A market-based approach to internet intermediary strict products liability

ABSTRACT

This essay proposes a way of dealing with the strict liability of Internet sellers of other manufacturers’ products, such as Amazon under its ‘Fulfillment by Amazon’ program. I discuss and reject two approaches to the problem that have been proposed by the courts, and advance a view according to which the relevant inquiry is whether Internet intermediaries such as Amazon could have prevented a defective product from reaching the US market. This view accounts in a satisfactory manner for the notion of responsibility that is at the core of US strict products liability law, and avoids the pitfalls of alternative policies. However, since this view also entails a de facto quasi-immunity to lawsuits for Internet intermediaries in many cases, safeguards to that quasi-immunity are also addressed. While the essay focuses on US law, the principles and policies under discussion should be applicable in other jurisdictions as well.

KEYWORDS:

• Internet
• Amazon
• strict products liability

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 RESTATEMENT (SECOND) OF TORTS § 402A (1965).

2 Id.

3 Id. cmt. f.

4 Erie Ins. Co. v. Amazon.com, Inc., 925 F.3d 135, 138 (4th Cir. 2019).

5 Id.

6 Oberdorf v. Amazon.com, Inc., 295 F.Supp.3d 496, 498 (M.D.Pa. 2017) (aff’d in part, vacated in part by Oberdorf v. Amazon.com, Inc., 930 F.3d 136, 3rd Cir.(Pa.), in turn vacated on rehearing en banc, Oberdorf v. Amazon.com, Inc., 936 F.3d 182, 3rd Cir.(Pa.)).

7 RESTATEMENT (SECOND) OF TORTS § 402A (1965).

8 The discussion in Brownsword & Somsen is focused on non-US law, but the points made are valid.

9 Likewise, as early as 1982, Durham identified the rise of the strict product liability doctrine as one of the results of new social and economic pressures (Durham 1982, 618, n. 118). For a critical view of the impact of product liability laws on the market competitiveness of US companies, see, e.g. Cortese and Blaner (1989).

10 Erie Ins. Co. v. Amazon.com, Inc., 925 F.3d 135, 138 (4th Cir. 2019).

11 Id.

12 Id. at 140, 143.

13 Id. at 140–144.

14 Id. at 141.

15 Vagueness will persist, of course, in cases where it is not clear whether title to property was transferred. But presumably those cases are marginal in commercial contexts such as the one at issue.

16 Indeed, the 1996 Defense of Marriage Act (DOMA) stipulated such a definition for ‘marriage’ (1 U.S.C. § 7), only to be struck down by the Supreme Court in US v. Windsor, 570 US 744 (2013).

17 For other criticisms of dictionary use, see Mouritsen (2010), who identifies a number of fallacies courts may be prone to, and illustrates them with the Supreme Court’s practice in Muscarello v. United States, 524 US 125 (1998).

18 State Farm Fire and Cas. Co. v. Amazon.com, Inc., 390 F. Supp. 3d 964, 969–70 (W.D. Wis. 2019).

19 Fox v. Amazon.com, Inc., 930 F.3d 415, 422 (6th Cir. 2019). Similarly, the court in State Farm Fire reached the conclusion that ‘neither the statutes nor dictionaries provide decisive, or even helpful, definitions’. State Farm Fire and Cas. Co. v. Amazon.com, Inc., 390 F. Supp. 3d 964, 970 (W.D. Wis. 2019).

20 RESTATEMENT (FIRST) OF TORTS INTRO. 1934. See also Schofield: ‘Among a people as closely bound together as the people of the United States a large measure of uniformity in private law is not only desirable, but practically necessary’. (Schofield 1907–1908, 512)

21 See Jobe (1981). Although states did not ultimately adopt statutes based on the MUPLA, the ideal of uniformity in the product liability area is acknowledged even by practitioners who are favorably disposed toward state experimentation and trial and error (see Chief Justice Feldman’s comments before the Senate Judiciary Committee in connection with the proposed Product Liability Fairness Act, in Committee on the Judiciary 1994, 11). In recent times, there has been a trend toward federalization of a number of tort law areas that were traditionally the province of the states (Eades 2000). I am neutral on the question whether uniformity is best achieved by federalizing areas of tort common law, by states passing laws tailored according to model statutes, or in a different way.

