state. Otherwise, could be unfair to person suing out-of-stater because he has to make sure he files early enough to get service made within the statute.
3. Does it affect pre-event behavior? Are people more likely to do anything differently because of the extra week? Might depend on the length of the statute (i.e., 3 mos. vs. 10 years).
4. Does it affect post-event behavior? Repose is important, for example, for an executor trying to disburse money (so it's not tied up). But a defendant can check the records to see if an action has been filed.
D. Byrd v. Blue Ridge -> Jury Trial is Procedural. State says judge makes decision whether person is employee or independent contractor for workmen's comp. Federal rule says you're entitled to a jury trial on all disputed facts (7th Amend.).
1. Balancing Test: Brennan uses a balancing test to weigh the state's interest in applying their rule against the federal interest. The federal interest in jury trials trumps the state's interest in the judge decision, so federal courts can use jury decision. (federalism concerns?)
2. Appears to be an objective test, but actually just a subjective determination -> judge makes up what the interests are, then weighs them.
3. Does it affect behavior? Jury trials liberalize the tort system, so it affects both pre-event, and especially post-event behavior.
E. Hanna v. Plummer -> Service of Process is Procedural. Federal rule (Rule 4) says you can use any state method, or you can leave it at a principal place or abode. State rule calls for "in hand" service.
1. It is not outcome determinative. Also, its a technical state rule that does not affect pre-event or post-event behavior. (Harlan) Therefore, Rule 4 should be applied.
2. Balancing test: State rule makes a statement about frozen money in estates, because of fear of lawsuits. So the federal rule infringes on state's interest.
3. Presumptively Procedural: Federal rules are presumptively procedural; so in diversity cases, the federal rule will apply when it conflicts with the state rule; unless its application would undermine the state's policy or its too substantive, such that it violates the constitution (safety valve).
a. Isn't this the distinction between legislative-made law and judge- made law that Erie meant to do away with? (Fed. is written down, so it's "real law").
b. It is assumed that you know when the federal rule is in conflict, but often the rule has to be construed to conflict.
c. By this time, litigants were looking for a rule, any rule, that they can plan by -> predictability.
F. Walker v. Armco Steel -> Ragan revisited; service is given too late. Rule 3 says that "an action is commenced when service is filed". State rule says service is required to stop statute of limitations. Are they in conflict? Rule 3 is read broadly to conflict with state rule.
G. Burlington Northern v. Woods -> State rule has automatic sanction for unsuccessful appeals of money damages (penalty for collecting interest on money you should be paying). Federal rule give a judge discretion to impose such sanctions (Fed. Rule of App. Proc. 38). Is this collision?
a. Does presumptively procedural really save time? As much argument goes into whether rules are in conflict as did the policy rules behind Erie, with about as much predictability.
b. It all depends on whether the judge reads the federal rule broadly (to be in conflict), or narrowly.
c. How do you determine whether there's a collision with a state rule? Is it an unarticulated use of the Harlan balancing test?
H. The Harlan test (pre-event/post-event behavior) is the most intellectually defensible. But it is too expensive and uncertain. Under the Harlan test, Walker comes out differently; is one wrong decision okay for superiority of the test)? Today the test is Hanna -> presumptively procedural. There are 4 tests:
1. Outcome Determination
2. Balancing Test (Byrd)
3. Pre/Post event behavior (Harlan in Hanna)
4. Presumptively Procedural
Hypo: State rule -> evidence is admissible; Federal rule -> evidence is admissible. How does it play out after Erie?
I. Stewart Organization v. RICOH -> Enforcement of forum selection clause. Federal law 28 U.S.C. § 1404(a) allows change of venue, but state law frowns on forum selection clauses. There are two questions considered:
1. Does the statute speak to the issue at hand?
2. Does it demonstrate the power of Congress to govern the issue (the constitutional authority of Congress to enact §1404 is not subject to serious question).
Statute was presumptively procedural -> Congress has the power to govern procedure in its own (federal) courts.
J. McKenna v. Ortho Pharmaceutical -> PA has "borrowing statute": if the cause of action is barred in the state in which it arose, it is barred in PA. Ohio law: Statute of limitations should not be "floating around", there should be a fixed date on which the statute begins to run. So with malpractice, the statute begins to run on last day of treatment, not discovery of symptoms. But the McKenna court predicted that Ohio law was going to change. Did federal judge get state law wrong? Is this just like creating federal common law?
