I. FACTS

Loomis ordered a hoverboard on Amazon’s website on November 28, 2015. The listing identified the seller to be TurnUpUp, a name used by SMILETO to sell its products on Amazon’s marketplace. SMILETO is allegedly a company based in China. The hoverboard was shipped to Loomis by Forrinx Technology (USA), Inc. Loomis was notified by Amazon that the product shipped on December 1, 2015. On December 11, 2015, Loomis sent an e-mail inquiring whether the hoverboard would be delivered in time for Christmas. The e-mail was sent through Amazon’s website. Loomis received the hoverboard on December 16, 2015. Loomis gifted the hoverboard to her son. On New Year’s Eve, he plugged it into an outlet in Loomis’s bedroom to

2 charge. Loomis’s boyfriend later discovered a fire burning in her bedroom. Her bed and the hoverboard were on fire. Loomis suffered burns to her hand and foot as a result of fighting the fire. A. Third Party Sales on Amazon.com Amazon.com is an online marketplace where Amazon and third party sellers list their products for sale. Amazon describes its marketplace as “an online mall” which provides an “online storefront” to third party sellers. Where Amazon is the seller of a product, it is identified as the seller on the product detail page, and it sources the product, sets the price, and holds title to it. This case does not involve an Amazon-listed product. Where a third party is the seller, it is identified as such on the product detail page and again on the order confirmation page before the user places the order. The third party sources the product, sets the price, and holds title to it. All third party sellers operate under the Amazon Services Business Solutions Agreement (BSA). The BSA requires a seller to “ensure that [it is] the seller of each of [its] Products” and to provide Amazon with accurate, updated product information in a specified format. A third party seller chooses what products to sell and at what price. However, the BSA requires pricing parity, where the price is “at least as favorable to Amazon Site users as the most favorable terms” offered by the seller elsewhere. Amazon provides payment processing for all third party sales. It remits the purchase price to the third party seller on a set schedule minus any service fees it may charge. Amazon collects a “referral fee,” a percentage of the sale price per item sold by the third-party seller, depending on the nature of the item sold. For toys, such as hoverboards, the referral fee is 15 percent of the sale price. The seller is required to route all payments and

3 refunds through Amazon, who may withhold payments, sometimes permanently, from the seller based upon its investigation of any disputes or claims. Refunds due to purchasers are calculated by the seller according to Amazon’s refund policies and routed through Amazon. The BSA further required all communications between the seller and buyer to be made through Amazon. Under the BSA, Amazon expressly reserves the right to control of its website and listings: “We have the right in our sole discretion to determine the content, appearance, design, functionality and all other aspects of the Amazon Sites, including by redesigning, modifying, removing or restricting access to any of them, and by suspending, prohibiting or removing any listing.” The BSA also allows Amazon in its sole discretion to refuse to process or cancel any transactions. The seller must also ensure its materials, products, offers and sales comply with all applicable laws. The BSA advises third party sellers they are “responsible for any non-conformity or defect in, or any public or private recall of, any of [their] Products or other products provided in connection with [the] Products.” In addition to Amazon’s own efforts to monitor recalls, the BSA requires third party sellers to notify Amazon “promptly” of any public or private recalls of their products. Sellers also agree to indemnify Amazon from any liability arising from its products. For those sellers whose gross proceeds exceed a specified threshold, which was applicable to TurnUpUp, the BSA also requires them to acquire excess insurance naming Amazon as an additional insured. Some third party sellers utilize Fulfillment by Amazon (FBA) services, which allow the seller to store its inventory in an

4 Amazon warehouse. If a product is sold under the FBA, Amazon packages and ships the product to the purchaser. TurnUpUp did not elect to utilize the FBA services. Amazon provides purchasers with what it calls an “A-to-z Guarantee:” “We want you to buy with confidence any time you make a purchase on the Amazon.com Website or use Amazon Pay. . . . The condition of the item you buy and its timely delivery are guaranteed under the unconditional A-to-z Guarantee.” Amazon, however, does not consider the A-to-z Guarantee to constitute a warranty for the products it lists. It instead warns purchasers in its Conditions of Use that third parties sell products through Amazon and that Amazon is “not responsible for examining or evaluating, and [does] not warrant, the offerings of any of these businesses or individuals . . . . Amazon does not assume any responsibility or liability for the actions, product, and content of all these and any other third parties.” B. The Sale of Hoverboards on Amazon More than 380,000 hoverboards were purchased through Amazon in 2015. The vast majority were sold by third party sellers. Under the BSA, Amazon charged TurnUpUp various fees for its services, including a $39.99 monthly nonrefundable subscription fee and a 15 percent referral fee that was calculated from the total sales price of each product. For the transaction at issue, Amazon received a referral fee of $55.50 from the $370 sale of the TurnUpUp hoverboard to Loomis. For the period between September 14, 2015, and December 16, 2015, TurnUpUp sold hoverboards totaling $736,366.68 through Amazon. Amazon received $110,645.92 in fees from those sales. In late November 2015, Amazon’s product safety team began investigating hoverboards in response to press reports that

5 hoverboards were involved in fires. It identified 17 reports of fire or smoke allegedly caused by hoverboards sold through Amazon. These 17 reports involved different models and manufacturers. On December 10, 2015, Amazon decided to remove all third party hoverboard listings from its website. It sent all prior hoverboard purchasers an e-mail notifying them of the reports of safety problems with hoverboards. Loomis testified in her deposition she did not recall receiving such an e-mail from Amazon. During this time, the Consumer Product Safety Commission (CPSC) conducted its own investigation into hoverboard safety and was in contact with Amazon about it. On February 18, 2016, the CPSC issued a letter stating that it regarded hoverboards that do not comply with a draft Underwriters Laboratories voluntary standard as presenting a potential “substantial product hazard.” The CPSC announced recalls of certain hoverboard models in July 2016. C. The Legal Proceedings Loomis brought suit against Forrinx1 and Doe defendants for products liability and fraud on September 2, 2016. In a form complaint, Loomis alleged three “counts” related to the product liability cause of action. Count one alleged strict products liability, count two alleged negligence, and count three alleged a breach of warranty. Loomis later amended her complaint to substitute Amazon into the lawsuit for a Doe defendant. Amazon moved for summary judgment on a number of grounds, including that it did not fall within the chain

1 Forrinx failed to appear and a default was entered against it.

6 of distribution for product liability purposes, it could not be liable under the “marketing enterprise theory” for those entities that fell outside the chain of distribution, and the federal Communications Decency Act (47 U.S.C. § 230) (CDA) barred Loomis’s claims.2 The trial court granted Amazon’s motion for summary judgment. Loomis timely appealed.

II. DISCUSSION

At issue in this appeal are Loomis’s strict and negligent product liability claims.3 She contends summary adjudication was improperly granted due to Amazon’s participation in the vertical chain of distribution for the product. Amazon disclaims any liability on the ground it is neither a manufacturer nor seller of the hoverboard. We conclude there exist triable issues of

2 Under the CDA, “[n]o provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider.” (47 U.S.C. § 230 (c)(1).) “No cause of action may be brought and no liability may be imposed under any State or local law that is inconsistent with this section.” (Id., (e)(3).) 3 Loomis acknowledges her fraud claim (second cause of action) and breach of warranty claim (count three of the first cause of action) are not at issue in this appeal. In the respondent’s brief, Amazon informed us that we need not address its CDA-based arguments as “[Loomis] has abandoned her fraud claim and frames her remaining claims as based solely on Amazon’s role in the transaction, not on the website’s content.” We understand this statement to mean Amazon concedes the CDA only applied to shield it from Loomis’s fraud claim and not her strict liability and negligence claims. Accordingly, we agree we need not address whether the CDA provides immunity to Amazon from the strict liability and negligence claims.

7 material fact which warrant reversal of summary adjudication as to these claims.4 A. Standard of Review A defendant moving for summary judgment must show that one or more elements of a cause of action cannot be established or that there is a complete defense to the cause of action. (Code Civ. Proc., § 437c, subd. (p)(2); Miller v. Department of Corrections (2005) 36 Cal.4th 446, 460.) Once the defendant’s burden has been met, the plaintiff is required to show a triable issue of material fact as to the cause of action or defense. (Code Civ. Proc., § 437c, subd. (p)(2).) A triable issue of fact is created when the evidence reasonably permits the trier of fact, under the applicable standard of proof, to find the purportedly contested fact in favor of the party opposing the motion. (Lugtu v. California Highway Patrol (2001) 26 Cal.4th 703, 722.) The plaintiff may not rely on the allegations in her pleadings but must set forth the specific facts showing the triable issue. (Code Civ. Proc., § 437c, subd. (p)(2).) We review the record de novo, considering all the evidence set forth in the moving and opposing papers except that to which

4 Loomis also contends her claim for punitive damages survives summary adjudication because her products liability claims do. Amazon does not address this issue in its respondent’s brief. Below, it sought summary adjudication of punitive damages on the ground Loomis had no viable underlying claims to support it. It presented no argument as to the merits of the claim. Loomis is correct that a punitive damages award may properly be entered in a strict products liability suit. (Grimshaw v. Ford Motor Co. (1981) 119 Cal.App.3d 757, 807-808, disapproved on a different ground by Kim v. Toyota Motor Corp. (2018) 6 Cal.5th 21.)

