PRODUCT LIABILITY AND ONLINE MARKETPLACES: COMPARISON AND REFORM

Abstract

This article analyses the challenges that online marketplaces and e-commerce pose to traditional product liability doctrines. It uses a comparative perspective to examine whether an online platform can be liable to a consumer for a defective product purchased on its platform, and the adaption of product liability law to this challenge in a series of jurisdictions. It reflects on the role of litigation and regulation, focusing on Europe and the United States, and considers reform in a number of jurisdictions in this area. It concludes with proposals for increasing the accountability of online marketplaces for products sold on their websites.

I. INTRODUCTION

Technological innovation has long posed challenges to the law, and the sphere of liability law is no exception. This is particularly the case in the specific area of product liability where there has been a lively doctrinal debate about how to take account of new technology and innovation.¹ Challenges generated by modern supply chains have also confronted the law, and in recent times this has been most clearly encapsulated by the thorny issue of the liability of online marketplaces for loss caused by defective products. Recent research has illustrated that non-compliant products are extensively commercialised on such websites,² which evidently exposes society to increased safety risks, and poses the question of how the law should be adjusted in order to tackle this significant public safety issue.

This article sets out to analyse from a comparative perspective the challenges that online marketplaces and e-commerce pose to traditional legal doctrines relating to product liability, by examining the question of whether an online platform can be liable to a consumer for a defective product purchased on its platform. After looking briefly at the historical development of product liability and how that has been affected by product innovation, a comparative examination of how product liability law has been adapted to take account of the role of online marketplaces in a series of jurisdictions will be undertaken. The role played in this sphere by litigation and regulation, respectively, will then be reflected upon, contrasting the different approaches in Europe and the United States (USA). Finally, the reform efforts in this area will be considered, and a series of proposals will be made in favour of increasing the accountability of online marketplaces for products sold on their websites.

In undertaking the comparative law analysis, a range of different civil law and common law jurisdictions will be examined, focusing on the most prominent product liability schemes across the globe, contrasting and analysing the difference in approach adopted to challenges of e-commerce.

II. TRADITIONAL APPROACHES OF PRODUCT LIABILITY TO TECHNOLOGICAL CHANGE

Product liability stemming from harm caused by defective products has undergone a rapid evolution across the globe. Originally, product liability was in many jurisdictions seen as simply an application of the law of contract or tort. In Europe, it was not until relatively late in the twentieth century that product liability was seen as raising distinct legal issues.³ The development of thinking in favour of reform of this area of the law was prompted by a series of mass product disasters, particularly in the healthcare sphere, which gave rise to discussions about a pan-European harmonised approach to product liability. This culminated in the adoption of the Product Liability Directive in 1985, ushering in a distinctive European approach to product liability based on a no-fault regime providing for liability for harm caused by defective products,⁴ which will be examined in further detail below. Despite some minor amendments of the Directive, and a formal review process integrated into the Directive, it has been increasingly accepted that the scheme ushered in by the Product Liability Directive has struggled to take account of developments in new technology in particular, including new supply-chain models such as online marketplaces. A revision process of the Directive is therefore underway,⁵ which will be elaborated below, to bring the legislation up to date, with a revised text likely to be formally adopted in early 2024.

The situation of the United Kingdom (UK) is now distinctive. Having implemented the Product Liability Directive into domestic law by virtue of the Consumer Protection Act 1987, which (for the moment) remains unaffected by Brexit, the UK is no longer party to the ongoing revision process of the Directive. Despite some indications made in favour of updating the legislation to take account of technological developments,⁶ the UK authorities have not yet launched a formal review process. The UK is thus in the somewhat paradoxical position of having outdated legislation originating from European Union (EU) instruments which are about to be updated.

The historical evolution of the distinct field of product liability occurred much earlier in the USA. Prior to the early twentieth century, liability for injuries resulting from defective products was governed by contract law.⁷ Damage-causing products could result in liability only as arising out of transactions between injured buyers and immediate vendors—those who stood in a relation of ‘privity’ of contract with one another.⁸ This excluded remote manufacturers in the ordinary case. The fall of privity came in the early twentieth century,⁹ transferring product liability cases from contract law into tort law based on negligence-based liability. This continued for several decades, during which time the sentiment grew among many scholars and jurists that negligence was insufficient.¹⁰ Around the middle of the twentieth century, Justice Roger Traynor of the Supreme Court of California authored an influential concurring opinion arguing that strict liability—not negligence principles—should govern manufacturer liability for defective products.¹¹ Justice Traynor's argument that public policy demands a manufacturer be responsible for a good's quality regardless of negligence was limited. It only applied in cases pertaining to a product's normal and proper use, and only to injuries traceable to a product as it reached the market. This rationale for strict liability was reaffirmed in 1963 in the case of Greenman v Yuba Power Products Inc ¹² where the Supreme Court of California—speaking again through Justice Traynor—held that a manufacturer is strictly liable in tort when a product it places on the market, knowing that it will be used without inspection for defects, proves to have a defect that caused a person injury who was using the product in a way it was intended.