K. Allstate Insurance v. Hague -> When can a state apply its own law? State A, an insurance claim is split by claimants. State B, coverage is for each individual claimant. P moves to A after the accident, to get A's law. Federal judge must decide whose law governs. Due process requires a significant grouping of contacts before state A's law applies. But states have extraordinary latitude in developing choice of law rules.
- See Shatts -> state's power is not unlimited.
The Doctrines of Preclusion
I. Stare Decisis: a stability norm; it is using precedent to create a form of external (outside the judge) stability.
A. Stare decisis is at its strongest in a commercial setting, when getting it certain is more important than getting it right. It's at its weakest in a constitutional setting, when getting it right is more important than getting it certain, so the Court feels freer to overturn itself.
B. If stare decisis is just predicting how a judge will rule, then a private party is able to keep trying to re-litigate and force the courts to change their mind.
1. But stare decisis makes it costly to keep going to court, so it forces decisions. It is usually cheaper to comply with stare decisis than to fight it.
2. Private parties -> can ignore stare decisis even if they keep getting kicked out of court. So it's different from violating a statute, there is no duty to comply with precedent.
3. Government Agencies -> when a precedent emerges, they have a duty to comply with it. They can't re-litigate, so what if the court got it wrong?
II. Res Judicata and Collateral Estoppel in General: Stare decisis is presumptive, res judicata and collateral estoppel are absolute.
A. Res judicata and collateral estoppel:
1. Reduce unnecessary litigation
2. Foster reliance on adjudication
3. promote comity between state and federal courts
B. Res Judicata (claim preclusion) - "the thing has been decided" -> says that your claim has already been decided, so you can't go into court and try to re-open it because you forgot something. It acts as a penalty for not bringing it up the first time, either through sloppiness, or game playing.
C. Collateral Estoppel (issue preclusion) -> a particular issue has already been decided against you, so you cannot try it again. (Res judicata deals with things that were never decided; collateral estoppel deals with things that were actually decided; stare decisis deals with an issue that has been decided, but not in my case).
- Direct Estoppel: the issue has been decided at an earlier trial between the same parties.
D. Preclusion (in general) concerns:
1. Due Process -> requires that nobody can be precluded until you've had your day in court. So there's always a question of whether a party was actually present in the first case (for example, class actions).
2. Raising it early -> Preclusion is a disfavored defense, so it must be raised early in the case.
3. Under Erie, must federal courts use state preclusion rules? 28 U.S.C. § 1738 -> federal judges must look down to forum in Case I to determine preclusion rules.
4. Pendent Jurisdiction gives you the opportunity to bring both claims in the same place.
a. So if you have the opportunity to bring both claims cheaply, should we force you to do so, for efficiency?
b. If deciding whether pendent jurisdiction should be granted, should we look to the likelihood that Case I will be preclusive on Case II?
III. Claim Preclusion: When a second suit is brought, the judgment in a prior suit will be considered conclusive, both on the parties to the judgment and those in privity with them, as to matters that actually were litigated or should have been litigated in the first suit. For claim preclusion to operate, three elements must be present:
(1) Only judgment that are "final", "valid", and "on the merits" have preclusive effect.
(2) The parties in the subsequent action must be identical to those in the first. (Distinguishes it from issue preclusion.)
(3) The claim in the second suit must involve matters properly considered included in the first action. (Turns on what the first action decided or should have decided.)
A. What is a claim?
1. Theory driven -> evolved form English system of writs, which forced you to choose a theory, or writ. If you lost, you could just buy another writ. In the U.S., the definition of claim was theory driven until the 1950's, so a tort theory was one claim, a contract theory was another.
2. Transaction -> Then the definition of claim was based on "same transaction". So they were based on a "cluster of facts" because theories were mixed and joined together when you brought a claim against someone.
a. So when you bring a claim now, you are forced to throw everything in, even if you don't need it, because you can't bring anything again.
b. Is the transaction theory too broad, so that it is inefficient?