8 objections were made and sustained. We liberally construe the evidence in support of the plaintiff opposing summary judgment and resolve doubts concerning the evidence in her favor. (Code Civ. Proc. § 437c, subd. (c); Yanowitz v. L’Oreal USA, Inc. (2005) 36 Cal.4th 1028, 1037; AARTS Productions, Inc. v. Crocker National Bank (1986) 179 Cal.App.3d 1061, 1064–1065.) B. The Doctrine of Strict Products Liability in California Greenman v. Yuba Power Products, Inc. (1963) 59 Cal.2d 57, 62 (Greenman) established the doctrine of strict products liability when it held “[a] manufacturer is strictly liable in tort when an article he places on the market, knowing that it is to be used without inspection for defects, proves to have a defect that causes injury to a human being.” “The purpose of such liability is to insure that the costs of injuries resulting from defective products are borne by the manufacturers that put such products on the market rather than by the injured persons who are powerless to protect themselves.” (Id. at p. 63.) The California Supreme Court extended the doctrine to retailers in Vandermark v. Ford Motor Co. (1964) 61 Cal.2d 256 (Vandermark), reasoning, “Retailers like manufacturers are engaged in the business of distributing goods to the public. They are an integral part of the overall producing and marketing enterprise that should bear the cost of injuries resulting from defective products. [Citation.] In some cases the retailer may be the only member of that enterprise reasonably available to the injured plaintiff. In other cases the retailer himself may play a substantial part in insuring that the product is safe or may be in a position to exert pressure on the manufacturer to that end; the retailer’s strict liability thus serves as an added incentive to

9 safety. Strict liability on the manufacturer and retailer alike affords maximum protection to the injured plaintiff and works no injustice to the defendants, for they can adjust the costs of such protection between them in the course of their continuing business relationship.” (Id. at pp. 262-263.) California courts must consider the policies underlying the doctrine to determine whether to extend strict liability in a particular circumstance. (Anderson v. Owens–Corning Fiberglas Corp. (1991) 53 Cal.3d 987, 995 (Anderson); O’Neil v. Crane Co. (2012) 53 Cal.4th 335, 362–363 (O’Neil).) The public policies articulated in Greenman and Vandermark that form the foundation for the application of strict liability are the following: (1) whether Amazon may play a substantial part in insuring that the product is safe or may be in a position to exert pressure on the manufacturer to that end, (2) whether Amazon may be the only member in the distribution chain reasonably available to the injured plaintiff, and (3) whether Amazon is in a position to adjust the costs of compensating the injured plaintiff amongst various members in the distribution chain. (Vandermark, supra, 61 Cal.2d at pp. 262-263.) Applying these policy considerations, courts have extended strict products liability to entities within the chain of distribution, including bailors and lessors (Price v. Shell Oil Company (1970) 2 Cal.3d 245, 248); wholesalers and distributors (Barth v. B. F. Goodrich Tire Co. (1968) 265 Cal.App.2d 228, 252- 253; Canifax v. Hercules Powder Co. (1965) 237 Cal.App.2d 44, 52 (Canifax)); and sellers of mass-produced homes (Kriegler v. Eichler Homes, Inc. (1969) 269 Cal.App.2d 224, 227). Courts have found these defendants were responsible for passing the product down the line to the consumer, had the ability to affect

10 product safety by exerting pressure on the manufacturer, and were able to bear the cost of compensating for injuries. (Arriaga v. CitiCapital Commercial Corp. (2008) 167 Cal.App.4th 1527, 1535 (Arriaga).) Courts, however, have declined to extend the doctrine to hotel proprietors (Peterson v. Superior Court (1995) 10 Cal.4th 1185); sellers of used products (Wilkinson v. Hicks (1981) 126 Cal.App.3d 515); and auctioneers (Tauber–Arons Auctioneers Co. v. Superior Court (1980) 101 Cal.App.3d 268 (Tauber-Arons)), who were found to have little to no relationship with the manufacturer and thus lacked the ability to affect product safety. A consumer injured by a defective product “may now sue ‘any business entity in the chain of production and marketing, from the original manufacturer down through the distributor and wholesaler to the retailer; liability of all such defendants is joint and several.’ ” (Wimberly v. Derby Cycle Corp. (1997) 56 Cal.App.4th 618, 628.) The purpose for this approach “is to extend liability to all those engaged in the overall producing and marketing enterprise who should bear the social cost of the marketing of defective products.” (Kaminski v. Western MacArthur Co. (1985) 175 Cal.App.3d 445, 455-456.) “The strict liability doctrine derives from judicially perceived public policy considerations, i.e., enhancing product safety, maximizing protection to the injured plaintiff, and apportioning costs among the defendants. [Citations.] Where these policy justifications are not applicable, the courts have refused to hold the defendant strictly liable even if that defendant could technically be viewed as a ‘ “link in the chain” ’ in getting the product to the consumer market. [Citation.] In other words, the facts must establish a sufficient causative relationship or connection between the defendant and the product

11 so as to satisfy the policies underlying the strict liability doctrine.” (Arriaga, supra, 167 Cal.App.4th at p. 1535.) The court in Bay Summit Community Assn. v. Shell Oil Co. (1996) 51 Cal.App.4th 762 (Bay Summit), set forth three factors to determine whether such a causative relationship or connection exists when the defendant falls outside the vertical chain of distribution. Under the marketing enterprise theory or stream of commerce approach, the plaintiff must show: “(1) the defendant received a direct financial benefit from its activities and from the sale of the product; (2) the defendant’s role was integral to the business enterprise such that the defendant’s conduct was a necessary factor in bringing the product to the initial consumer market; and (3) the defendant had control over, or a substantial ability to influence, the manufacturing or distribution process.” (Id. at p.776; Kasel v. Remington Arms Co. (1972) 24 Cal.App.3d 711 (Kasel).) C. Bolger v. Amazon.com LLC In Bolger, the Fourth District applied strict products liability analysis to substantially identical facts as presented in this case. There, the plaintiff bought a replacement battery for her laptop from a third party seller through the Amazon website. The battery allegedly exploded several months later, causing severe burns. (Bolger, supra, 53 Cal.App.5th at p. 437.) The trial court granted summary judgment in favor of Amazon on the ground it did not distribute, manufacture, or sell the product in question. The Fourth District reversed. (Id. at p. 439.) The Bolger court found Amazon was “a direct link in the chain of distribution, acting as a powerful intermediary between the third party seller and the consumer.” (Bolger, supra, 53 Cal.App.5th at p. 438.) “As a factual and legal matter, Amazon

12 placed itself between [the seller] and Bolger in the chain of distribution of the product at issue here. Amazon accepted possession of the product from [the seller], stored it in an Amazon warehouse, attracted Bolger to the Amazon website, provided her with a product listing for [the seller’s] product, received her payment for the product, and shipped the product in Amazon packaging to her. Amazon set the terms of its relationship with [the seller], controlled the conditions of [the seller’s] offer for sale on Amazon, limited [the seller’s] access to Amazon’s customer information, forced [the seller] to communicate with customers through Amazon, and demanded indemnification as well as substantial fees on each purchase. Whatever term we use to describe Amazon’s role, be it ‘retailer,’ ‘distributor,’ or merely ‘facilitator,’ it was pivotal in bringing the product here to the consumer.” (Ibid.) After concluding Amazon was a link in the chain of distribution, the Bolger court found the policies articulated in Greenman, Vandermark, and their progeny supported imposition of strict liability. The court found Amazon was the only member of the enterprise reasonably available to an injured consumer in some cases, it played a substantial part in ensuring the products listed on its website were safe, it could and did exert pressure on upstream distributors like the instant seller to enhance safety, and it had the ability to adjust the cost of liability between itself and its third party sellers. The court concluded Amazon should be held liable if a product sold through its website turned out to be defective, since strict liability in the instant case afforded maximum protection to the injured plaintiff and worked no injustice to Amazon. (Bolger, supra, 53 Cal.App.5th at pp. 453- 455.)