In 1965, the American Law Institute's Restatement of Torts (Second) incorporated a general principle of strict product liability.¹³ The relevant provision—section 402A—embraces a ‘consumer expectations’ test as governing all products.¹⁴ Under this standard, the litmus test for manufacturer liability is the average, ordinary consumer's expectation with respect to a product's harmful characteristics. By contrast, under the negligence-inflected ‘risk-utility’ test articulated by the Supreme Court of California in Barker v Lull Engineering Co,¹⁵ a manufacturer's liability for design defects or for failure to warn consumers about a product's dangerousness is based upon a balancing of a product's utility features and risk features.¹⁶ In 1998, the American Law Institute published the Restatement (Third) of Torts: Products Liability, which set up the tripartite classification of manufacturing defect, design defect and failure to warn. Following the majority position in US courts, only manufacturing defects are governed by strict liability; design defect and failure to warn claims are governed by negligence-inflected risk–utility tests.¹⁷ In modern tort litigation for product liability, the question of whether the consumer expectation test or risk–utility test—or some combination thereof—is applied depends on the jurisdiction.¹⁸ For design defects, the Third Restatement embraced a particular form of the risk–utility test, the ‘reasonable alternative design’ test, according to which a product:

is defective in design when the foreseeable risks of harm posed by the product could have been reduced or avoided by the adoption of a reasonable alternative design by the seller or other distributor, or a predecessor in the commercial chain of distribution, and the omission of the alternative design renders the product not reasonably safe.¹⁹

The Third Restatement's ‘reasonable alternative design’ test has received an uneven reception in American courts.²⁰

As has been noted above, throughout the historical evolution of the law of product liability, the changing nature of commercial practices and the modification of supply chains have resulted in changes in the law.²¹ In recent times, the development of e-commerce has posed new challenges, in particular due to the rise of online marketplaces. Given their unique nature, these platforms pose specific difficulties in terms of liability for defective products sold on their websites. Despite the commonality of the basic service provided, namely connecting sellers of products with consumers via an online marketplace, the specific organisation of the relevant transactions can vary a good deal. The underlying contractual arrangements differ considerably with certain online platforms being the direct contracting party with the consumer and others not. Additionally, there are also notable variations in product dispatch methods: in certain instances they are sent direct by the third-party vendor to the purchaser, while in others they are routed via the platform's fulfilment centre, thereby placing the platform in physical possession of the item. The position of civil liability of these platforms will now be examined.

III. ARE ONLINE MARKETPLACES LIABLE FOR DEFECTIVE PRODUCTS ON THEIR WEBSITES?

There are a variety of different approaches to the potential liability of online marketplaces. Although the advent of e-commerce platforms is not new, the regulatory response has lagged behind changes in the market. Product liability is just one of many challenges posed by the exponential rise in e-commerce, and it has by no means been the priority of policymakers in this area. There are few countries where legislative initiatives have been undertaken, though the issue has started to feature on law reform agendas. This can be seen from the current debate accompanying the revision of the Product Liability Directive in Europe, which will be explored below.

From a comparative law perspective, the following typology of approaches to this issue can be identified: jurisdictions in which there is no liability for online marketplaces; countries where the response is uncertain; and, finally, a small number of countries where liability may be found in particular circumstances.

A. No Liability for Online Marketplaces

An absence of liability for online marketplaces seems to be the predominant approach at present. According to this approach, online platforms or marketplaces are not, for a variety of reasons, subject to the classic rules relating to liability for products sold online which cause injury.

This is the approach in Europe. The original EU Product Liability Directive was designed to be a harmonising influence in light of the diverse provisions across Europe.²² In terms of potential defendants under the Product Liability Directive, the ‘producer’ of the product is the primary entity liable, a concept which is defined principally as the manufacturer of the good.²³ The supplier of a product is conceived only as incurring subsidiary liability,²⁴ though an importer is considered to be in the same position as a producer in the case of import into the Community of a product for sale.²⁵ The Product Liability Directive was drafted in an era well before the modern development of e-commerce. As a result, the approach and concepts found in the Directive do not necessarily adapt well to modern models of supply chains, in particular the rise of e-commerce.

Despite the growing role of platforms, there are important hurdles to considering an online marketplace as a producer. This is likely to be the case even where the online marketplace provides fulfilment services, and even if shipping packaging features the online platform's trademark. Courts might not give weight to such branding on shipping packaging if the product packaging itself clearly identifies its ‘actual’ producer. The one exception to this position is where own-branded products are distributed by online platforms, where liability would potentially arise.²⁶ In Germany, such own-branders are known as ‘quasi-manufacturers’, and whilst, in some cases, online marketplaces do sell products under their own brand, this is not their distinctive feature as they typically facilitate sales by third parties.²⁷ Such an exception is thus only applicable to a small minority of products purchased on platforms.

Might an online marketplace be considered to be an importer under the provisions of the Product Liability Directive? It has been argued that, in certain scenarios, an online platform might be considered an importer under the definition given by the Directive. This could be the case where there have been fulfilment services provided by the platform, as the product will thus have been in the possession of the platform, and could be considered to be ‘a form of distribution’ as envisaged by the Directive, even though the platforms are likely to maintain the traditional position that they are merely hosting sales, rather than being a supplier.²⁸ From this perspective, the relevant legislation in Belgium is very open as it refers to ‘any form of distribution’ which has been interpreted as being satisfied by the mere fact of placing a product at one's disposal even for a very limited period of time.²⁹ On the other hand, in Germany, it has been argued that online marketplaces cannot be considered to be importers under German law.³⁰ Reference is made to the German Product Safety Act, which draws a distinction between fulfilment centres and importers. This could constitute an argument against the liability of online marketplaces, as under the German legislation, providers of fulfilment services are not to be considered as importers, but rather a separate category.

Where no fulfilment services have been undertaken, it is even more difficult to see how the platform could be characterised as an importer under the Product Liability Directive. In the absence of fulfilment services being provided in respect of the product, the online platform would have no involvement in the transporting of the physical product into the EU.