22 See McKisson v. SalesAffiliates, Inc., 416 S.W.2d 787 (Tex. 1967), Galindo v. Precision American Corp., 754 F.2d 1212 (5th Cir. 1985).

23 See Delaney v. Towmotor Corp., 339 F.2d 4 (2d Cir. 1964), Beattie v. Beattie, 786 A.2d 549 (Del. Super. 2001), First Nat. Bank of Mobile v. Cessna Aircraft Co., 365 So. 2d 966, 967 (Ala. 1978).

24 See Francioni v. Gibsonia Truck Corp., 372 A.2d 736, 738 (Pa. 1977).

25 Fox v. Amazon.com, Inc., 930 F.3d 415, 423–24 (6th Cir. 2019).

26 Id. at 14.

27 Erie Ins. Co. v. Amazon.com, Inc., 925 F.3d 135, 144–145 (4th Cir. 2019).

28 Fox v. Amazon, at 424.

29 Id. at 425.

30 Id.

31 Id. In Fox, the Court was concerned mainly with an application of the Tennessee Products Liability Act, but the opinion makes it clear that consistency with products liability law in general is important: ‘Defendant also has the capability to spur the manufacturing and sale of safer products in the future, which is a primary purpose behind the doctrine of products liability in general’. Id. at 424. Moreover, consistency with decisions in other states that adopted control approaches to seller liability was another significant consideration, as the Court quotes several cases from various jurisdictions relying on the notion of control. These observations support the view that the discussion in Fox is relevant to determinations of strict products liability under § 402A of the Restatement, and is not limited to an application of Tennessee law.

32 Francioni, 372 A.2d at 739.

33 Id.

34 See RESTATEMENT (SECOND) OF TORTS § 402A cmt. c (1965). See also Henderson (1980, 1040–41).

35 Francioni, 372 A.2d at 739.

36 Id. at 740, note 3.

37 Musser v. Vilsmeier Auction Co., Inc., 562 A.2d 279, 281 (Pa. 1989).

38 Id. at 280.

39 Id. at 282.

40 Id. at 283.

41 The seller in Fox was not reachable by the plaintiff. Fox v. Amazon.com, Inc., 930 F.3d 415, 424 (6th Cir. 2019). That is one instance of a case where the only party the plaintiff could have recovered from was Amazon.

42 In all fairness, the court very likely did not need to bother with these distinction for the purpose of deciding the issue of lessors. But the question of which kind of circulation is supposed to be prevented by the seller becomes relevant in the context of the marketing of products on a global scale by complex business entities such as Amazon.

43 See the analysis in Henderson (1980). For certain categories of products (those substitutable non-commercially), an increase in the price charged by Amazon for its services (reflected, presumably, in an increase in the price of the product offered by the initial seller) will lead a certain number of purchasers (the ones for whom slight increases or decreases in price matter) to look for the product elsewhere. And sometimes products obtained non-commercially will present greater risks of injury than commercially-obtained products (precisely because their price has not internalized the risk of lawsuit, like their commercial counterparts). The frequency of occurrence of this kind of consumer behavior will depend on the kind of product at stake.

44 As a result, referring to this cluster of views as the ‘multi-factor approach’ seems better than the ‘balancing factors approach’.

45 As the court in Erie rightly observed, ‘Although Amazon’s services were extensive in facilitating the sale, they are no more meaningful to the analysis than are the services provided by UPS Ground, which delivered the headlamp … ’ Erie Ins. Co. v. Amazon.com, Inc., 925 F.3d 135, 142 (4th Cir. 2019).

46 See RESTATEMENT (SECOND) OF TORTS § 402A cmt. c (1965).

47 Id.

48 The phrases ‘what it is to be a seller for purposes of strict products liability’, and ‘seller liability’ are used interchangeably.

