Trial Magazine

Federal courts are often confronted with complex factual and legal issues that make the adjudication of disputes extraordinarily challenging. This is especially true in multidistrict litigation (MDL), in which one federal judge is assigned to oversee and manage mass litigation affecting hundreds or thousands of claimants from various states. Significant difficulties arise in diversity cases when there is an undecided issue of state law and the MDL judge must predict how the highest court in the state would decide the issue. Such predictions—commonly referred to as “Erie guesses”¹—can meaningfully affect the substantive rights of MDL litigants.

The Erie doctrine stems from the U.S. Supreme Court’s 1938 decision in Erie Railroad Co. v. Tompkins, which was intended to rein in “an unconstitutional assumption of powers” by the federal courts that were exercising their own judgment about states’ common law.² The Court noted that “the federal courts exercising jurisdiction in diversity of citizenship cases [should] apply as their rules of decision the law of the state, unwritten as well as written.”³ Under Erie, when the state law is clear, federal courts have a duty to accept it and the state’s policy choices, even if the court thinks the decision is wrong. And when the state law is unclear and an Erie guess is required, that prediction still must be made in accordance with state law.

As noted by the Erie Court: “There is no federal general common law.”⁴ The twin aims of the Erie doctrine are to discourage forum-shopping and avoid inequitable administration of the law.⁵ Erie’s emphatic pronouncement, it was thought, would achieve these goals. It has not.⁶

All federal courts, including transferee MDL courts, “routinely decide unsettled questions of state law, and often decide them incorrectly.”⁷ Some errors are to be expected. After all, absent clear guidance on the issue from the state’s highest court, a federal court’s prediction is a guess nonetheless. What should not be expected—or accepted—are errors in the constitutional application of the law with regard to principles of federalism and separation of powers.⁸ Federal courts must not intrude on the legitimate lawmaking function of the state court.⁹

Problems with Erie guesses arise when federal courts sitting in diversity fail to undertake any meaningful state law analysis and instead opt to rely solely on federal court decisions for guidance. By doing so, these courts are essentially creating new state law based on a perceived federal common law that the Supreme Court stated in Erie does not exist and should not be used.¹⁰

When Guesses Lead to Dueling Outcomes

Decisions in two recent drug MDLs—In re Zantac (Ranitidine) Products Liability Litigation and In re Fluoroquinolone Products Liability Litigation—illustrate the potential for conflict resulting from different federal court approaches to Erie questions. Both courts were asked to decide the same issue: Whether brand-name manufacturers may be liable under Illinois law for injuries caused by their drug’s warning label when the label is affixed to a generic version of the drug. Though presented with the same legal issue, the courts took very different approaches and came to opposite conclusions. The practical consequence for injured plaintiffs is that their claims receive inequitable treatment when they should have been subject to the same law.

In re Zantac. The In re Zantac MDL is centralized in the Southern District of Florida.¹¹ The plaintiffs allege that Zantac can form the carcinogen N-nitrosodimethylamine when exposed to heat or ingested, increasing users’ risk of cancer. The Erie issue before the court was whether consumers who took a generic version of Zantac may hold brand-name manufacturers liable under various state tort theories.¹²

Specifically, the MDL court was tasked with making an Erie guess as to whether the highest courts in 35 jurisdictions—including Illinois—would recognize the plaintiffs’ liability theory. This issue had never been decided by the Illinois Supreme Court or Illinois’s appellate courts. And the Seventh Circuit, which encompasses Illinois, also had not addressed the issue.¹³

The MDL court began its analysis by doing what many federal courts do in these situations—looking to other federal court decisions for guidance.¹⁴ In doing so, the court noted the “overwhelming national consensus” among federal courts was that there was no such brand-name manufacturer liability.¹⁵ With respect to Illinois in particular, however, the court was faced with three federal court decisions predicting whether the Illinois Supreme Court would hold brand-name manufacturers liable for ingestion of their generic equivalent drugs.