A final question is that of subsidiary or ‘ancillary’ defendants, such as suppliers. Under the Product Liability Directive, a supplier may be found liable where neither the producer nor the importer can be identified and the supplier fails to provide the relevant details within a reasonable time period.³¹ The problem with this provision is that it is dependent upon the online platform being characterised as a ‘supplier’, which is an uncertain premise, and one, moreover, that the platforms contest. In any case, even if they were so characterised, it would be relatively simple for them to escape liability by identifying the manufacturer, who will in many cases be based abroad, and often in the case of products on online platforms, on the other side of the world, making a product liability claim not realistic or viable.

In this context, it should also be mentioned that the recently adopted Digital Services Act (DSA)³² does not contain an explicit provision on product liability of online marketplaces. In particular, it is not likely that the liability exemption regulated in Article 6 of the DSA will provide a shield against product liability claims. The provision stipulates that hosting providers under certain conditions ‘shall not be liable for the information stored’ at the request of a third party.³³ In other words, the safe harbour provision of the DSA only provides a conditional exemption from liability related to information stored online, but not from liability related to their involvement in the physical distribution of dangerous products that takes place offline.³⁴

However, the DSA does not permit online marketplaces to remain (legally) indifferent to the quality and safety of products sold by third-party traders. Specifically, the particular measure obligates very large online platforms as well as platforms operated by enterprises other than micro and small enterprises³⁵ to require that traders have obtained ‘a self-certification by the trader committing to only offer products or services that comply with the applicable rules of Union law’³⁶ before allowing them to use their services to sell products. Of course, this is (very) different from imposing product liability on relevant platforms. Moreover, the DSA, in particular Article 30, stops short of obligating platforms to allow traders to use their services only if they also provide proof of having secured product liability insurance.

In sum, the current EU regime found in the Product Liability Directive is not well adapted to modern supply chains and e-commerce in particular. This position has been explained by Professor Busch as deriving from the fact that the Directive in question ‘harks back to the pre-Internet era when supply chains where mainly organized as “pipelines” involving importers, wholesalers and retailers’.³⁷ What this means in practice is that online platforms are generally not liable in case of defective products bought on their marketplaces. Given the prevalence of products manufactured out of jurisdiction on these platforms, and the evident difficulties of identifying and then pursuing a manufacturer on the other side of the world, consumers are likely to be left without a remedy in such scenarios. That is a problematic position given the important and growing market-share of such platforms. The severity of the situation is underscored by recent research indicating that product safety is a notable concern across many of these platforms.³⁸

A French decision from 2020 illustrates this phenomenon. It arose from a defective phone charger bought on a marketplace run by a French company. The product exploded and caused injury to the claimant. In a claim brought inter alia against the online marketplace based upon the implemented Product Liability Directive, the Paris Court of Appeal held that the claimant had failed to demonstrate that the online marketplace had acted as either the seller, producer or supplier of the product in question, making reference to the website's terms and conditions as indicating that the latter was acting solely as an intermediary. The fact that the platform had received the payment for the product did not change the court's position as the purchase order stated expressly that the platform was acting on behalf of the identified third-party seller.³⁹

B. Uncertainty as to Current Position

In other countries, the legal position remains uncertain given that novel legal issues are involved and there is a lack of case law on point. Canada is a prime example. As there is no strict liability regime for defective goods in Canada, claimants have to rely on shoe-horning any claim against online marketplaces into orthodox common law concepts such as duty of care.⁴⁰ It is uncertain whether any Canadian court would find a sufficiently proximate relationship for a duty of care to arise.

There is also uncertainty in other systems as to whether traditional legal concepts could be adapted to such factual scenarios. In respect of France, the possibility of the French Consumer Code applying to these cases has been mooted, as the provisions obligate producers and distributors to take ‘all useful measures’ to ‘ensure compliance with all safety obligations’.⁴¹ Whilst there is a very broad definition given of a distributor,⁴² there may be difficulties in establishing that the platform is really acting ‘in the marketing chain’ given that the traditional defence is that the latter is only an intermediary, bringing together sellers and buyers without determining the price.

Other potential routes to liability have been discussed. One of those is organisational liability of the online platforms. This issue has been raised by commentators including Professor Büyüksagis in a recent article in the Journal of European Tort Law,⁴³ and will be examined further below. A similar approach has been mooted in Japan, where liability for damage occurring from defective products can be based on contract or tort law under the Civil Code, as well as on a strict liability regime contained in the Japanese Product Liability Act. In Japan, it has been mooted that online platforms could be subject to ‘systemic liability’, according to which they would bear tort liability if they fail to provide an online platform system which prevents damage from occurring to the users.⁴⁴

C. Tentative Finding of Liability

Two jurisdictions in which there have been developments in favour of platform liability are the USA and Portugal. This has also been suggested during debate of the reforms proposed at EU level.

The USA has been in the vanguard of developments in this area. Recent US case law suggests an increasing willingness on the part of courts to hold e-commerce companies liable for defective goods sold on digital marketplaces. Before 2019, American courts uniformly dismissed product liability suits filed against Amazon by consumers injured by products purchased from third-party vendors on Amazon.com.⁴⁵ In most instances, the primary impediment to plaintiffs’ claims was Amazon's ability to skirt classification as a ‘seller’ or ‘distributor’ of goods—as opposed to an intermediary auctioneer of sorts from which legal title to a particular product is not transferred to a buyer upon its purchase—for purposes of state law-based strict product liability rules.⁴⁶ Courts’ rationales for locating Amazon outside the ‘chain of distribution’⁴⁷ often emphasised, in addition to the non-transference of title, Amazon's insufficient exercise of control over the third-party products sold on its website.⁴⁸

But the tide turned in Oberdorf v Amazon.com Inc,⁴⁹ where the US Court of Appeals for the Third Circuit held that Amazon was a ‘seller’ subject to a state strict product liability statute based on inter alia the company's being the ‘only member of the marketing chain available to the injured plaintiff for redress’, ‘in a better position than the consumer to prevent the circulation of defective products’ and readily in a position to ‘distribute the cost of compensation for injuries resulting from defects by charging for it in [their] business’.⁵⁰ However, this decision was subsequently vacated and Amazon's petition for rehearing en banc was granted, whereupon the Third Circuit certified to the Pennsylvania Supreme Court the question of whether Amazon is a ‘seller’ under Pennsylvania law.⁵¹ The case was settled before the state's high court could answer the question.