13 In Bolger, Amazon presented many of the same arguments to disclaim liability as it does in this case. (Bolger, supra, 53 Cal.App.5th at pp. 455-458.) The court rejected each of these arguments. Amazon initially focused on dictionary or Commercial Code definitions of “seller” and “distributor” to argue those definitions did not apply to it. It asserted it was a service provider that was not subject to strict liability and that had no control over the selection of the product or its pricing. (Bolger, supra, 53 Cal.App.5th at p. 456.) The court explained strict products liability was created judicially because of the economic and social need for the protection of consumers in an increasingly complex and mechanized society. The scope of strict liability was expanded when necessary to account for market realities and to cover new transactions, such as this one, in widespread use in today’s business world regardless of labels or dictionary definitions. (Bolger, supra, 53 Cal.App.5th at p. 456.) The court rejected Amazon’s assertion it was merely a service provider with no control over the product. It held, “Nothing aside from Amazon’s own choices required it to allow [the seller] to offer its product for sale, to store [the seller’s] product at its warehouse, to accept Bolger’s order, or to ship the product to her. It made these choices for its own commercial purposes. It should share in the consequences.” (Bolger, supra, 53 Cal.App.5th at p. 457.) The court concluded Amazon’s control over both the product and the sales transaction formed the basis for its liability. (Id. at pp. 458-459.) Additionally, Amazon argued, as it does here, that the Legislature, not the court, was the appropriate forum to address

14 whether those policies would be served in new contexts. The Bolger court found this argument ran “directly contrary to California law” on strict liability, which was created by the courts and expanded and contracted where warranted by its purposes. (Bolger, supra, 53 Cal.App.5th at p. 459.) Lastly, the court found the federal CDA did not shield Amazon from strict liability because liability was based on Amazon’s own conduct, not the content of the seller’s product listing. (Bolger, supra, 53 Cal.App.5th at p. 465.) Amazon contends Bolger was erroneously decided because it ignores long-standing limitations on strict liability law. According to Amazon, these long-standing limitations include a “threshold requirement” for strict liability expressed in O’Neil and case law excluding service providers from strict liability. We explain below (post, sec. D,3) how Amazon has misinterpreted O’Neil and the cases regarding liability of service providers. We are also not persuaded by Amazon’s argument that Bolger improperly applied strict liability to “facilitators.” Amazon argues, “Before the Fourth District’s Bolger decision, the lower courts uniformly . . . rejected claims against service providers that facilitated, or were even necessary to consummate, sales.” In each of those cases, however, the courts did not refuse to impose strict liability simply because the defendant facilitated a sale. The courts each examined the underlying policies and determined strict liability as to finance lessors and auctioneers was not appropriate. (Arriaga, supra, 167 Cal.App.4th at p. 1537 [“The policy considerations behind imposing strict products liability on those involved in the vertical distribution of consumer goods are not furthered by including finance lessors.” (Italics omitted.)]; Tauber-Arons, supra, 101 Cal.App.3d at p. 274 [“The

15 real issue in this case is whether the policy of the doctrine of strict liability established by the decisions of the appellate courts in this state justifies the imposition of strict liability under such circumstances.”]; Brejcha v. Wilson Machinery, Inc. (1984) 160 Cal.App.3d 630 [relying on Tauber-Arons].) Amazon next criticizes Bolger for relying on “vague and ill- defined policy notions” that were “cited six decades ago.” As Bolger observed, this argument directly conflicts with the Supreme Court’s repeated observation that “[t]he question whether to apply strict liability in a new setting is largely determined by the policies underlying the doctrine.” (O’Neil, supra, 53 Cal.4th at pp. 362–363; Anderson, supra, 53 Cal.3d at p. 995; see also Bay Summit, supra, 51 Cal.App.4th at p. 774.) We conclude Bolger has properly applied well-established strict liability law to the facts of its case and was correctly decided. D. Analysis 1. Vertical Chain of Distribution As technology advances, innovation is paving the way to new business practices. Amazon is on the leading edge of e- commerce. Based on our review of Amazon’s third-party business model under the BSA, we are persuaded that Amazon’s own business practices make it a direct link in the vertical chain of distribution under California’s strict liability doctrine. Contrary to Amazon’s assertion that it merely provided an online storefront for TurnUpUp and others to sell their wares, it is undisputed Amazon placed itself squarely between TurnUpUp, the seller, and Loomis, the buyer, in the transaction at issue. When Loomis wanted to buy a hoverboard for her son, she perused product listings on Amazon’s website. Amazon took Loomis’s order and processed her payment. It then transmitted

16 the order to TurnUpUp, who packaged and shipped the product to Loomis. When Loomis wondered whether the hoverboard would arrive in time for Christmas, she communicated her concerns through Amazon. TurnUpUp was not allowed to communicate with Loomis directly. If Loomis had wanted to return the hoverboard, the return would have been routed through Amazon. Amazon remitted Loomis’s payment to TurnUpUp after deducting its fees, including a 15 percent referral fee based on the total sale price. These facts undermine Amazon’s characterization of its marketplace as an online mall providing online storefronts for sellers. Owners of malls typically do not serve as conduits for payment and communication in each transaction between a buyer and a seller. Moreover, they do not typically charge a per-item fee rather than a fixed amount to rent their storefronts. Instead, these actions – 1) interacting with the customer, 2) taking the order, 3) processing the order to the third party seller, 4) collecting the money, and 5) being paid a percentage of the sale – are consistent with a retailer or a distributor of consumer goods. 2. Stream of Commerce Approach Although we conclude Amazon is a link in the vertical chain of distribution, we nevertheless recognize e-commerce may not neatly fit into a traditional sales structure. The stream of commerce approach or market enterprise theory offers an alternative basis for strict liability. “[U]nder the stream-of-commerce approach to strict liability no precise legal relationship to the member of the enterprise causing the defect to be manufactured or to the member most closely connected with the customer is required

17 before the courts will impose strict liability. It is the defendant’s participatory connection, for his personal profit or other benefit, with the injury-producing product and with the enterprise that created consumer demand for and reliance upon the product (and not the defendant’s legal relationship (such as agency) with the manufacturer or other entities involved in the manufacturing- marketing system) which calls for imposition of strict liability. [Citation.]” (Kasel , supra, 24 Cal.App.3d at p. 725.) Thus, a defendant may be strictly liable under the stream of commerce approach if: “(1) the defendant received a direct financial benefit from its activities and from the sale of the product; (2) the defendant’s role was integral to the business enterprise such that the defendant’s conduct was a necessary factor in bringing the product to the initial consumer market; and (3) the defendant had control over, or a substantial ability to influence, the manufacturing or distribution process.” (Bay Summit, supra, 51 Cal.App.4th at p. 778.) Viewing the evidence in the light most favorable to Loomis, there exists a triable issue of material fact as to each of the three factors identified above. First, Amazon does not dispute it received a direct financial benefit in the form of fees, including a monthly subscription fee and a 15 percent referral fee, from the sale of the product. However, it contends the fees it charges must also be connected to its activities in bringing the product to the initial consumer market. (Bay Summit, supra, 51 Cal.App.4th at p. 776 [“Shell directly profited from its activities in creating a market for the polybutylene plumbing system and from each sale of the plumbing system.”].) It argues it has engaged in no such activities, and has derived no financial benefit, because the market for hoverboards existed regardless of Amazon.

18 Contrary to Amazon’s broad interpretation of what constitutes “the initial consumer market,” caselaw tells us “the relevant market in which the defendant was deemed a participant was the initial distribution to the consuming public of the particular defective product of a given manufacturer,” “not just products of the same classification.” (Tauber-Arons, supra, 101 Cal.App.3d at p. 276; see also Kasel, supra, 24 Cal.App.3d at p. 725 [defendant’s participation in the enterprise which distributed and created consumer demand for “Remington Express” shot shells]; Bay Summit, supra, 51 Cal.App.4th at p. 776 [polybutylene plumbing system].) Therefore, we need to determine whether Amazon received a direct financial benefit from its activities to bring TurnUpUp hoverboards to market. To this end, Amazon presented no evidence that it did not play a role in establishing a market for TurnUpUp hoverboards. Instead, it presented testimony that it had no information what other sales channels TurnUpUp used to sell its hoverboards. By contrast, the evidence shows Amazon received $110,645.92 from its sale of TurnUpUp hoverboards from September 14, 2015 to December 16, 2015. At a minimum, this evidence creates a triable issue of material fact as to whether Amazon received a direct financial benefit from its activities and sales of the product. For the same reason, there exists a triable issue with respect to the second factor—whether Amazon’s role was integral to the business enterprise. Amazon has failed to meet its burden to demonstrate this factor cannot be established because it has presented no evidence of its role in bringing TurnUpUp’s hoverboards to market. Third, the record demonstrates Amazon had substantial ability to influence the manufacturing or distribution process