49 See, e.g. the analysis of the court in Fox v. Amazon.com, Inc., 930 F.3d 415 (6th Cir. 2019).

50 Note that, on this proposal, a certain principle of seller liability may be prima facie unfair, as long as it is not all- things-considered unfair.

51 RESTATEMENT (SECOND) OF TORTS § 402A cmt. c (1965). Similarly, the American Law of Products Liability states,

The relevant inquiry is whether the seller's conduct justifies a conclusion that it has undertaken a special responsibility for product safety, and whether it is equitable [emphasis added] to impose on the seller the loss caused by the product and the burden of spreading that loss as a cost of doing business. (§ 5:10.Strict liability– Who is a “seller”, American Law of Products Liability 3d § 5:10).

52 The acknowledgement of the notion of responsibility does not turn strict liability into negligence. In the absence of responsibility, there would be nothing to prevent liability attaching not only to the seller, but to UPS, the US. Postal Service, or the city along whose roads the product arrived at the customer’s door. So the Restatement’s talk of responsibility in the context of strict liability is justified.

53 Id.

54 Musser v. Vilsmeier Auction Co., Inc., 562 A.2d 279, at 283. (Pa. 1989).

55 State Farm Fire and Cas. Co. v. Amazon.com, Inc., 390 F. Supp. 3d 964, 967–68 (W.D. Wis. 2019). The district court in that case did not rely on the Restatement, but on Wisconsin statutes and precedent.

56 Id. at 972. Note that the court talks about both the American marketplace, and Wisconsin, as the relevant markets. The proposal that follows is focused on the US marketplace, not on particular states.

57 The court may be interpreted as using this as merely one factor among others. The proposal in this article is that something like this should be the cornerstone of the analysis in internet intermediary product liability cases.

58 On this restricted view of responsibility, the status of Amazon will literally be like that of the auctioneer in Musser, helpless to intervene when a product enters the market, and constituting simply an ‘extension’ of the seller’s business. See Musser, 562 A.2d at 283.

59 In the event Amazon refused to offer its platform to a seller of a certain product, the seller would sell the defective product elsewhere, so it would end up with a different customer.

60 See Musser, 562 A.2d at 283 (rejecting the extension of the quality of being a seller to auctioneers, under 402A). Auctioneers, in many cases, could have prevented a particular product from reaching a particular consumer, by refusing to offer the product at auction. That product would then have been sold elsewhere (maybe at another auction), probably to another consumer. Thus, on the strict view of what could have been prevented by the intermediary, auctioneers would have been liable in many situations.

61 Adding to the injustice, if Amazon is held to be a seller, it may very well have to pay more sales tax in various states. The Supreme Court, in South Dakota v. Wayfair, Inc., held that the state of South Dakota was within its rights to request an out-of-state online retailer with no physical presence in the state to collect and remit sales tax for sales of products to consumers in South Dakota. 138 S.Ct. 2080 (2018). As a consequence of that decision, more state legislatures will have an incentive to pass similar laws requiring online retailers like Amazon to pay sales taxes to the state. Of course, legislatures can do that irrespective of whether Amazon is deemed a seller under 402A, since something can be a seller for tax law purposes but not for product liability purposes. Nevertheless, holding Amazon liable for every sale it makes to a consumer will provide a further reason to justify such laws. The point is not that state tax laws (such as the one in South Dakota) are themselves unjust; it is only that they represent a burden on online retailers like Amazon, so their adoption should not be encouraged further by unjust strict product liability laws.

62 Moreover, they are probably irrelevant. Suppose it were possible to somehow identify all the members of the class of headlamp wearers, and that the law protected only members of that class. If a non-headlamp wearer decided one day to buy a headlamp, and it turned out to be defective, it is not clear why that person should not receive the protection of the law as well.

63 This is not to be confused with any requirement of citizenship or residence. Physical presence on the US territory is sufficient, on this proposal, to receive the protection of 402A.