In 2020, Bolger v Amazon.com LLC ⁵² was decided, a case in which a California state intermediate appellate court reversed a lower court's grant of summary judgment in favour of Amazon in a product liability suit, finding dispositive a number of factors demonstrating Amazon's ‘capacity to exert its influence on third party sellers to enhance product safety’.⁵³ And in 2021, Loomis v Amazon.com LLC ⁵⁴ continued the trend. Here the same California state appellate court affirmed its previous holding in Bolger as ‘correctly decided’⁵⁵ and determined that e-commerce vendors such as Amazon may be held liable under a theory of strict liability for defective products purchased by consumers from third-party sellers. Though the precise rationales underpinning these decisions differ in some respects, they share the common feature of affirming e-commerce companies’ susceptibility to product liability suits based on their ability to anticipate and minimise downstream injuries occasioned by products sold on their online marketplaces.⁵⁶

While not every jurisdiction has adopted the approach endorsed in Bolger and Loomis,⁵⁷ a recent administrative enforcement action filed by an independent federal regulatory agency—the Consumer Products Safety Commission—relied on a similar rationale by determining that Amazon constitutes a ‘distributor’ subject to its enforcement authority under its organic statute, in part, based on the significant role that Amazon plays in overseeing the ‘holding, storing, sorting, and preparing for shipment’ of potentially hazardous goods.⁵⁸

In sum, while the status of e-commerce companies as subject to state strict product liability law remains in development in the USA, there is an identifiable trajectory, in several jurisdictions, that favours the imposition of liability on companies like Amazon as sellers of defective products.⁵⁹ In much the same way that a societal transition occurred with respect to the production of goods through non-automated means, a more recent shift away from consumers’ use of brick-and-mortar stores and toward reliance upon e-commerce has taken place, giving rise to the discussion as to the proper place for laying liability when products sold through online marketplaces cause injury to consumers.⁶⁰

It should in addition be noted that, in the USA, there have also been some regulatory and legislative developments at the state level. In California, two proposed legislative interventions, had they been successful, would have extended strict product liability to online marketplaces.⁶¹ From a regulatory perspective, in 2021, the US Consumer Product Safety Commission (CPSC) filed an administrative complaint against Amazon.com in order to obligate Amazon.com to accept responsibility for the recall of unsafe products sold on its website on the grounds that it is a distributor of products in the USA.⁶² The process is ongoing.⁶³

In Portugal, there is an innovative regime concerning the liability of platforms for the lack of delivery or conformity of goods, digital content or digital services. Whilst this does not relate to the Product Liability Directive specifically, involving instead direct liability of platforms under Directives 2019/770 and 2019/771, it is a unique development in Europe.⁶⁴

A recent development at the EU level, which should be highlighted, is the revised Product Liability Directive which was finalised in December 2023.⁶⁵ This text, which involves a substantial and wide-ranging reform of the Directive, includes some innovative provisions relating to online platforms. Article 7 of the revised Directive refers to ‘Economic operators liable for defective products’ and lays down that when an economic operator in the EU cannot be identified, then ‘any provider of an online platform that allows consumers to conclude distance contracts with traders and that is not an economic operator’ may be liable provided that the aforementioned carve-out from the immunity of hosting services providers found in Article 6(3) of the DSA is satisfied. This proposal will be examined in greater detail below, but it is a significant development in the sense that it does allow for the potential liability of online marketplaces in product liability cases.

IV. REGULATORY MODELS

In undertaking comparative law analysis, it is important to account for contextual factors which can make an impact on, or influence, the solutions adopted in various jurisdictions. One such factor is the policy choice made by individual systems in adopting regulatory or litigation solutions for resolving questions of public policy. In this section, the respective roles of litigation and regulation in this sphere will be reflected upon from a macro perspective, contrasting European and US approaches, before the use of other measures such as informal or self-regulatory solutions to provide remedies is analysed.

A. Regulation versus Litigation

From a global perspective, there are two distinct approaches to consumer protection, either following the regulatory model or the litigation model of dispute resolution. Neither of these approaches need necessarily be mutually exclusive, and indeed joined-up policy responses should ideally be made up of a combination of both regulation and remedies available through the courts. In practice, however, there has been a tendency for individual legal systems to opt for either a predominantly regulatory or litigation-based approach. What are the relative advantages of each model?

On the one hand, the regulatory-based model places governmental intervention at the centre of economic oversight by means of controls on entry to market, prevention, monitoring and post-market supervision. The advantage of such an approach is that it results in power being harnessed by the State in terms of supervision and enforcement. It is also preventive in the sense that there is an ability to project into the future to mitigate impact of market dominance or emerging risks. This contrasts with the essentially reactive approach of litigation, which involves an ex post facto analysis. Regulatory action at a European level could provide a harmonised approach across the various Member States, thereby creating greater certainty for business.

The regulatory model does, however, require the deployment of significant public funds to work effectively, and in rapidly moving areas of the economy such as the technology sector, regulators will need to be nimble on their feet in order to keep up with swift developments. Experience has shown that this can sometimes be challenging—as evidenced, for instance, by the regulatory gap in respect of online marketplaces.