19 through its ability to require safety certification, indemnification, and insurance before it agrees to list any product. For example, the BSA allows Amazon to require certification of products it lists from the Underwriter’s Laboratories, which, among other things, establishes standards for manufacturing practices. Amazon’s contention that it has no relationship with the manufacturer or the distributors has no bearing on whether it can influence the manufacturing process. We are persuaded the trial court erroneously granted summary adjudication on the strict liability claim based on a stream of commerce approach. 3. Service Provider Amazon disputes it plays a role in the vertical distribution chain or the stream of commerce because it is not a manufacturer, seller, or supplier, but merely a service provider. It seizes on a single sentence in O’Neil to create a “threshold requirement” for the imposition of strict liability. The sentence in O’Neil reads: “That the defendant manufactured, sold, or supplied the injury-causing product is a separate and threshold requirement that must be independently established.” (O’Neil, supra, 53 Cal.4th at p. 362.) Amazon would have us interpret that sentence in O’Neil to limit strict liability only to manufacturers, sellers, and suppliers. Read in context, however, it is clear O’Neil did not intend to overturn five decades of case law extending strict liability to lessors, bailors, and others within the stream of commerce who may not bear the label of manufacturer, seller, or supplier. (See, e.g., Fortman v. Hemco, Inc. (1989) 211 Cal.App.3d 241, 251 [“entities in the stream of commerce for purposes of strict liability are not limited to those

20 readily identifiable as designer, manufacturer, or vendor of the defective product”].) In O’Neil, the plaintiff sought to impose strict liability on a manufacturer who supplied a nondefective part that was later used with a defective product that caused injury. By this sentence, the high court merely indicated an innocent manufacturer may not be held strictly liable for a defective product made, sold, or supplied by someone else. (O’Neil, supra, 53 Cal.4th at p. 362.) We likewise reject Amazon’s argument it is merely a service provider who is not subject to strict products liability. We have identified how it was instrumental in the sale of the hoverboard to Loomis. Even if Amazon may be characterized as a service provider, Murphy v. E. R. Squibb & Sons, Inc. (1985) 40 Cal.3d 672 (Murphy) and Hernandezcueva v. E.F. Brady Co., Inc. (2015) 243 Cal.App.4th 249 (Hernandezcueva), cited by Amazon, are instructive on the issue of when strict liability attaches to service providers. In both cases, the court determined the defendant provided both a service and sale of a product. Therefore, “[t]he propriety of imposing strict liability on a party that both supplies and installs a defective component hinges on the circumstances of the transaction.” (Hernandezcueva, supra, 243 Cal.App.4th at p. 260.) In Murphy, the court determined the service aspect of the defendant pharmacist’s role predominated over its sale of prescription drugs. (Murphy, supra, 40 Cal.3d at p. 675.) In Hernandezcueva, the court determined the subcontractor that installed drywall in a commercial project provided both a service (the installation) and the sale of a product (the drywall). Despite

21 its dual role, the subcontractor was a participant in the stream of commerce for strict liability purposes. (Hernandezcueva, supra, at p. 263.) Here, Amazon provides a service to TurnUpUp in the form of a website to list its product and, as described above, was also instrumental in the sale of the product by placing itself squarely between TurnUpUp and Loomis. That it did not hold title to the product and did not have physical possession of the hoverboard does not automatically render it solely a service provider and remove it from strict liability. Canifax, supra, 237 Cal.App.2d 44 is instructive on this issue. In Canifax, the customer purchased the defective fuse (and related supplies) from a jobber, who in turn placed an order with the defendant, who passed on the order for the fuse to the manufacturer. The manufacturer shipped the fuse directly to the jobber and the defendant paid the manufacturer from the proceeds it received from the jobber. (Canifax, supra, 237 Cal.App.2d 44 at p. 48.) The defendant disputed strict liability on the basis it did not manufacture the fuse, sell the fuse, and never had possession of the fuse. The court held, “The fact that it chooses to delegate the manufacture of [the] fuse to another and that it causes the manufacturer to ship the product directly to the consumer cannot be an escape hatch to avoid liability.” (Id. at p. 52.) The circumstances surrounding the sale of the hoverboard are materially similar to those surrounding the sale of the fuse in Canifax. As in Canifax, Amazon did not manufacture the hoverboard, did not sell the hoverboard, and never had physical possession of the hoverboard. These facts, however, “cannot be an escape hatch to avoid [strict] liability.” (Canifax, supra,

22 237 Cal.App.2d at p. 52.) This is because, like the defendant in Canifax, Amazon took the order for the hoverboard, took the payment, and passed the order up the chain of distribution. (Id. at p. 50.) 4. Out-of-State Cases We are not persuaded by Amazon’s reliance on those decisions that restrict strict liability to sellers or manufacturers by applying out-of-state law. (Erie Ins. Co. v. Amazon.com, Inc. (4th Cir. 2019) 925 F.3d 135, 140 [“products liability under Maryland law. . . is imposed on sellers and manufacturers (a manufacturer also being a seller)”]; Stiner v. Amazon.com, Inc. (Ohio 2020) 2020-Ohio-4632 [Ohio Products Liability Act]; Fox v. Amazon.com, Inc. (6th Cir. 2019) 930 F.3d 415, 425 [Tennessee Products Liability Act]; Milo & Gabby LLC v. Amazon.com, Inc. (Fed.Cir. 2017) 693 F.Appx 879, 890 [addressing federal intellectual property law]; Garber v. Amazon.com, Inc. (N.D.Ill. 2019) 380 F.Supp.3d 766, 779 [Illinois law on products liability]; State Farm Fire and Cas. Co. v. Amazon.com, Inc. (W.D.Wis. 2019) 390 F.Supp.3d 964 [Wisconsin products liability statute].) We need not stray so far afield when California courts have provided extensive analysis of strict liability doctrine in California. We have identified the acts constituting Amazon’s participation in the distribution chain. “Whatever term we use to describe Amazon’s role, be it ‘retailer,’ ‘distributor,’ or merely ‘facilitator,’ it was pivotal in bringing the product here to the consumer.” (Bolger, supra, 53 Cal.App.5th at p. 438.) These acts support our conclusion that Amazon is in the vertical chain of distribution of the alleged defective hoverboard. Next, we assess whether applying strict liability to Amazon’s third-party seller

23 business model supports the relevant public policy considerations. 5. Policy Considerations Underlying the Doctrine Are Furthered by Imposing Strict Products Liability in This Case In analyzing whether strict liability is appropriate in new circumstances, courts assess whether relevant public policy goals are furthered by its application. (Bay Summit, supra, 51 Cal.App.4th at p. 774.) As noted earlier in this opinion, the relevant public policy considerations are: (1) whether Amazon may play a substantial part in insuring that the product is safe or may be in a position to exert pressure on the manufacturer to that end, (2) whether Amazon may be the only member in the distribution chain reasonably available to the injured plaintiff, and (3) whether Amazon is in a position to adjust the costs of compensating the injured plaintiff amongst various members in the distribution chain. (Vandermark, supra, 61 Cal.2d at pp. 262- 263.) We address each in turn. As to product safety, Amazon contends it has no proactive authority over product design or manufacture because its relationship is typically with the distributor or retailer, not the manufacturer. It asserts it can only reactively address safety issues by removing or suspending sellers after a product has been shown to be unsafe. Thus, imposing strict liability on it will not enhance product safety. By its very argument, Amazon acknowledges it plays a role in ensuring product safety. Indeed, sellers are required under the BSA to comply with all applicable laws, including, presumably, consumer safety laws and regulations. The evidence shows Amazon may require proof that a product complies with

24 recognized safety standards before it is listed. For example, Amazon may require documentation or certification from the Underwriter’s Laboratories, which sets forth standards for manufacturing practices and devises safety tests for products. Additionally, Amazon on occasion has consulted with the CPSC to determine if there are specific regulations or issues related to products it lists that are “outside of the norm.” These steps, which Amazon has taken to ensure product safety in limited circumstances, refute its contention it has no ability to proactively affect product safety. Application of strict liability in this case may incentivize Amazon to expand its safety compliance requirements to more products and thus further the goal of product safety. Moreover, we agree with Bolger that “[j]ust like a conventional retailer, Amazon can use its power as a gatekeeper between an upstream supplier and the consumer to exert pressure on those upstream suppliers (here, third party sellers) to enhance safety.” (Bolger, supra, 53 Cal.App.5th at p. 454.) Under the BSA, “Amazon sets fees that it would retain for the sale of a third-party product, protects itself by requiring third- party vendors to indemnify Amazon should any ‘claim, loss, damage, settlement, cost, expense or other liability’ occur, and reserves the right to refuse to provide . . . services for a product that does not comport with Amazon’s policies. With the rights retained, Amazon could halt the placement of defective products in the stream of commerce, deterring future injuries.” (Gartner v. Amazon.com, Inc. (S.D.Tex. 2020) 433 F.Supp.3d 1034, 1044, citation omitted.) As to consumer compensation, Amazon may be the only member of the distribution chain reasonably available for an

