Erie and the source of multi-part tests

Here is an under-discussed Erie problem: Whether, and how to analyze whether, a federal court must apply an atextual judicially created multi-factor balancing test over contrary state law. I thought of it in looking at this Third Circuit case that Eugene flagged. I wondered what courts had done in diversity cases and found this 2021 New Mexico case in which plaintiffs urged the court to apply the state near-absolute guarantee of anonymity in sexual abuse cases rather than the federal multi-factor balancing. The court's analysis, in language that would fail my Civ Pro exam, came in a footnote:

The Court rejects Plaintiffs’ request for “deference to the State of New Mexico's judicial system's ‘blanket finding’ that childhood sexual abuse survivors be permitted to remain anonymous until their trials.” Doc. 33 at 4. When exercising its diversity jurisdiction, this Court must apply New Mexico substantive law. See Hayes Family Tr. v. State Farm Fire & Cas. Co., 845 F.3d 997, 1005 (10th Cir. 2017) (citing Hanna v. Plumer, 380 U.S. 460, 465 (1965)). But this mandate does not allow Plaintiffs to argue that this Court should defer to a state court procedural practice. Federal courts are bound to follow federal procedural law. Hanna, 380 U.S. at 465 (1965).

This case and the general problem provided two insights on Erie.

First, when courts make the "federal courts apply federal procedural law" move, they mean federal courts apply the FRCP or other statutorily created procedural law. That is the lesson of Hanna--the Erie doctrine does not act as a "check" on the FRCP. But Hanna requires a different approach (and perhaps conclusion, although not in Hanna itself) when the federal rules--including procedural laws--derives from a source other than the FRCP. Courts have unfortunately (and inaccurately) short-handed the analysis. But I think that is the idea they are getting at.

Second, courts have not identified clear rules for determining when an atextual judicial gloss derives from sparse statutory text (and thus becomes part of the text) and when it reflects judge-made common law. That distinction determines whether the federal standard certainly applies under the REA (because no federal rule is invalid) or whether the court performs a far less certain "relatively unguided Erie" analysis that is more likely to require it to apply state law. This was the point of departure between Ginsburg's majority and Scalia's partial dissent in Gasperini--whether the judicially created "seriously erroneous result"/"miscarriage of justice" standard for new trials derived from an interpretation of FRCP 59(a) (Scalia) or whether the courts created it to fill statutory gaps (Ginsburg). But neither explained why their preferred understanding was correct. But it is not always obvious. Why is Twiqbal plausibility an interpretation of FRCP 8(a)(2) while the four-part Winter test for a preliminary injunction is a stand-alone test independent of FRCP 65? (the latter one is historical and predates the FRCP, but the point stands).

Although not explicit and not framed in these terms, courts seem to understand the multi-factor balance for pseudonymity as an interpretation and thus part of FRCP 10(a), which requires that the case caption include the party names. The Third Circuit explained: 1) "Rule 10(a) requires parties to a lawsuit to identify themselves in their respective pleadings" and 2) "[w]hile not expressly permitted under Federal Rule of Civil Procedure 10(a), in exceptional cases courts have allowed a party to proceed anonymously." It follows (although the court did not say this much) that the balancing of nonexhsaustive competing interests for and against anonymity implements FRCP 10. Again, however, it is not obvious why this is the case.

So how should the District of New Mexico have resolved the plaintiff's argument for applying the more plaintiff-friendly pseudonymity rule, beyond "we must apply federal procedure." It should have said that under the Rules of Decision Act, FRCP 10(a), as interpreted through the multi-factor balancing, constitutes an Act of Congress that "otherwise requires or provides" and serves as the rule of decision; as interpreted, it is sufficiently broad to answer the question of when a plaintiff can proceed pseudonymously. The court then should have analyzed whether the rule was valid under § 2072--whether it really regulates procedure and whether it does not abridge, enlarge, or modify a substantive right. The answer would be "yes, it is valid," because every FRCP is valid. But the substantive policies underlying state law (protecting the privacy of sexual-assault victims and encouraging them to come forward) would have forced the court to either rely on "incidental effects" on substantive rights being permissible or adopt the Scalia view that a procedural rule never, in any meaningful sense, can AEM. Either way, that analysis is a lot more complicated than the footnote allows.

And what if the multi-factor test were not part of FRCP 10? The unguided Erie analysis must consider whether ignoring state law would cause the "character or result" to differ, whether a party might go to federal court to avoid less favorable state law, and whether the state law is "bound up" with substantive policy. That analysis usually points towards state law and probably would here--a defendant might remove to avoid automatic pseudonymity and the state law has underlying substantive concerns.

But this example shows why "apply federal procedure," without more, cannot be the extent of the analysis. A litigant-name rule is unquestionably "procedural"--it relates to the manner and means through which rights are enforced in court. But if the conflict with state law comes from a federal legal rule unmoored from the FRCP, the federal court may be required to apply that state law.