Further, regulation is not equipped to deal with the issue of compensation for harm caused to individual market participants. Even if financial penalties such as fines may be available as part of the tools of regulatory action, this does not satisfy the restitutio in integrum principle for those who have suffered harm.⁶⁶ As a consequence, the issue of potential liability will inevitably be raised.

The litigation model is premised on a less State-centric and more decentralised approach. Whilst the primary objective of litigation arising from defective products is generally to gain compensation for consumers by means of damages payments, the litigation model can also encapsulate the desire of influencing market behaviour by means of a deterrence model, whereby ex post facto litigation is expected to influence the way market actors operate. In the products space, this entails business, platforms and their insurers being given an incentive to monitor the safety of products, thereby deploying their considerable expertise and resources to improve product safety on online platforms.⁶⁷

In common law jurisdictions it is possible for courts to develop the law to take account of evolving commercial realities, although the view may be taken that matters of consumer protection should be left to the legislature. The US case law examined above illustrates such a dynamic approach to the rise of e-commerce. This also has the advantage of meeting the expectations of consumers that platforms should exercise a degree of control over third parties selling products via their platforms, particularly given the challenge of pursuing traders established in far-flung parts of the globe who are nonetheless active on online marketplaces and shipping direct to customers. Over and above the pre-existing relationship between the parties, given that the platforms act as a gateway to the European market generally, they have the necessary market power to influence traders’ behaviour so that the requiring of appropriate insurance in respect of claims might ensure that the potential liability of the platform is channelled back to the contracting parties active on the marketplace.

However, in many jurisdictions, litigation remains an extremely onerous option in terms of resources and time, and in business-to-consumer cases, the asymmetry of the parties is a strong disincentive on the bringing of claims by consumers. While the development of collective redress procedures such as European Directive 2020/1828 on representative actions have been introduced to try to rectify that position, it is the case that few if any (published) products cases have been brought against online platforms.

Regulation and liability are not the only approaches available. Self-regulation has been adopted at a European level as a flexible way to improve the overall position. There are also other policy options, as will be discussed below.

B. Soft-Law Initiatives in Respect of Online Marketplaces

In the shadow of the law, soft-law initiatives have been launched in this sphere. The clearest example is the 2018 EU Product Safety Pledge, which is a voluntary scheme for online platforms facilitated by the European Commission with the objective of increasing the safety of products sold online by third-party sellers through online marketplaces.⁶⁸ Increasing numbers of platforms have signed up to the Product Safety Pledge over time. The initiative includes a commitment to follow 12 specific actions, including: (i) consulting information on recalled/dangerous products available on the EU Rapid Alert System for Non-Food Products (RAPEX) and other sources and taking appropriate action in relation to the safety of products once identified; (ii) cooperating with Member State authorities in identifying, as far as possible, the supply chain of dangerous products; and (iii) having internal mechanisms in place for notice and ‘take-down’ procedures for dangerous products.

Importantly, in March 2023, the Product Safety Pledge was superseded by the Product Safety Pledge Plus, which is an extended commitment framework consisting of 20 actions where signatories make commitments regarding products marketed on their interfaces. These actions include ‘allow[ing] access to their interfaces for the online web-crawling tools operated by the national market surveillance authorities and/or the Commission to identify dangerous products’⁶⁹ as well as a modernised reporting mechanism. All 11 online marketplaces which had signed the original Pledge⁷⁰ have also signed the extended Pledge Plus, thereby reconfirming their commitment to cooperate with EU national authorities and take measures towards reducing product safety risks for consumers. Signatories started applying the new framework as of 1 December 2023. It is also worth noting that in November 2023 the Pledge Plus was integrated into the newest Consumer Protection Pledge⁷¹ which extends to voluntary consumer protection commitments in the online environment other than product safety.

The Product Safety Pledge was not, however, legally binding and thus did not create any liability or rights, and the Product Safety Pledge Plus did not introduce any changes in that regard. It therefore has limited relevance to product liability-related issues, other than as a preventive tool for encouraging unsafe products to be removed before any harm is caused to consumers. Views about the practical impact of the Pledge are mixed. In various European Commission progress reports, the Pledge is said to have given rise to tangible results.⁷² The recent UK Product Safety Review reported, however, more mixed reactions, noting that ‘[a] significant majority of stakeholder responses were sceptical of voluntary agreements in the sector’.⁷³ Moreover, in that Review, it was noted that ‘[e]nforcement authorities noted that the signatories varied in their approach to their pledged responsibilities and how efficiently they tackled referrals of unsafe products’.⁷⁴

It should be noted that some voluntary measures in the Product Safety Pledge and the Product Safety Pledge Plus will soon be reclassified as mandatory, since they are now included in the General Product Safety Regulation which entered into force in June 2023,⁷⁵ and which replaces the previous General Product Safety Directive.⁷⁶

Beyond Europe, Australia has also adopted a Product Safety Pledge, inspired by the EU Pledge.⁷⁷ As in the EU, this is a voluntary initiative committing signatories to product safety-related responsibilities and reporting. A 2022 Review by the regulator suggests that platforms who have signed the pledge are responsive to regulator requests to review unsafe products.⁷⁸ It reports that platforms are also proactive in identifying unsafe products, including through the use of technology such as ‘machine learning, self-updating algorithms, and image recognition tools’ to detect unsafe products.⁷⁹ However, it remains unclear what impact these measures are having as compared to the volume of unsafe products being offered through e-commerce methods. Similarly to its EU counterpart, the Australian Pledge does not deal with liability issues which fall outside its scope.