25 injured consumer to recover damages. Amazon contends there is no evidence to show how frequently an injured plaintiff is truly left without recourse. The record shows, however, that Forrinx, the only other defendant in this matter, failed to appear and a default was taken against it. Indeed, the same issue arose in Bolger, where two defendants failed to appear, and a third could only be served in China. (Bolger, supra, 53 Cal.App.5th at p. 453; see also Fox v. Amazon.com, Inc., supra, 930 F.3d at p. 424.) As to loss spreading, Amazon can adjust the costs of consumer protection between it and third party sellers through its fees, indemnity requirements, and insurance. Amazon does not argue it is unable to spread the costs in these, and other, ways. It instead contends these costs will result in higher prices to be paid by consumers and small businesses will also be required to bear these higher costs. Amazon’s argument obfuscates the goal of loss spreading. At issue is whether the loss should be borne solely by the injured consumer who may have no recourse against other defendants, or whether it may be spread among those who profited from the sale of the defective product. In our view, these policy considerations support the application of strict liability to Amazon’s third party seller business model. Lastly, Amazon contends the policies articulated by the California Supreme Court in Greenman and Vandermark are incomplete and have been eroded by subsequent events. Indeed, Amazon questions whether strict liability has brought benefits commensurate with its substantial costs. These are issues best directed to the California Supreme Court, who created strict products liability doctrine almost 60 years ago and whose decisions we are bound to follow. (Auto Equity Sales, Inc. v. Superior Court (1962) 57 Cal.2d 450, 455.)

26 Accordingly, we hold the application of strict liability to Amazon’s third-party seller business model is supported by the relevant public policy consideration discussed in Vandermark.

III. Summary Adjudication Was Improperly Granted as

to the Negligent Products Liability Claim Loomis also challenges the summary adjudication of her negligence claim, contending Amazon failed to meet its burden to demonstrate that one or more elements of the claim cannot be established or that there is a complete defense to the claim. Amazon argues it does not owe Loomis a duty of care for the same reason it is not strictly liable: it did not manufacture or sell the product. It asserts “[a]n element of a product-liability claim is that the defendant manufactured or sold the product. Her failure to establish that element doomed her . . . strict-liability and negligence claims.” Amazon’s position is reasonable given that “[t]he theories of negligence and strict liability ‘parallel and supplement each other’ [citation], and the same policy considerations that militate for or against imposition of strict liability may apply with equal force in the context of negligence. [Citations.]” (Bettencourt v. Hennessy Industries, Inc. (2012) 205 Cal.App.4th 1103, 1118.) However, Amazon provides no legal support for its argument that negligent products liability may only be imposed on manufacturers and sellers. Instead, a duty of care may be imposed on a defendant who is not a manufacturer or seller in the context of a negligent products liability claim if certain policy factors are met. The O’Neil court explained, “Courts of this state have traditionally considered several factors in determining the existence and scope of duty: ‘the foreseeability of harm to the plaintiff, the degree of certainty that the plaintiff suffered injury, the closeness of the

27 connection between the defendant’s conduct and the injury suffered, the moral blame attached to the defendant’s conduct, the policy of preventing future harm, the extent of the burden to the defendant and consequences to the community of imposing a duty to exercise care with resulting liability for breach, and the availability, cost, and prevalence of insurance for the risk involved.’ ” (O’Neil, supra, 53 Cal.4th at p. 364.) Aside from those factors which parallel strict liability doctrine, Amazon provided no analysis as to how it did not owe a duty of care to Loomis under those factors. For this reason, the trial court erred in granting summary adjudication as to the negligent products liability count.5

5 Our conclusion that Amazon has failed to meet its burden on summary adjudication is not the equivalent of a finding that Amazon owed Loomis a duty of care or was negligent under these circumstances.

28

DISPOSITION

The judgment is reversed. The trial court is directed to vacate its order granting summary judgment. A new order shall be entered (1) denying summary adjudication of Loomis’s strict products liability and negligent products liability claims as well as her claim for exemplary damages, and (2) granting summary adjudication of the remaining claims. Upon entry of the new order, the trial court shall conduct further proceedings not inconsistent with this opinion. Loomis shall recover her costs on appeal.

OHTA, J.*

I concur:

STRATTON, Acting P. J.

* Judge of the Los Angeles Superior Court, assigned by the Chief Justice pursuant to article VI, section 6 of the California Constitution.

29 Wiley, J., Concurring. The Amazon is the world’s largest river. Amazon.com supposedly chose its trademark because it aimed to create the world’s largest river of commerce. Amazon.com can control what it created. Unbeknownst to Amazon, a manufacturer may use Amazon’s site to sell a defective product that will cause future accidents. Perhaps it is a computer battery that can explode. (Bolger v. Amazon.com, LLC (2020) 53 Cal.App.5th 431, 444 (Bolger).) Perhaps it is a hoverboard that ignites fires. Whatever it is, Amazon is situated swiftly to learn of and to contain the emerging problem, thereby reducing accidental injuries. Amazon can cabin the danger by stopping sales. Amazon can alert past buyers who have yet to experience the lurking hazard: Amazon has information about its customers and their purchases. Other measures are possible. Once Amazon is convinced it will be holding the bag on these accidents, this motivation will prompt it to engineer effective ways to minimize these accident costs. Tort law will inspire Amazon to align its ingenuity with efficient customer safety. Customers will benefit. Amazon maintains it already safeguards customer safety. (E.g., Bolger, supra, 53 Cal.App.5th at pp. 442–443.) On appeal, Amazon’s brief says it “voluntarily implemented a product safety system years ago, independent of the specter of tort liability, and . . . it spends hundreds of millions of dollars annually to ensure that products offered on the website are safe, compliant, and authentic.” The cited source of this assertion is not the record, but rather an online press release “written by Amazon” posted in “response to a Wall Street Journal story about the safety of

1 products offered in [the Amazon.com] store.” (Amazon, Product safety and compliance in our store (Aug. 23, 2019), [as of Apr. 5, 2021], archived at (AboutAmazon.com).) Amazon’s cited post makes the following admissions: “[W]e [at Amazon.com] have developed, and continuously refine and improve, our tools that prevent suspicious, unsafe, or non-compliant products from being listed in our store. [¶] Our proactive measures begin when a seller attempts to open an account. Our new seller account vetting includes a number of verifications and uses proprietary machine learning technology that stops bad actors before they can register or list a single product in our store. All products offered in our stores must comply with applicable laws and regulations, and our own policies. For example, we require toys to be tested to relevant safety standards set by the Consumer Product Safety Commission. We have a dedicated global team of compliance specialists that review submitted safety documentation, and we have additional qualification requirements that sellers must meet to offer products. In 2018, our teams and technologies proactively blocked more than three billion suspect listings for various forms of abuse, including non-compliance, before they were published to our store. [¶] Once a product is available in our store, we continuously scan our product listings and updates to find products that might present a concern. Every few minutes, our tools review the hundreds of millions of products, scan the more than five billion attempted daily changes to product detail pages, and analyze the tens of millions of customer reviews that are submitted weekly for signs of a concern and

2 investigate accordingly. Our tools use natural language processing and machine learning, which means new information is fed into our tools daily so they can learn and constantly get better at proactively blocking suspicious products. [¶] In addition, we provide a number of ways for regulatory agencies, industry organizations, brands, customers, and our customer service teams to report safety issues. When we receive these reports, we move quickly to protect customers, remove unsafe products from our store, and investigate. For example, if a customer reports a concern with a product, a customer service associate can instantly trigger an investigation. Additionally, because of our direct relationships with customers, we are able to trace and directly notify customers who purchased a particular product online and alert them to a potential safety issue—our systems are far more effective than other online and offline retailers and customers can feel confident they’ll have the information they need. [¶] We also regularly work with agencies including the Food and Drug Administration and Consumer Product Safety Commission, and the information we share helps them identify trends and develop regulations. We are active in industry working groups and committees that are dedicated to developing new solutions and guidelines that will benefit all retailers and consumers. [¶] We invest significant resources to protect our customers and have built robust programs designed to ensure products offered for sale in our store are safe and compliant. We want customers to shop with confidence and if ever a customer has a concern, they can contact our customer service team, and we will investigate.” (AboutAmazon.com, supra.)