Two of those opinions—Dolin v. SmithKline Beecham Corp.¹⁶ and Garner v. Johnson & Johnson, Janssen Research & Development LLC¹⁷—were issued by Illinois federal district courts, which, applying Illinois law, predicted the state’s high court would hold brand-name manufacturers liable under such circumstances. The third opinion was from outside Illinois: The Sixth Circuit’s decision in In re Darvocet, Darvon & Propoxyphene Products Liability Litigation rejected the reasoning of Dolin and predicted that the Illinois Supreme Court would reject this brand-name manufacturer liability theory.¹⁸

Concluding that the Sixth Circuit’s opinion was “sound and more persuasive” than the Illinois federal court decisions, the MDL court found that the plaintiffs’ claims would fail under Illinois law, thereby extinguishing the substantive rights of the Illinois MDL claimants who had taken generic Zantac.¹⁹ Notably, despite Erie’s mandate that federal courts sitting in diversity should apply the law of the state in making such predictions, there was no discussion, analysis, or citation to any Illinois state court case law in the MDL court’s opinion. Instead, the court relied solely on the federal court common law outlined in the Sixth Circuit’s opinion in In re Darvocet.

In re Fluoroquinolone. The In re Fluoroquinolone MDL is pending in the District of Minnesota.²⁰ That litigation involves the antibiotic class of drugs known as fluoroquinolones, which allegedly can cause peripheral neuropathy, a severe nerve injury. As in the In re Zantac litigation, the federal judge in In re Fluoroquinolone was tasked with deciding whether Illinois’s highest court would find the plaintiffs’ brand-name manufacturer liability theory viable.²¹

The court began its analysis by discussing many of the pivotal federal and state court cases that had considered brand-name manufacturer liability claims and their rationales for accepting or rejecting the theory, recognizing that permitting such claims represented the minority view.²² Notwithstanding this precedent, the court acknowledged its constitutional limitations under Erie and emphasized that “irrespective of decisions in other jurisdictions, the Court is obligated to apply Illinois law.”²³ After painstakingly analyzing each of the plaintiffs’ specific claims in light of Illinois law, the court found “that warning label claims against a brand-name drug manufacturer are viable pursuant to
Illinois law.”²⁴

Recognizing its disagreement with the In re Zantac decision, the court explained: “Because the Southern District of Florida relied solely on the Sixth Circuit’s analysis [in In re Darvocet] rather than examining the underlying Illinois case law,” the recent In re Zantac decision did not affect the court’s analysis here.²⁵ The court’s dismissive treatment of the In re Zantac decision is a not-so-subtle reminder that, pursuant to Erie, there is no such thing as federal common law in diversity cases. Indeed, unlike In re Zantac, the court’s meaningful examination of the plaintiffs’ individual claims under Illinois case law resulted in three of the plaintiffs’ five state law claims surviving.

Alternatives to Erie Guesswork

While it is not uncommon for federal judges to disagree about undecided issues of state law in diversity cases, the hope and expectation is that they would at least agree on the Erie requirement to rely solely on the law of the state in deciding the issue. As the conflicting outcomes of In re Zantac and In re Fluoroquinolone make clear, however, that is not always the case. So, what can MDL plaintiffs and their counsel do to try to avoid such inconsistent results?

One available, yet infrequently implemented, course of action is certification.²⁶ In the absence of an authoritative state court decision, a federal court may turn to state procedures for certifying undecided questions of state law to the state’s highest court. In general, this procedure permits the state’s highest court to consider questions certified to it by federal courts of appeal, federal district courts, or other federal courts.²⁷

Certification can be beneficial because it obviates the need for federal courts to make Erie guesses at all and allows the state’s supreme court the opportunity to interpret or change its own law—all while fostering cooperative judicial federalism.²⁸ Moreover, as a practical matter, state supreme courts are typically in a better position and have a more significant interest in evaluating the application of a state’s own laws than the federal judiciary.²⁹

Even though certification may be appropriate, successfully invoking certification in the context of an MDL can prove difficult.³⁰ In the rare instances when it has been successful, it has been because the parties pursued the certification issue at the earliest opportunity, the issue concerned a geographically distant state outside of the circuit presented with the question, and the court was less familiar with the other state’s law.³¹ Under such circumstances, the “practical considerations [can] militate in favor of certification.”³²

It is much more common, however, for MDL courts to refuse to certify questions of state law—for many equally practical reasons. First, the federal judge has the discretion to decide whether to certify a state law question; it is “not obligatory.”³³ And while the Supreme Court has endorsed the use of certification procedures,³⁴ it has not provided guidance to federal courts on the factors to be considered in exercising their discretion.³⁵

Thus, there is considerable conflict among the circuits on the approach to be taken. Some federal courts, citing the Supreme Court’s support for it, take a favorable view toward certifying questions for review;³⁶ others take a restrictive view and resort to certification only sparingly; and still others have charted a middle ground.³⁷ Regardless of the approach taken, the practical availability of certification in an MDL proceeding is inextricably tied to the jurisdiction to which it is transferred—a decision over which the litigants have no real control because it is determined by the Joint Panel on Multidistrict Litigation.