Japan followed a similar path in 2023. The Japanese Product Safety Pledge,⁸⁰ based on the Organization for Economic Cooperation and Development ‘Communiqué on Product Safety Pledges’, was developed by related ministries and agencies and operators of major online marketplaces. It is a product safety pledge aimed at protecting consumers from the risks to life and health posed by recalled and unsafe products listed and sold on online marketplaces. Being a ‘voluntary initiative of public–private collaboration’ that goes beyond legal frameworks, it has been signed (as of February 2024) by seven online marketplace operators.

There has also been voluntary action undertaken in the USA by the online platform Amazon in which the latter has agreed to cover up to $1000 of injuries or property damage caused by products sold by third-party vendors on their platform.⁸¹ The scheme is accompanied by mandatory insurance undertaken by the third-party vendors. It does not seem to be available in any other jurisdiction and seems to have been launched in the USA in parallel with the developing US case law extending liability in some circumstances to platforms.

Aside from these specific initiatives, other remedies to consumers seem relatively limited. Online and alternative dispute resolution (ODR and ADR) schemes are generally restricted to contractual claims and are therefore not available for product liability cases. Credit card guarantees may in some circumstances be applicable,⁸² but those are necessarily limited in scope and there may be exclusions.

V. REFORMING THE CURRENT APPROACH

It has been observed above that despite the exponential increase in e-commerce, very few jurisdictions provide for effective remedies against online platforms for harm caused by products sold on their websites. It may perhaps be seen as incongruous that online marketplaces escape liability despite playing an increasingly significant role in the supply chain, and evidence suggesting that they offer non-compliant products for sale.⁸³ Remedies for consumers are thus very limited. Bringing a claim against the original producer of the item in question, which is often based in a different jurisdiction, is generally neither feasible nor cost-effective. It is somewhat surprising that such a gap in the law has arisen despite e-commerce being by no means a new phenomenon. The exception is, as has been seen, the USA, where there is an emerging jurisprudence in favour of strict product liability claims against online platforms for harms due to unsafe products.⁸⁴ It is, however, restricted to certain US states and has only been applied in a handful of cases.

In most jurisdictions, little recourse is available. Europe is one example. Despite having one of the most advanced systems of consumer protection in the world, few remedies are currently available against online platforms in most European countries, other than contractual actions (explored further below). This has attracted a good deal of criticism from observers,⁸⁵ and has prompted reform proposals to be made.

In addition to important academic contributions,⁸⁶ one of the most prominent proposals is that of the European Law Institute (ELI). After having published an influential paper setting out the ‘Guiding Principles for Updating the Product Liability Directive for the Digital Age’,⁸⁷ in July 2022 the ELI released their Draft of a Revised Product Liability Directive.⁸⁸

Article 8 of the ELI Draft concerns ‘Liable Economic Operators’ and sets out what is presented as a ‘cascade scheme for the liability of the economic operators’. Over and above the primary defendants in product liability cases, Article 8 extends potential liability to online marketplaces in cases where ‘neither an authorised representative nor an importer exists in the Union or cannot be identified’. In such a case, liability may arise in the following circumstances pursuant to Article 8(3):

where there exists an online marketplace where such an online marketplace presents the product or otherwise enables the specific transaction at issue in a way that would lead a consumer to believe that the product that is the object of the transaction is provided either by the online marketplace itself or by a trader who is acting under its authority or control, the online marketplace will be deemed an economic operator which has enabled the making available of the product on the Union market, and shall also be liable.⁸⁹

This ‘novel’ provision is described in the accompanying commentary as representing ‘a significant innovation as it addresses the relevant role of online marketplaces in making products available on the market’.⁹⁰

The key aspect of the provision is whether the presentation of the product/transaction leads the consumer to believe that the product is provided by the marketplace or by a trader under their control. Whilst this approach is said to be aligned with the carve-out from the immunity of hosting services providers (such as online marketplaces) found in Article 6(3) of the DSA, there are grounds to doubt whether it is well adapted to cases of product liability. The concrete problem is that it seems relatively easy for the platform in question to bypass liability by means of a disclaimer or other form of words to make it clear that the transaction is not undertaken by the online platform or a trader under its control. This would defeat the object and allow immunity of such platforms to be achieved simply. Indeed, this approach seems to be very different to that encapsulated in an earlier ELI Model Rules on Online Platforms project.⁹¹

The ELI proposal, nonetheless, does have the merit of placing the position of online platforms squarely within the public realm. Not surprisingly, therefore, the issue of online platforms has thus featured during the current reform process of the Product Liability Directive by the European Commission. Whilst the Commission did not initially show a great degree of enthusiasm for grasping this particular nettle,⁹² there was a degree of surprise when, in September 2022, the Commission published a proposal for a revised Product Liability Directive which tackled the issue head on.

Online platforms feature in Article 7 of the revised Directive, which concerns ‘Economic operators liable for defective products’. Under Article 7, when an economic operator in the EU (including a manufacturer) cannot be identified, then ‘any provider of an online platform that allows consumers to conclude distance contracts with traders’ may be liable provided that the carve-out from the immunity of hosting services providers found in Article 6(3) of the DSA is satisfied. This entails in concrete terms that the online platform will need to show it has presented the specific product or otherwise enables the specific transaction ‘in a way that would lead an average consumer to believe that the information, or the product or service that is the object of the transaction, is provided either by the online platform itself or by a recipient of the service who is acting under its authority or control’. This approach, however, is vulnerable to the criticism made above in respect of the ELI proposal: it will in practice be relatively easy for the platforms to bypass liability by means of a disclaimer.