3 These words are good. The admissions confirm the obvious: Amazon can control its river. It can undertake cost-effective steps to minimize accidents from defective products sold on its website. Strict tort liability will underline the priority Amazon places on its safety efforts. Thus we have an easy case that beautifully illustrates the deep structure of modern tort law: a judicial quest to minimize the social costs of accidents—that is, the sum of the cost of accidents and the cost of avoiding accidents. Judges have been applying this social cost-benefit analysis as a felt instinct for a long time, as we shortly will survey. That deep structure makes this case simple to decide. When efforts to minimize accident costs are relatively inexpensive and apt to be effective, courts impose tort duties. Amazon has cost-effective options for minimizing accident costs. Amazon therefore has a duty in strict liability to the buyers from its site, including Kisha Loomis. I In large measure, tort law is cost-benefit analysis. Comparing the cost to the benefit of an activity is rational thinking. All else equal, it is rational to invest in ventures where the expected return exceeds the cost. Conversely, it is irrational to spend $2 of your own money to get something you value only at $1. The judicial use of cost-benefit analysis is familiar throughout tort doctrine, in the doctrines of negligence as well as strict liability. In contrast to formal cost-benefit analyses in other spheres, this judicial practice does not compare quantified sums of dollar and cents on a balance sheet; case records nearly never contain precise data of the right kind. Rather, the

4 litigation setting forces judicial reasoning to be more seat-of-the- pants, more commonsensical than that. We see this commonsensical method over and over through the decades. The intellectual history of cost-benefit analysis in tort law goes back at least a century. (E.g., Berkovitz v. American River Gravel Co. (1923) 191 Cal. 195, 199 (Berkovitz); see also Chicago, Burlington & Quincy Railroad Co. v. Krayenbuhl (Neb. 1902) 65 Neb. 889, 902–904 [91 N.W. 880, 882–883].) Shortly we return to the 1923 Berkovitz decision. In 1944, Justice Roger Traynor—California’s most esteemed jurist—explained tort doctrine must aim to minimize the social costs of accidents. (Escola v. Coca Cola Bottling Co. (1944) 24 Cal.2d 453, 462 (conc. opn. of Traynor, J.) [“public policy demands that responsibility be fixed wherever it will most effectively reduce the hazards to life and health inherent in defective products”] (Escola); see also Greenman v. Yuba Power Products, Inc. (1963) 59 Cal.2d 57, 63–64 (Traynor, J.) [unanimous court adopts logic of Traynor’s concurring Escola opinion] (Greenman).) Justice Traynor’s Greenman decision had enormous impact. “Within a decade of Greenman, a majority of jurisdictions in the United States had adopted causes of action in strict product liability. Today all but a handful of states employ some version of products liability law.” (Goldberg et al., Tort Law: Responsibilities and Redress (3d ed. 2012) p. 887.) Shortly after Escola, Judge Learned Hand formalized tort law’s efficiency calculus with enduring precision. (United States v. Carroll Towing Co. (2d. Cir. 1947) 159 F.2d 169, 173 [defendant’s duty is a function of three variables: (1) the

5 probability of an accident; (2) the gravity of the injury from an accident; versus (3) the burden—that is, the cost—of adequate precautions]; see Posner, A Theory of Negligence (1972) 1 J. Legal Stud. 29, 32–33.) In 1970, Professor (and now Senior United States Circuit Judge) Guido Calabresi treated tort law’s deep structure in his landmark book, The Costs of Accidents (Calabresi). In Calabresi’s approach, courts can use cost-benefit analysis to appraise whether defendants can make cost-effective accident avoidance investments. Where the benefit of investments in accident avoidance outweighs the cost, courts should impose tort duties on defendants. But courts refrain from imposing tort duties in situations where the options for accident avoidance are so limited that the costs of accident avoidance do not outweigh the benefits. (Cf. Posner, Guido Calabresi’s The Costs of Accidents: A Reassessment (2005) 64 Md. L.Rev. 12, 15 [Calabresi’s framework approximated cost-benefit analysis].) In 1978, the California Supreme Court’s Barker decision explicitly put cost-benefit analysis into the heart of tort law. (Barker v. Lull Engineering Co. (1978) 20 Cal.3d 413.) The question was whether a vehicle called a high-lift loader was a defective product. Lumber tipped off a loader working on a hilly construction site. The falling lumber injured the driver, who sued the manufacturer, saying the loader was a defective product: the manufacturer could have made the loader with stabilizing mechanical arms, which the driver maintained would have prevented the tipping and his injury. The manufacturer disagreed, saying its loader design was not defective at all; rather the problem was the bad decision to use the loader on a steep hill, where of course things would tip over. To resolve the issue, our

6 Supreme Court crafted a cost-benefit standard: do the benefits of the challenged design outweigh the risk of danger inherent in such design? (Id. at pp. 418–421 & fn. 2, 426–427, 430–435.) Thus, a product is defective if its design embodies excessive preventable danger: that is, unless the benefits of the design outweigh the risk of danger inherent in such design. (Soule v. General Motors Corp. (1994) 8 Cal.4th 548, 567, 570.) The Barker decision also formulated the alternative “consumer expectations” test. (Barker, at pp. 429–430.) By the 1980s, the notion of tort cost-benefit analysis was mainstream. (See Evra Corp. v. Swiss Bank (7th Cir. 1982) 673 F.2d 951, 958 (Posner, J.) [“The amount of care that a person ought to take is a function of the probability and magnitude of the harm that may occur if he does not take care”].) In 2016, the California Supreme Court cited Calabresi’s enduring cost-benefit approach in its unanimous Kesner opinion. (Kesner v. Superior Court (2016) 1 Cal.5th 1132, 1153, citing Calabresi, supra [courts should assign tort liability to ensure those best situated to prevent injuries are incentivized to do so] (Kesner).) We return to Kesner shortly. This developing tort doctrine brings us to today and to Amazon. Under this doctrine, Amazon owes its customers a duty in strict liability because Amazon’s position in the distribution chain allows it to take cost-effective steps to reduce accidents. The cost- benefit analysis in Amazon’s case is not a close call: the benefits of the actions Amazon can take to minimize accidents vastly outweigh the costs of these actions to Amazon. Amazon’s options are practical and cost-effective; indeed, Amazon says it is already

7 taking these actions. Amazon thus must face strict liability for Loomis’s fiery encounter with the hoverboard she bought from Amazon’s site. Imposing this duty on Amazon creates financial incentives that back up Amazon’s good words about its concern for customer safety. Some suggest considerations of moral justice can compete with tort law’s calculus of social benefit. (E.g., Simons, Tort Negligence, Cost-Benefit Analysis, and Tradeoffs: A Closer Look at the Controversy (2008) 41 Loyola L.A. L.Rev. 1171, 1172–1173; Schwartz, Mixed Theories of Tort Law: Affirming Both Deterrence and Corrective Justice (1997) 75 Tex. L.Rev. 1801, 1819–1820.) This case presents no potential conflict like that. The only time Amazon’s brief uses the word “justice” is to write “the Bolger Court saw the ‘novelty of these issues’ as an opening to use law as ‘an instrument of justice’ to implement ‘current concepts of what is right and just.’ 53 Cal. App. 5th at 462 (quoting Kriegler v. Eichler Homes, Inc., 269 Cal. App. 2d 224, 227 (1969)).” If they ever do, moral justice and cost-benefit analyses do not conflict in this case. II The case law jibes completely with this cost-benefit analysis. This is true for decisions that impose tort duties, as well as for decisions that refuse to impose tort duties. We see this consistency in both categories of decisions. We begin with decisions that impose duties. A Courts impose tort duties on entities situated to take cost- effective measures to reduce the costs of accidents. Six precedents illustrate this point.

8 1 Justice Traynor’s Escola opinion proposed making Coca Cola strictly liable for an exploding cola bottle. The company was positioned to fix the problem through efficient accident avoidance measures. Coca Cola could have pressurized bottles more carefully, bought stronger bottles, tested more thoroughly, switched to cans, or used any combination of tactics. (See Escola, supra, 24 Cal.2d at p. 459 [“The bottle was admittedly charged with gas under pressure, and the charging of the bottle was within the exclusive control of [Coca Cola]”]; id. at p. 461 [Coca Cola “had exclusive control over both the charging and inspection of the bottles”]; id. at p. 462 (conc. opn. of Traynor, J.) [the bottler “is best situated to afford such protection” from the problem of exploding soda bottles].) Coca Cola made the key decisions and so was positioned to make soda containers safer—or, if cost- effective fixes were impossible, to stop marketing the product altogether. Justice Traynor thus urged strict liability to force the company to internalize the accident costs it had been imposing on its customers. The Supreme Court unanimously affirmed the judgment against Coca Cola. (Id. at p. 461.) 2 In the same way, the Greenman decision imposed strict liability on a lathe maker that could have used better set screws to hold spinning wood more securely. Improved screws could have made the wood less likely to fly off the lathe and hurt the operator. (See Greenman, supra, 59 Cal.2d at p. 60.) Again, tort law forced the manufacturer to internalize the costs of accidents from its product.