3. Liability facts -> less broad
4. If you define a claim broadly, you will force a plaintiff to bring everything, to frontload issues that even he thinks suck. This wastes the court's time, because there are no more small claims. It is also hard on defendants, who have to oppose even stupid claims.
a. So how you define "claim" is actually a policy argument about how much strategic game-playing you want.
b. How broad you define "claim" is how much you want each side to put on the table.
Vasu v. Kohlers -> Property damage and personal injury are different claims (Dictum). P's insurance co. settles with him for property damage, gets subrogation, sues D to get their money back, and loses. P wants to go after D for personal injury. Court allows this "split".
1. P did not lose his right to trial just because his insurance co. lost, because he did not get his day in court -> due process.
2. If the insurance co. had won, P can use issue preclusion against D. In second trial D can't win, but he can lose -> No Mutuality (shows D's incentive for class actions).
Rush v. City of Maple Heights -> P sues the city for damage to her motorcycle; the court finds the city was negligent, awarding her $100. She then takes that judgment into court for issue preclusion against the city for her personal injuries, amounting to $12,000. P's lawyer relies on dicta from Vasu. City argues that she split her claim, so it should be precluded, court agrees.
1. City's lawyer claims that because first claim was for $100 he didn't defend it as well as he would have for $12,000. But most courts would say he should have raised something at the first trial. A "good lawyer" would have settled the first claim, then said see you in court for the $12,000.
2. You can distinguish Vasu because there were different parties to the second action. But don't persons and their insurance co. have interests that are close enough to be considered the "same party"?
3. D can argue that all Vasu meant was that property and personal injury claims could be brought by separate people. In Rush, claims were in the hands of the same person. Both arose form the same facts, so one "claim".
a. In Vasu, if they weren't found to be "separate" and therefore non- preclusive, then if 2 people held the claims, it would become a race to judgment, with the second-place finisher's claim being either barred or merged into the second claim.
b. Merger: If you get a judgment on a piece of your claim, everything else arising from that claim (or set of facts) is merged. It "disappears" into the judgment, and you've got nothing left to sue on. This forces people to bring all their theories at once.
4. But Vasu said that claims = rights, different rights violated bring up different claims. So no issue preclusion because parties were different, and no claim preclusion because there were different rights violated.
5. Claim Splitting -> can also lead to game playing, you can run a small claim by the court to get issue preclusion later. Or if you lose, you get to see defendant's case for your big claim.
6. Retroactive Overruling -> when you change the rules for policy reasons, you may harm the parties in front of you who have relied on prior rule. Changing rules retroactively is damaging; people are reluctant to rely again. [So can you uphold Vasu, and give the parties something to go on?]
7. Prospective Overruling -> is against the policy that courts decide the case in front of them. It also makes it too easy to abandon precedent, because it doesn't hurt anyone.
8. Holding vs. Dicta -> Rush lost because his lawyer relied on dicta. Shows that the price of changing rules is borne by the individual parties.
Cromwell v. County of SAC -> County issue bonds, to each bond are attached coupons (interest payments). P owns 25 coupons, county defaults.
Case I -> P sues County on one coupon, D says it was obtained fraudulently: P didn't prove he paid value for it - judgment for D.
Case II -> Same P then sues for second coupon, and proves he paid value for it.
D argues P split his claim - it should have been raised in Case I, but court says P is not barred. [If it was found that he didn't give value in case I - issue preclusion].
1. Court says each coupon is a different "transaction", each coupon could be held by a different person; so coupon 2 is not barred under claim preclusion.
a. The fact that you didn't give value for one doesn't mean that you didn't for the rest.
b. If you got them from different parties on different days -> different clusters of facts ("transaction").
c. If they weren't separate claims, you would get a Vasu race, where a decision against one party would bar or merge the claims of all others.
2. So Cromwell was decided on potentially different clusters of facts; courts today would probably decide on actual different clusters of facts.
3. Issue Preclusion -> Defendant can try to use issue preclusion, but the question of whether P gave value for the coupon was not Actually Adjudicated.
a. To see what was actually decided, you look to the holding.
b. Issue preclusion is absolute, so even if there's a fundamental change in the law, P can't re-litigate his decision. Stare decisis is not so absolute, so P may be the only person who can lose.