The wording of Article 7 also provides that the potential liability of the platform is subject to the economic operator (including the product manufacturer) not being identified in the EU. This also provides a simple way to avoid liability. Indeed, a similar exception to distributor liability is already present under the current rules and can easily be avoided by good record keeping. In practice, distributors almost always share suppliers’ contact details.

A. Alternative Approaches

A series of different proposals has been mooted to increase the accountability of platforms for products sold on their websites. Erdem Büyüksagis has argued that an alternative solution would be to deploy national consumer laws and tort rules to fill the legal vacuum in respect of online platforms. He has thus raised the issue of extending vicarious liability to e-retailers as an alternative to classifying them as producers,⁹³ arguing that online marketplaces can be considered as organisations potentially subject to vicarious liability. The online marketplace's level of involvement in the conclusion of contracts should be taken into account and in cases where it is heavily engaged in transactions, the online marketplace should be held liable for defective products sold on its marketplace whether or not it engaged in conduct that breached a legally specified standard or in a business that involves a defect originating from a third party.⁹⁴

Büyüksagis also argues that in some countries contractual liability could be a vehicle as well, as in some jurisdictions, given the simple fact that the online marketplace collects, on behalf of the supplier, the money paid by the consumer, and transfers it to the former before the sold item has been delivered is enough to hold the online marketplace jointly liable with the supplier.⁹⁵

Professor Busch has put forward the proposition, drawing upon the relevant US case law, that the decisive factor determining whether a platform should be subject to product liability or not, should revolve around whether the platform has ‘control’ over the transaction.⁹⁶ He also notes that from a comparative perspective, this would parallel the ‘Uber test’ applied by the Court of Justice of the European Union (CJEU) in its recent case law on digital platforms,⁹⁷ thereby representing a general principle of responsibility transferrable to the field of platform liability. Busch also considers that in applying the criterion of control (or similar variants such as the notion of ‘predominant influence’ or ‘decisive influence’), it is justifiable to distinguish between platforms that offer fulfilment services and those that do not, with the degree of control being considerably higher in respect of the former than the latter.

Consumer organisations have also pointed out the unsatisfactory nature of the current position. In a series of reports,⁹⁸ BEUC, the European consumer organisation, has argued that online marketplaces should bear subsidiary liability as suppliers under the Product Liability Directive, and be liable where at least one of the following conditions are fulfilled: (1) the producer cannot be identified; (2) the marketplace fails to inform the consumer in due time of the identity of the producer and does not enable communication between the consumer and the producer by providing him or her with the relevant contact details; (3) the marketplace received clear evidence about a non-compliant product on its platforms; (4) the producer is identified but does not take measures to remedy the harm, or if (5) the marketplace has a predominant influence or control in the transaction chain.⁹⁹

At a European institutional level, the Committee on the Internal Market and Consumer Protection of the European Parliament in an Opinion in July 2020 made it clear that online marketplaces should be held liable under the Product Liability Directive.¹⁰⁰

B. Policy Proposals

Based on the various proposals set out above, a series of different parameters may be identified as a potential basis for founding the liability of online platforms.

1. Contractual relationship between the platform and consumer

One approach is to determine whether the platform is selling goods/services itself or instead acting on behalf of, or in the name of, a trader. There has traditionally been a focus on objective factors to determine whether platforms act as an ‘entity acting on behalf of the trader’.¹⁰¹ The Commission has stipulated that such is the case, inter alia, when the consumer can: (1) directly purchase the products compared or displayed (ie without rerouting the consumer to the manufacturer or another reseller); or (2) when a comparison tool is provided to consumers as a service in return for remuneration; or (3) when there is a material connection with a trader, such as advertising or sponsorship.¹⁰²

2. Impression given to the average consumer

The current approach in the various European reform projects outlined above indicates a preference for drawing parallels with the DSA and basing the relevant test for online platform liability on whether the average consumer is led to believe that the product is provided by the platform.¹⁰³ The CJEU applied a similar approach in the case of Sabrina Wathelet in the context of the Sale of Goods Directive 1999/44/EC.¹⁰⁴ It held that a person acting as an intermediary for a private individual could be regarded as being the ‘seller’ where he had not duly informed the purchaser of the identity of the owner of the goods.¹⁰⁵ Yet, with regard to the Consumer Rights Directive which applies the same concept of trader/seller, the CJEU followed a different logic in the case of Tiketa. The Court held that Directive 2011/83 does not determine the identity of the parties to the contract concluded with the consumer where the principal trader uses an intermediary, nor does it govern the attribution of liability between the principal trader and the intermediary in the event of failure to fulfil the obligations which it lays down. The CJEU held that the question whether the natural or legal person who is acting as an intermediary in the name of or on behalf of another trader has informed the consumer that he was acting in that capacity has no bearing on the classification of that intermediary as a ‘trader’. According to the definition, an intermediary is a ‘trader’ and thus must comply with the requirements laid down by that Directive.¹⁰⁶

3. Control

Another common aspect of many of the alternatives proposed is that of ‘control’ or ‘influence’. BEUC refers to the test of whether a marketplace has ‘a predominant influence or control in the transaction chain’.¹⁰⁷ Busch also proposes that the decisive factor for extending liability should be whether the platform has ‘control’ over the transaction or not.¹⁰⁸ In the USA, the extent to which platforms can superintend the sale of goods on their websites—and are therefore capable of intervening to prevent potentially dangerous products from reaching consumers—has played an important part in courts’ decisions finding Amazon strictly liable for harms caused by unsafe products sold by third-party vendors.¹⁰⁹ Indicia of control identified by US courts include placing restrictions (and requiring compliance therewith) on third-party vendors,¹¹⁰ warehousing third-party products and processing customer payments,¹¹¹ and retaining the power to block the sale of certain goods or goods sold by certain third-party vendors.¹¹²

