On practically no issue has this Blog been more insistent than on the principle of Erie conservatism when federal courts sitting in diversity undertake to “predict” state tort law. Our posts on this subject go back to 2006. At that time, we said:
In both of these decisions, novel questions of state law, involving state statutes intended to reduce tort liability, were answered so as to expand liability in ways that no state court had endorsed. Under established principles of federalism, that should not have happened. The Supreme Court made clear in Day & Zimmerman, Inc. v. Challoner, 423 U.S. 3 (1975), that “[a] federal court in diversity is not free to engraft onto those state rules exceptions or modifications which may commend themselves to the federal court, but which have not commended themselves to the State in which the federal court sits.” Id. at 4.
[A] federal court is not free to apply a different rule however desirable it may believe it to be, and even though it may think that the state Supreme Court may establish a different rule in some future litigation.
Hicks v. Feiock, 485 U.S. 624, 630 n.3 (1988).
Since then, we’ve returned to this proposition time and again – addressing it in the context of First Circuit, Third Circuit, Fourth Circuit, Fifth Circuit, Sixth Circuit, and Eleventh Circuit law − even proposing legislation to codify the Supreme Court’s position.
Now we’re aware of another appeal that has put the Erie conservatism principle front and center. We’ve mentioned before the winning streak that Amazon.com was on in product liability cases, and wondered whether that online business model could eventually affect the market for prescription drugs. Well, the case that prompted that post, Oberdorf v. Amazon.com, Inc., 295 F. Supp.3d 496 (M.D. Pa. 2017), was recently reversed, 2-1, by a Third Circuit panel. See Oberdorf v. Amazon.com Inc., ___ F.3d ___, 2019 WL 2849153 (3d Cir. July 3, 2019) (“applying” Pennsylvania law).
Until then, no court anywhere had found Amazon (or similar entities) to be a product liability “seller” for purposes of strict liability. See Fox v. Amazon.com, Inc., ___ F.3d ___, 2019 WL 2896326, at *7 (6th Cir. July 5, 2019) (applying Tennessee law); Erie Insurance Co. v. Amazon.com, Inc., 925 F.3d 135, 141-42 (4th Cir. 2019) (applying Maryland law); Milo & Gabby LLC v. Amazon.com, Inc., 693 F. Appx. 879, 885 (Fed. Cir. 2017) (applying federal copyright law); Stiner v. Amazon.com, Inc., 120 N.E.3d 885, 893-94 (Ohio App. 2019); Garber v. Amazon.com, Inc., 380 F. Supp.3d 766, 776-78 (N.D. Ill. 2019); Carpenter v. Amazon.com, Inc., 2019 WL 1259158, at *5 (N.D. Cal. March 19, 2019); Eberhart v. Amazon.com, Inc., 325 F. Supp.3d 393, 398-400 (S.D.N.Y. 2018); Allstate N.J. Insurance Co. v. Amazon.com, Inc., 2018 WL 3546197, at *7-12 (D.N.J. July 24, 2018); McDonald v. LG Electronics., USA, Inc., 219 F. Supp.3d 533, 542 (D. Md. 2016); Inman v. Technicolor USA, Inc., 2011 WL 5829024, at *6 (W.D. Pa. Nov. 18, 2011).
Thus, for the Third Circuit to predict that Pennsylvania law would be to the contrary was a huge leap into the unknown that a federal court sitting in diversity jurisdiction should never have made. As Judge Scirica, in dissent, pointed out:
This case implicates an important yet relatively uncharted area of law. No Pennsylvania court has yet examined the product liability of an online marketplace like Amazon’s for sales made by third parties through its platform. Our task, as a federal court applying state law, is to predict how the Pennsylvania Supreme Court would decide the case. We must take special care to apply state law and not to participate in an effort to change it.” In my view, well-settled Pennsylvania products liability law precludes treating Amazon as a “seller” strictly liable for any injuries caused by the defective [product].
Oberdorf, 2019 WL 2849153, at *12 (citation and quotation marks omitted) (dissent).
Here’s a link to Amazon’s petition for en banc review, filed recently. The petition leads with the Erie conservatism principle:
[T]he panel decision is contrary to decisions of the Supreme Court and this Court holding that federal courts sitting in diversity cannot “act as … judicial pioneer[s]” by deciding “whether and to what extent they will expand state common law.” City of Phila. v. Lead Indus. Ass’n, 994 F.2d 112, 123 (3d Cir. 1993). The majority’s creation of an unprecedented form of “seller” liability under Pennsylvania law − with far-reaching consequences not just for Amazon but also for countless other businesses and service providers − conflicts with Lead Industries Association, as well as Day & Zimmermann v. Challoner, 423 U.S. 3 (1975), Sheridan v. NGK Metals Corp., 609 F.3d 239 (3d Cir. 2010), Travelers Indemnity Co. v. Dammann & Co., 594 F.3d 238 (3d Cir. 2010), Werwinski v. Ford Motor Co., 286 F.3d 661 (3d Cir. 2002), City of Philadelphia v. Beretta U.S.A. Corp., 277 F.3d 415 (3d Cir. 2002), Adams v. Madison Realty & Development, 853 F.2d 163 (3d Cir. 1988), Falcone v. Columbia Pictures Industries, 805 F.2d 115 (3d Cir. 1986), Bruffett v. Warner Communications, 692 F.2d 910 (3d Cir. 1982), and McKenna v. Ortho Pharmaceutical Corp., 622 F.2d 657 (3d Cir. 1980).