Second, the availability of certification procedures is not uniform. Every state (except North Carolina), the District of Columbia, and three territories have established certification procedures by court rule or by statute. However, its use in practice varies wildly, with some jurisdictions’ procedures allowing for liberal certification,³⁸ and others allowing it in limited circumstances only.³⁹

The result is that some litigants will “receive the constitutional benefits of certification” while others will not.⁴⁰ For MDL claimants, all of whom sink or swim together, that is an all-or-nothing choice.

For example, assume an MDL has been centralized before a federal judge in the District of Minnesota and that judge is presented with a novel question of state law involving a group of MDL claimants residing in Arkansas. That MDL judge, who is geographically distant from Arkansas and may be unfamiliar with its law, has the power to refuse to certify the question to the Arkansas Supreme Court and, instead, make an Erie guess as to how the state’s highest court would decide the issue of state law.

Further suppose that the federal judge decided against the Arkansas MDL claimants on that state law issue—and decided it incorrectly. Under these circumstances, the MDL judge’s Erie prediction would still be controlling on the Arkansas claimants’ state law claims in the MDL.

The inequity of such a result is palpable. As one federal judge has noted, “such incorrect predictions inevitably skew the decisions of persons and businesses who rely on them and inequitably affect the losing federal litigant who cannot appeal the decision to the state supreme court.”⁴¹

But consider the opposite scenario. If that MDL had been centralized before a different federal judge in Minnesota—one who favored certification—that state law question could have been certified to the Arkansas Supreme Court and, if accepted, those Arkansas claimants would have retained their state law claims.

Also, despite the Supreme Court’s endorsement of certification, there is considerable disagreement (even within the same circuit court) over whether certification accomplishes what it is designed to do—save time, cost, energy, and resources—especially in the MDL context.⁴² Thus, despite the “special competence”⁴³ state courts likely possess over issues of state law, it is unlikely that certification will succeed in MDLs unless it is pursued early in the litigation.⁴⁴

Accordingly, in the absence of a clear preference for the MDL judge to grant certification early in the litigation and a willing state supreme court to accept a certified question, practitioners should insist that the MDL court adhere to its constitutionally limited role in deciding an Erie question.

Emphasize to the court that, pursuant to Erie, there is no federal general common law in diversity cases. Remind the court that the statute on which Erie is based, the Rules of Decision Act, provides that “the laws of the several states, except where the Constitution or treaties of the United States or Acts of Congress otherwise require or provide, shall be regarded as rules of decision in civil actions in the courts of the United States, in cases where they apply.”⁴⁵ Thus, state law generally supplies the rules of decision in federal diversity cases, so that is where the court should look for the answer.

It is inappropriate to rely solely on the federal judiciary’s analysis of what the state’s highest court would do. Consistent with Erie’s rule, a federal court must instead look to and apply the law of the forum state. Only by doing so can the constitutional mandate of Erie be realized and plaintiffs’ viable state law claims protected.

R. Jason Richards is a partner at Aylstock, Witkin, Kreis and Overholtz in Pensacola, Fla., and can be reached at jrichards@awkolaw.com.