From a European perspective, the ELI proposal combines the notion of control with the issue of how the product or transaction is presented, as noted above, with liability incurred if a consumer would be led to believe that the ‘product … is provided either by the online marketplace itself or by a trader who is acting under its authority or control’.¹¹³ The aspect of control relates broadly to control over the actual transaction. As noted above, this would present the advantage of consistency with the ‘Uber test’ applied by the CJEU in its case law on digital platforms,¹¹⁴ forming a coherent approach with a general principle of responsibility transferrable to the field of platform liability. What would be the relevant factors in determining whether that level of control has been reached? The crucial aspect of the test for platform responsibility in the Uber judgment was whether the platform (in that case Uber) ‘exercises decisive influence over the conditions under which that service is provided’.¹¹⁵ The relevant elements may thus include whether the platform sets prices and offers payment services, whether fulfilment services are provided in respect of the product, the contractual relationship with the platform/service provider, and the relevant actions of the platform in creating the marketplace itself.¹¹⁶ It is to be noted also that the E-Commerce Directive distinguishes between platforms that assume a passive role and those that assume an active role in transmitting data from its original source to the consumer,¹¹⁷ such as by ranking the options in a certain manner or encouraging loyalty and repeat purchases.

4. Power of the platform

Another consideration is the power exercised by the platform in question, with the basic premise being that the greater the size of the platform and its relevant market power, the greater the degree of responsibility that it consequently incurs.¹¹⁸ It is to be noted that this aspect has been mentioned in some of the US cases having extended liability to online platforms. For instance, the California state appellate court in Bolger v Amazon.com held that ‘Amazon … act[s] as a powerful intermediary between the third-party seller and the consumer [and] is the only member of the enterprise reasonably available to an injured consumer in some cases’.¹¹⁹

5. Effects on competition between different business models

Another parameter which might be taken into account when designing a product liability regime for the platform economy is how the regulatory model will affect competition between different market actors and thus the structure of the market.¹²⁰ In particular, the design of the liability regime may influence the decision of businesses regarding which position along the continuum between pure reseller and pure marketplace they choose.¹²¹ If, for example, stricter liability rules apply for those platforms that offer fulfilment services and for those who only offer direct shipping of products by third-party sellers, new entrants and smaller firms might shy away from offering fulfilment services in order to avoid exposure to product liability. As a consequence, they would offer consumers a poorer retail experience (no fast delivery, no seamless returns management, etc). In contrast, large platforms like Amazon would still be able to shoulder the risks of offering fulfilment services. It should be noted that these competition policy considerations are not per se an argument against a liability model that distinguishes between different platform business models. Rather, competition law and product liability law should be aligned to offer an adequate response to the competitive consequences of a particular regulatory choice in product liability law.

These factors are by no means exhaustive of course, and depending on the jurisdiction, as well as contextual factors such as the pre-existing regulatory environment or the underlying test relating to liability, the parameters may vary.

C. Caveats

Policymakers considering platform liability reform may want to take into account the extent to which increased liability exposure may result in higher operation prices for e-commerce companies and, resultantly, higher costs for consumers.¹²² Overall, ‘there is a general concern that if platforms like Facebook, Google, and Amazon face liability for user harm then these platforms will no longer find it worthwhile to spend resources to innovate and enhance the value of the user experience’.¹²³ At the same time, however, policymakers should bear in mind that there exists little to no concrete empirical evidence to support this concern, and that conjecture alone should not dictate the decision on whether to impose a liability rule.

Policymakers should also note that increased liability might also result in anti-competitive effects by benefiting incumbent e-commerce companies—particularly wealthy, established titans in the industry such as Amazon—insofar as a liability rule would disadvantage smaller start-ups to a greater extent.¹²⁴ The state of California's proposed legislation (discussed above) provides a salient example.¹²⁵ Although Amazon initially opposed the proposed legislation, the company backed it once it was revised to apply to other online retailers.¹²⁶ A statement released by Etsy's chief executive officer raised alarm bells concerning the impact the bill would have on Amazon's smaller competitors, stating that it would be so ‘expensive for any small or mid-sized business to try to comply with’ that ‘it could be crushing to smaller e-commerce players’ while only a mere ‘inconvenience to the e-commerce behemoth’.¹²⁷ For this reason, policymakers may want to complement changes in platform liability with policies designed to encourage platform competition.¹²⁸

VI. CONCLUSION

The liability of online marketplaces is an area that is ripe for reform. In grasping the reform nettle, policymakers will have to decide between regulatory and litigation models as outlined above, or, as in many countries, an approach combining the two. If liability is to be extended to platforms, then policy decisions will need to be made as to how exactly liability is framed in each jurisdiction and in particular the relevant test for liability. As has been seen above, there are different approaches to imposing liability, underpinned by different rationales, which will potentially affect the impact on platforms. If the choice is made to extend liability under a bespoke regime for product liability, then a series of factors may be seen as relevant to liability such as the degree of control or influence exercised over the transaction or the producer, the ability of the platform to influence the product supplier in favour of increasing product safety, the way that the transaction is presented and the associated impression given to the consumer, or economic factors such as the market power of the relevant online platform.

This comparative study provides the relevant information and reflection for that process in terms of presenting the status quo in various different systems, identifying the advances made in certain jurisdictions towards the limited recognition of the liability of online marketplaces, and highlighting the relevant factors to consider when implementing such a regime.