9 3 The Vandermark opinion extended strict liability to a car retailer that sold a new Ford to one Chester Vandermark. (Vandermark v. Ford Motor Co. (1964) 61 Cal.2d 256.) Vandermark claimed the car’s braking defect caused it to crash. For the unanimous court, Justice Traynor applied strict liability to the retailer as well as to the manufacturer, because the retailer was positioned to take actions to reduce the costs of accidents from defective products. The retailer was part of the manufacturer’s overall marketing enterprise and in some cases would be the only member of that enterprise reasonably available to injured plaintiffs. The retailer’s continuing business relationship with the manufacturer would allow dealers to exert pressure on the possibly remote manufacturer as an added safety incentive. (See id. at pp. 261–263.) The dealer never suggested its ability to exert pressure on Ford was so costly and ineffective as to be impractical. 4 Rowland v. Christian (1968) 69 Cal.2d 108 is another important case imposing tort liability on a decision maker positioned to undertake efficient accident precautions. The opinion put a duty on Nancy Christian, who knew the bathroom faucet handle in her apartment was cracked and needed replacing. Christian invited James Rowland to the apartment. Rowland said he was going to the bathroom. The handle broke and cut Rowland when he tried to use it. The court held Christian owed a duty to warn Rowland of the faucet crack. (Id. at pp. 110–112.) This ruling made perfect cost-benefit sense. It would have cost Christian little to share her knowledge of the latent danger

10 with Rowland. The information would have allowed Rowland to take suitable care. Imposing this accident avoidance duty on the knowledgeable property possessor was efficient and thus socially rational. The fact this was a negligence case shows cost-benefit analysis in tort law is not confined to the context of strict liability. 5 The previously mentioned Kesner case illustrated the logic of tort law’s cost-benefit analysis, and again outside the strict liability setting. Employees unwittingly carried workplace asbestos dust home to family members, who breathed it in and developed mesothelioma from the take-home exposure. The unanimous Supreme Court ruled the employers had a duty to the family members. (Kesner, supra, 1 Cal.5th at pp. 1141, 1156, 1165.) The tort goal was “incentivizing reasonable preventative measures.” (Kesner, supra, 1 Cal.5th at p. 1151.) The employers did “not claim that precautions to prevent transmission via employees to off-site individuals—such as changing rooms, showers, separate lockers, and on-site laundry—would unreasonably interfere with business operations.” (Id. at p. 1153.) Employers “are generally better positioned than their employees or members of their employees’ households to know of the dangers of asbestos and its transmission pathways, and to take reasonable precautions to avoid injuries that may result from on-site and take-home exposure.” (Ibid.; see also id. at p. 1156 [“Businesses making use of asbestos were well positioned, relative to their workers, to undertake preventive measures, and [the employers] cite no evidence to suggest such measures would have been unreasonably costly.”].)

11 The cost-benefit analysis thus was simple: the employers had cost-effective accident avoidance measures available to them, so they owed a duty in tort to family members facing take-home asbestos exposure. (See Kesner, supra, 1 Cal.5th at pp. 1154– 1156.) 6 Amazon repeatedly cites a Court of Appeal decision: Bay Summit Community Assn. v. Shell Oil Co. (1996) 51 Cal.App.4th 762 (Bay Summit), which concerned Shell Oil’s liability for leaky polybutylene plumbing systems. This decision is instructive, as is the whole polybutylene saga. Both cut against Amazon. Polybutylene was a plastic used for water piping. “Beginning in the late 1970s, polybutylene plastic plumbing systems—touted as being cheaper and more durable than copper pipe systems—were installed in new homes nationwide . . . . Over the years, several million homes, many of them mobile homes, were built with polybutylene plumbing systems. Before long, the plumbing systems began to experience failures of the fittings and of the pipe itself.” (Hensler et al., Class Action Dilemmas: Pursuing Public Goals for Private Gain (2000) p. 375, fn. omitted (RAND).) Polybutylene piping was the classic defective product. It seemed like a good thing when introduced, but over time it literally failed to hold water. Chlorine in drinking water apparently caused it eventually to crack and leak, leading to nationwide litigation and massive settlements involving some 220,000 replumbing jobs and many millions of dollars. (RAND, supra, at p. 390.) In 1977, Shell Oil Company had begun manufacturing polybutylene resin—the raw material for the pipes—and it was

12 the sole manufacturer of polybutylene resin. But Shell entirely withdrew the product from the U.S. market in 1996. (RAND, supra, at p. 375.) The Bay Summit decision was one chapter in the nationwide litigation that drove Shell out of the polybutylene market. In Bay Summit, the case’s record was very limited and in conflict as to (1) whether Shell did in fact control the manufacturing or distribution process and (2) whether it was an integral factor in bringing a polybutylene plumbing system to market. (Bay Summit, supra, 51 Cal.App.4th at p. 778.) The court remanded the case for a factual determination of whether Shell “had control over, or a substantial ability to influence, the manufacturing or distribution process.” (Ibid.) Absent this ability to control or influence the manufacturing or distribution system, Shell could not be strictly liable for leaking polybutylene plumbing systems. Bay Summit is a disastrous precedent for Amazon because, unlike in Bay Summit, there is no doubt about Amazon’s ability to control the distribution system Amazon invented. Amazon is the distribution system. It thus should be strictly liable for defective products people buy from its site. Beyond the particular Bay Summit decision, the entire polybutylene saga is an example of how tort law can lead to socially rational decisionmaking by forcing commercial actors to internalize the costs of their actions. According to the RAND analysis, the final outcome was that polybutylene piping disappeared from the U.S. marketplace because it was faulty. Apparently it was cheaper to switch to different kinds of plastic than to bear the tort liability for polybutylene. If the product is

13 defective, the sooner it is fixed or disappears, the better. This saga favors Loomis, not Amazon. These six cases show courts have used cost-benefit analysis to decide whether to impose tort duties. Six cases on the other side of the coin show the same method. B Courts refrain from imposing tort duties where measures to avoid accidents are burdensome and unlikely to be cost-effective. This notion initially may seem off-putting: do we not seek to avoid accidents at all costs? No, we do not. As Calabresi explained, “[w]e take planes and cars rather than safer, slower means of travel.” (Calabresi, supra, at p. 18.) At some point, a quest for perfect safety becomes irrationally expensive. Six cases show how judicial cost-benefit analysis can locate this point. 1 The previously cited Berkovitz decision illustrates this judicial instinct. It used the cost-benefit method long before that terminology appeared in tort decisions. The Berkovitz accident was in 1919. Two couples in a “Dodge touring car” were breaking the speed limit at 2:00 a.m. in a city and rear-ended a gravel truck going about 10 miles per hour. (Berkovitz, supra, 191 Cal. at pp. 197–198.) People in the Dodge sued the truck company, claiming the truck’s coal-oil tail lamp was out. If true, that would have violated the law requiring a night light and thus would have been negligence per se. The evidence conflicted about the coal-oil lamp. The truck driver said he saw the reflection of the tail light three or four blocks before the accident scene and it was working then, so by his account the light must have failed in those three or four blocks—just moments before the crash. (Id. at p. 198.) The Supreme Court

14 ruled that, under these circumstances, the truck company was entitled to an excuse instruction. (Id. at pp. 199–200.) The Berkovitz court permitted the excuse defense because it, in essence, decided the conduct the trucker described was efficient. The trucker claimed he made sure his light was working just before the accident. There was no other better and yet practical way for the trucker to check his tail light: in 1919, dashboard alerts about failed tail lights were things of the future. In that year, the only way the trucker could have been more cautious would have been to post a person “over the rear light to observe whether it is constantly burning.” (Id. at p. 199.) This lookout notion was obviously preposterous in the court’s view. This extreme measure would have been socially uneconomical—like forcing all interstate traffic today to drive at 10 miles per hour to minimize accidents. That much safety would be socially irrational. Because the precaution cost would have been high while the accident risk from a very-briefly-unlit tail light was low, the Berkovitz decision refused to impose an absolute per se duty on the trucker. The cost of the additional precaution was socially irrational: posting lookouts on the back of every motor vehicle would have been more costly than the safety benefit would justify. 2 A second example is O’Neil v. Crane Co. (2012) 53 Cal.4th 335 (O’Neil). Amazon relies on O’Neil, but that decision contradicts Amazon’s argument. The key feature about the O’Neil case was that defendant Crane made valves but was being sued for replacement asbestos gaskets Crane did not make. When workers refurbish an old