Notes

1. Connor Shaull, An Erie Silence: Erie Guesses and Their Effects on State Courts, Common Law, and Jurisdictional Federalism, 104 Minn. L. Rev. 1133, 1134–35 (2019).
2. Erie R.R. Co. v. Tompkins, 304 U.S. 64, 71, 78–79 (1938), overruling Swift v. Tyson, 41 U.S. 1 (1842) and emphatically rejecting the notion that federal courts sitting in diversity “are free to exercise an independent judgment as to what the common law of the state is—or should be.” See also Judiciary Act of 1789, ch. 20, §34, 1 Stat. 73, 92.
3. Erie, 304 U.S. at 73.
4. Id. at 78.
5. Hanna v. Plumer, 380 U.S. 460, 468 (1965).
6. See Shady Grove Orthopedic Assoc. v. Allstate Ins. Co., 559 U.S. 393, 456 (2010) (Ginsburg, J., dissenting); McCarthy v. Olin Corp., 119 F.3d 148, 159 (2d Cir. 1997) (Calabresi, J., dissenting); John L. Watkins, Erie Denied: How Federal Courts Decide Insurance Coverage Cases Differently and What to Do About It, 21.2 Conn. Ins. L.J. 455, 475 (2015); see also Patrick J. Borchers, The Real Risk of Forum Shopping: A Dissent from Shady Grove, 44 Creighton L. Rev. 29, 30 (2010).
7. Watkins, supra note 6, at 475. See, e.g., Corban v. United Servs. Auto. Ass’n, 20 So. 3d 601, 616–18 (Miss. 2009) (rejecting the Fifth Circuit’s interpretation and application of an anticoncurrent condition clause in Hurricane Katrina cases); State ex rel. Elvis Presley v. Crowell, 733 S.W.2d 89, 95–99 (Tenn. Ct. App. 1987) (rejecting the federal decisions as incorrect under state law).
8. Erie’s holding was deeply rooted in the fundamental constitutional principles of federalism and separation of powers. See Lindenberg v. Jackson Nat’l Life Ins. Co., 919 F.3d 992, 996 n.4 (6th Cir. 2019) (Bush, J., dissenting from the denial of rehearing en banc); Bradford R. Clark, Federal Common Law: A Structural Reinterpretation, 144 U. Pa. L. Rev. 1245, 1259 (1996).
9. Erie, 304 U.S. at 79.
10. See N.W. Airlines, Inc. v. Transp. Workers Union of Am. AFL-CIO, 451 U.S. 77, 95 (1981) (“[F]ederal courts, unlike their state counterparts, are courts of limited jurisdiction that have not been vested with open-ended lawmaking powers.”); Am. Cent. Ins. Co. v. Ouachita Coca-Cola Bottling Co., 149 F.3d 1173 (5th Cir. 1988) (“We thus heed our Erie directive and, being unable to determine that Louisiana has a public policy prohibiting this insurance coverage, we refrain from creating state law.”); Joyce v. A.C. & S., Inc., 785 F.2d 1200, 1208 (5th Cir. 1986) (Swygert, J., dissenting) (noting that, consistent with Erie, “this court has the duty to avoid creating state law if possible and leave the task to the Virginia Supreme Court and the legislature”).
11. In re Zantac (Ranitidine) Prods. Liab. Litig., 437 F. Supp. 3d 1368 (J.P.M.L. 2020) (Transfer Order to the Southern District of Florida).
12. See In re Zantac (Ranitidine) Prods. Liab. Litig., 510 F. Supp. 3d 1175, 1195–96 (S.D. Fla. 2020).
13. See Dolin v. GlaxoSmithKline LLC, 901 F.3d 803, 816 (7th Cir. 2018) (observing that Illinois courts have not yet considered the brand-name manufacturer theory but refusing to decide whether Illinois would adopt such a theory because it found that the plaintiff’s claim was preempted by federal regulations).
14. In re Zantac (Ranitidine) Prods. Liab. Litig., 437 F. Supp. 3d at 1196–98.
15. Id. at 1195.
16. See generally Dolin v. SmithKline Beecham Corp., 62 F. Supp. 3d 705, 714 (N.D. Ill. 2014).
17. Garner v. Johnson & Johnson, Janssen Research & Dev. LLC, 2017 WL 6945335, at *8 (C.D. Ill. Sept. 6, 2017).
18. In re Darvocet, Darvon & Propoxyphene Prods. Liab. Litig., 756 F.3d 917, 941–54 (6th Cir. 2014).