Amazon petition at 2; see id. at 9-12 (discussing this precedent in more detail).
So how did we do? Comparing our 2009 Erie-in-the-Third-Circuit post with the Amazon brief, Amazon relies on two later-decided Third Circuit cases, which of course we couldn’t cite. But as for the rest, Amazon, had four we didn’t have (Adams, Falcone, Bruffett, and McKenna), but we also had four that Amazon didn’t cite: Lexington National Insurance Corp. v. Ranger Insurance Co., 326 F.3d 416, 420 (3d Cir. 2003); Camden County Board of Chosen Freeholders v. Beretta, U.S.A. Corp., 273 F.3d 536, 541-42 (3d Cir. 2001); Northview Motors, Inc. v. Chrysler Motors Corp., 227 F.3d 78, 92 n.7 (3d Cir. 2000); and Leo v. Kerr-McGee Chemical Corp., 37 F.3d 96, 101 (3d Cir. 1994). Call it a draw – but the larger point is that there is a long line of Third Circuit authority should have prevented the leap taken in Oberdorf. In that respect, Oberdorf reminds us of the same court’s disregard of its own prior precedents that preemption was a legal issue in In re Fosamax (Alendronate Sodium) Products Liability Litigation, 852 F.3d 268, 288 & n.106 (3d Cir. 2017), rev’d, 139 S. Ct. 1668 (2019) (on precisely this point) – something we commented on here.
What’s equally frustrating to us was the supposed “test” for being a strict liability “seller” that the Third Circuit majority employed. The Oberdorf majority “consider[ed] whether the following four factors apply”:
(1) Whether the actor is the only member of the marketing chain available to the injured plaintiff for redress;
(2) Whether imposition of strict liability upon the actor serves as an incentive to safety;
(3) Whether the actor is in a better position than the consumer to prevent the circulation of defective products; and
(4) Whether the actor can distribute the cost of compensating for injuries resulting from defects by charging for it in his business. . . .
Oberdorf, 2019 WL 2849153, at *4, see Francioni v. Gibsonia Truck Corp., 372 A.2d 736, 739 (Pa. 1977) (articulating factors).
Talk about a test biased in favor of finding liability. Factors one and four are blatant cost-spreading-based liability shifters dating from the 1970s heyday of strict liability. Factors two and three, while ostensibly safety-related, likewise almost unerringly point to liability. Any liability, strict or otherwise, can virtually always be construed as an “incentive” to safe conduct or a “deterrent” to its opposite. Likewise, when if ever will a “consumer” be “better positioned . . . to prevent the circulation of defective products”? The fix was in with this test.
The Pennsylvania Supreme Court recognized the bias of this test, however, in a case subsequent to what Oberdorf relied on – and that Oberdorf majority failed to cite – a case that just happens to be a prescription medical product liability decision. Coyle v. Richardson-Merrell, Inc., 584 A.2d 1383, 1387 (Pa. 1991). Coyle rejected the imposition of strict liability on pharmacies as supposed “sellers” of drugs, cautioning:
In support of their argument, [plaintiffs] can assert no more than the [defendant’s] ability to obtain insurance and/or indemnification as a means of distributing the potential cost of liability. Reliance on cost-shifting as the only factor to be considered in whether a given party should be exposed to liability, however, would result in absolute liability rather than strict liability.
Id. at 1387. See also Cafazzo v. Central Medical Health Services, Inc., 668 A.2d 521, 526 (Pa. 1995) (reiterating that “[t]o assign liability for no reason other than the ability to pay damages is inconsistent with our jurisprudence”) (citing Coyle).
How online websites fit into product liability is an intriguing question, as we discussed in our original Oberdorf post. Some of our clients would support liability, some would not − so we take no position on that question. However, from day one, the DDL Blog has been foursquare in our opposition to federal courts adopting expansive state-law theories of liability in diversity jurisdiction cases. The Third Circuit violated fundamental Erie principles in its rush to liability in Oberdorf, and that needs correcting. It’s been thirty years since the United States Supreme Court last reined in liability overreach in diversity jurisdiction, but perhaps it will get another chance in Oberdorf.