15 leaky valve, they predictably put new replacement gaskets in the valve. But the O’Neil decision held valve maker Crane owed no tort duty, either in negligence or strict liability, to people repairing its valves who breathed dust from the replacement asbestos gaskets Crane did not make. (O’Neil, supra, 53 Cal.4th at p. 342.) Crane could foresee workers repairing its valves would use asbestos gaskets made by other companies. (O’Neil, supra, 53 Cal.4th at p. 342.) So why did Crane have no tort duty? Because the Supreme Court said the cost of imposing this duty on one firm to warn of dangers from products from other manufacturers would outweigh the benefit of resulting injury reductions. This duty would be tremendously and ineffectively overbroad: like requiring “match manufacturers to warn about the dangers of igniting dynamite.” (Id. at p. 361.) But when every firm must warn everybody about everything, the costly exercise does no good. “ ‘To warn of all potential dangers would warn of nothing.’ ” (Id. at p. 363, quoting Andre v. Union Tank Car Co., Inc. (1985) 213 N.J. Super. 51, 67 [516 A.2d 277, 286].) O’Neil’s cost-benefit analysis meant the court refused to impose a tort duty where it would be socially irrational to do so. Crane did not control the marketing of gaskets made by other firms. Amazon, by contrast, completely controls its river. There is nothing socially irrational or ineffectively redundant about making Amazon strictly liable for accidents from products bought from its website. 3 The same principle governed Peterson v. Superior Court (1995) 10 Cal.4th 1185, 1188–1189 (Peterson), which ruled land and hotel owners are not strictly liable for injuries to their

16 tenants and guests caused by an unknown defect in the premises. The owners remained liable in negligence; injured guests retained a strict liability action against the manufacturer, distributors, and retailers of the allegedly defective products. But to extend the strict liability duty to land owners for unknown defects owners did not create and could not discover through a reasonable inspection would mean imposing costly duties without a compensating benefit in accident reduction. “A landlord or hotel owner, unlike a retailer, often cannot exert pressure upon the manufacturer to make the product safe and cannot share with the manufacturer the costs of insuring the safety of the tenant, because a landlord or hotel owner generally has no ‘continuing business relationship’ with the manufacturer of the defective product. . . . ‘The cost of insuring risk will not be distributed along the chain of commerce but will probably be absorbed by tenants who will pay increased rents.’ ” (Id. at p. 1199.) “Because of the practical impossibility of obtaining expert knowledge of all the components of an apartment, landlords must rely on others for their safe manufacture, installation and repair. In this respect, landlords are in no better position to know of defects than are tenants.” (Id. at p. 1209.) By contrast, Amazon is in a better position than its customers to learn of and to combat defects in products on its website. 4 Parsons v. Crown Disposal Co. (1997) 15 Cal.4th 456 (Parsons) is conceptually similar to O’Neil and Peterson. Darrell Parsons was riding his horse on an urban bridle path when a truck noisily lifted a nearby trash bin. The crashing sound made the horse bolt, throwing Parsons to the ground. (Id. at p. 462.)

17 Parsons sued the trash company. The Supreme Court ruled the company owed Parsons no tort duty. The Supreme Court used a “social utility analysis” to evaluate accident avoidance measures the trash company could have taken: “changing the hours of collection, temporarily ‘blocking off’ the area with warning cones or tape, posting warning signs, providing riders with a schedule of collection times, or a combination of these methods.” (Id. at p. 474.) The court rejected Parsons’s proposals because they would increase “the burden on machine operators over what was considered reasonable.” (Ibid.) These precautions “unreasonably would impair the utility” of the trash company, which ran a business “of high social utility.” (Ibid.) And imposing these duties on the trash company would imply similar restrictions on a wide “range of socially useful activities that may produce such noises and provoke such fright.” (Id. at pp. 474–475.) The Supreme Court illustrated its cost-benefit analysis with commonsense examples: “Should a homeowner whose property abuts this extensive bridle path (or who has horse-owning neighbors) be obligated to peek over his or her six-foot fence to make sure that no horse is near before starting a power lawn mower? Should he or she be required somehow to keep a constant lookout while mowing the lawn, lest he or she frighten a horse that approaches shortly after lawn mowing begins? Should a homeowner or building contractor be obligated to undertake similar procedures before and during use of chain saws, leaf blowers, or other loud power tools? What about noise from passing cars and trucks on the adjacent highways, or from a picnicker’s radio, or from an emergency siren or alarm, or from a jetliner’s sonic boom? We conclude that imposing a duty in the present case to guard against fright to a

18 horse might well subject all manner of actors to the same duty and potential liability, with obvious and detrimental consequences stifling to the community.” (Parsons, supra, 15 Cal.4th at p. 475, emphasis in original.) In short, this seat-of-the-pants cost-benefit analysis showed the accident avoidance Parsons had proposed for the trash company was not worth its cost. These proposed safety measures were just as obviously inefficient as lookouts on 1919 trucks to see if the coal-oil lamp was still lit. (Cf. Grady, The American Negligence Rule (2019) 53 Val.U. L.Rev. 545, 576 [Parsons basically held the untaken precautions “flunked the Learned Hand formula”].) When accident avoidance will be costly and ineffective, the Supreme Court refuses to burden defendants with tort duties. Amazon loses under this rule, for it can take cost-effective actions to minimize the cost of accidents. Indeed, it claims it already does. The same rule explains the lower court California cases on which Amazon relies. These Court of Appeal decisions concern two situations: secondhand dealers and finance lessors. We treat each in turn. 5 The secondhand dealer cases declined to impose a duty in strict liability because the dealers were unable to take cost- effective measures to minimize accidents from the secondhand machines they sold. Wilkinson v. Hicks (1981) 126 Cal.App.3d 515 (Wilkinson) is representative. Robert Wilkinson injured his hand operating a 50-year-old Niagara punch press at the Mac Smith Company. Benmatt Industries owned the press for many years. Then

19 defendant Roy Hicks, a dealer in secondhand industrial machinery, bought it from Benmatt and sold it “as is” to Mac Smith Company. Wilkinson sued Hicks, but the trial court disallowed the strict liability claim. Only the negligence action went to the jury, which ruled against plaintiff Wilkinson. (Id. at pp. 516–519.) On appeal, the court affirmed secondhand dealer Hicks owed no strict liability duty to Wilkinson. Imposing strict liability on secondhand dealers would require them “routinely to dismantle, inspect for latent defects, and repair or recondition their products.” (Wilkinson, supra, 126 Cal.App.3d at p. 521.) That rule “would effect a radical change in the nature of the used product market, which would deprive that market of [its] desirable flexibility.” (Ibid.) The proposed rule would make used goods dealers the insurers of the goods they sold. (Ibid.) This would increase the prices of secondhand goods. (Id. at p. 520.) In a nutshell, plaintiff Wilkinson proposed requiring secondhand dealers to take expensive safety measures: dismantling, inspecting, repairing, and granting a mandatory warranty. The court, quite reasonably, was not convinced these steps were cost-effective in light of the existing strict liability duties on the original manufacturer and its original distribution network. Wilkinson’s proposed safety measures were too expensive and ineffective to be socially desirable. So Hicks owed no strict liability duty to Wilkinson. (Accord, Brejcha v. Wilson Machinery, Inc. (1984) 160 Cal.App.3d 630; Tauber–Arons Auctioneers Co. v. Superior Court (1980) 101 Cal.App.3d 268.) In contrast, the measures Amazon can take to minimize the cost of accidents are cost-effective and socially efficient.

20 6 The same principle guided the finance lessor decision in Arriaga v. CitiCapital Commercial Corp. (2008) 167 Cal.App.4th 1527. Guillermo Arriaga suffered a work injury while operating a glue spreader machine. Black Bros. manufactured the glue spreader and sold it to Klors, which resold it to CitiCapital (actually, its predecessor, which we ignore), which leased it to AVP, which sold it to Orepak, which employed Arriaga. (Id. at pp. 1532–1533.) The injured Arriaga sued many parties, including CitiCapital. The court analogized CitiCapital’s role to that of a bank that loaned purchase money to AVP. (Id. at p. 1536.) The court refused to put a duty of strict liability on CitiCapital in part because the court could not see what CitiCapital might do to “enhance product safety.” (Id. at p. 1537.) “[T]he finance lessor is not in any position to either directly or indirectly exert pressure on the manufacturer to enhance the safety of the product.” (Id. at p. 1538; see also ibid. [“the finance lessor does not have control over, or a substantial ability to influence, the manufacturing or distribution process”]; id. at 1539 [“the finance lessor’s relationship with a particular manufacturer does not, in the normal course, possess the continuity of transactions that would provide a basis for indirect influence over the condition and safety of the product”].) In short, there was no action, let alone any cost-effective action, the finance lessor could take to reduce the likelihood of glue spreader accidents. The Arriaga court thus refused to impose tort duties on the finance lessor. The contrast with Amazon’s situation is obvious. *******

21 Amazon’s citations thus offer it no support. All involve defendants who, unlike Amazon, had no cost-effective way to reduce the costs of accidents. This case is easy. Amazon is well situated to take cost- effective measures to minimize the social costs of accidents. Strict liability will prompt this beneficial conduct. Loomis wins this appeal. The case will return to the trial court for resolution of issues the appeal has not addressed.

WILEY, J.

22