64 Fox v. Amazon.com, Inc., 930 F.3d 415, 419 (6th Cir. 2019).

65 I want to thank a journal reviewer for pressing this point.

66 This is an essential reason why the view advanced in this article differs from the court’s holding in State Farm Fire. Even though the court upheld a test for what it is to be a ‘seller’ that depends on whether Amazon could have prevented the seller’s product from reaching the US market, the court appeared to consider only what Amazon itself did, and how it structured the contractual transaction with the seller under the ‘Fulfillment by Amazon’ policy. It seems that the court contemplated only the actions of Amazon, and, once it looked at the role of Amazon in the transaction, the inquiry was over (and Amazon turned out to be liable). On the contrary, on my proposal, if the seller would have reached the US market anyway, then the contractual terms between Amazon and the initial seller will not be dispositive, and Amazon will not be liable.

67 This applies in particular to corporations or other types of organizations, but in the rare case it can apply to individuals as well.

68 For an informative description of major provisions of Amazon’s Business Solutions Agreement with sellers, see the court’s description in Papataros v. Amazon.com, Inc., CV179836KMMAH, 2019 WL 4011502 (D.N.J. Aug. 26, 2019) (stayed by Papataros v. Amazon.com, Inc., 217CV9836KMMAH, 2019 WL 4740669 (D.N.J. Sept. 3, 2019)).

69 See generally Erie Ins. Co. v. Amazon.com, Inc., 925 F.3d 135 (4th Cir. 2019).and Fox v. Amazon.com, Inc., 930 F.3d 415 (6th Cir. 2019). The court in Eberhard v. Amazon noted that there appears to be ‘an emerging consensus against construing Amazon as a “seller” or “distributor” – and, therefore, against holding Amazon strictly liable for defective products sold on its website’. Eberhart v. Amazon.com, Inc., 325 F. Supp. 3d 393, 400 (S.D.N.Y. 2018). But see State Farm Fire and Cas. Co. v. Amazon.com, Inc., 390 F. Supp. 3d 964 (W.D. Wis. 2019) for the opposite result.

70 This is of course no legal immunity, but is basically a de facto immunity stemming from the fact that Amazon has no control over most US sellers’ ability to reach the US market. As section C argues, this quasi-immunity needs to be tempered by adopting adequate safeguards.

71 See e.g. South Dakota v. Wayfair, 138 S.Ct. 2080, 2099 (2018).

72 As will be recalled, that is why the formal approach to the problem is a failure.

73 Thanks to a reviewer for bringing this up.

74 In the context of rebuttable negligence rules in European tort law, van Dam notes: ‘If the defendant cannot prove that he was not at fault, his liability is not based on proven fault but solely on the presumption of fault. This is, in fact, a form of strict liability’ (van Dam 2013, 305). He also goes on to say that ‘[r]ules of liability for rebuttable negligence are often applied in cases in which the defendant possesses more information about the cause of the damage than the claimant’, a rationale that explains, at least in part, why Amazon should be held liable in these cases.

75 Fox v. Amazon.com, Inc., 930 F.3d 415, 418 (6th Cir. 2019).

76 Id. at 419–20.

77 For Amazon’s reaction to the problematic hoverboards, see Fox v. Amazon at 420.

78 If one worries that this condition imposes a negligence standard in what was supposed to be a case of strict liability, one should note, first, that the standard is brought in only as a safeguard against Amazon’s quasi-immunity, and that it does not replace Amazon’s potential strict liability. The easiest way to see this is to consider a perfectly possible situation where Amazon does its duty to inform the customers of the defective products, but it is still held strictly liable (because it could have prevented the product from entering the US market in the first place). Moreover, even if the approach under discussion may ultimately be characterized as some form of negligence liability, terminology is not the issue, and the proposal should stand or fall on its own merits, irrespective of what it is called. The line between negligence and strict liability is not always clear, and it has been argued that the concepts are closer than they are traditionally made out to be. See Abraham (2012), Salvador-Coderch, Garoupa, and Gómez-Ligüerre (2009).

79 See RESTATEMENT (SECOND) OF TORTS § 402A cmt. c (1965). See also Henderson, supra note 16, at 1041.

80 This assumption, reasonable as it seems, must await confirmation from empirical research in economics.