19. In re Zantac (Ranitidine) Prods. Liab. Litig., 510 F. Supp. 3d at 1210–11. The ruling would affect the Illinois plaintiffs already in the MDL, as well as those who registered claims with the MDL but have not filed them yet.
20. In re Fluoroquinolone Prods. Liab. Litig., 122 F. Supp. 3d 1378 (J.P.M.L. 2015) (Transfer Order to District of Minnesota).
21. See In re Fluoroquinolone Prods. Liab. Litig., 517 F. Supp. 3d 806, 815 (D. Minn. 2021).
22. Id. at 817.
23. Id.
24. Id. at 825.
25. Id. at 819 n.7.
26. “Certification is not to be routinely invoked whenever a federal court is presented with an unsettled question of state law.” Armijo v. Ex Cam, Inc., 843 F.2d 406, 407 (10th Cir. 1988). Under the diversity statutes, federal courts have a duty to decide questions of state law whenever necessary to render a judgment, even if the issues are difficult or uncertain. See Copier v. Smith & Wesson Corp., 138 F.3d 833, 838–39 (10th Cir. 1998); see also Lehman Bros. v. Schein, 416 U.S. 386, 394 (1974) (Rehnquist, J., concurring) (observing that certification “entails more delay and expense than would an ordinary decision of the state question on the merits by the federal court”).
27. See generally 17A Charles A. Wright & Arthur R. Miller, Federal Practice & Procedure §4248 (3d ed. 2021).
28. See, e.g., Geib v. Amoco Oil Co., 29 F.3d 1050, 1060–61 (6th Cir. 1994).
29. See, e.g., Oyens Feed & Supply, Inc. v. Primebank, 2015 WL 2082963, at *4 (N.D. Iowa May 5, 2015).
30. See, e.g., In re Zimmer NexGen Knee Implant Prods. Liab. Litig., 884 F.3d 746, 752 (7th Cir. 2018).
31. See, e.g., In re Amazon.com, Inc., 942 F.3d 297, 300 (6th Cir. 2019).
32. Id. at 300–01.
33. Lehman Bros., 416 U.S. at 390–91.
34. See Arizonans for Official English v. Arizona, 520 U.S. 43, 76 (1997).
35. Dolores K. Sloviter, A Federal Judge Views Diversity Jurisdiction Through the Lens of Federalism, 78 Va. L. Rev. 1671, 1685 (1992); see also Fiat Motors of N. Am. Inc. v. Mayor of Wilmington, 619 F. Supp. 29, 33 (D. Del. 1985).
36. See, e.g., Roth v. The Evangelical Lutheran Good Samaritan Soc’y, 147 F. Supp. 3d 806, 813 (N.D. Iowa 2015).
37. Watkins, supra note 6, at 479–80.
38. See, e.g., Ariz. Rev. Stat. Ann. §§12-1861–1867 (1993).
39. See, e.g., Wyo. Stat. Ann. §§1-13-104–107 (1977); see also Matter of Certified Question, 549 P.2d 1310, 1311 (Wyo. 1976) (holding that it will not answer a certified question until there is nothing left for the federal court to do but apply the state court’s answer to the question and enter judgment consistent with the answer).
40. Watkins, supra note 6, at 481.
41. Sloviter, supra note 35, at 1681.
42. Compare Toner v. Lederle Labs., 779 F.2d 1429, 1432 (9th Cir. 1986) (Certification “saves time, energy, and resources and helps build a cooperative judicial federalism.”) with Kremen v. Cohen, 325 F.3d 1035, 1043–44 (9th Cir. 2003) (Kozinski, J. dissenting) (“Certifying the case shifts the difficult work of deciding it to the state court, which is often so busy keeping its own house in order that it scarcely has time for our overflow laundry. Certification also burdens litigants, forcing them to reargue the case in a different forum—a process that is costly and full of delay.”).
43. R.R. Comm’n of Tex. v. Pullman Co., 312 U.S. 496, 499 (1941).
44. See, e.g., Lehman Bros., 416 U.S. at 394–95; Am. Fidelity Bank & Trust Co. v. Heimann, 683 F.2d 999, 1002 (6th Cir. 1982).
45. 28 U.S.C. §1652 (1944) (emphasis added). For example, in Guaranty Trust Co. v. York, 326 U.S. 99, 111–12 (1945), the Court held that Erie mandated that federal courts apply New York statutes of limitations to state law claims heard pursuant to the courts’ diversity jurisdiction and not the federal courts’ own ad hoc notions of equitable timeliness.