The Economic Burden of the Revised Product Liability Directive on EU and Non-EU Countries

2. Overview of the New EU Product Liability Directive

The PLD entered into force in December 2024 and will apply to products placed on the EU market from December 2026 onwards. Significantly expanding the scope and liabilities of the 1985 Product Liability Directive, the new Directive aims to modernise EU product liability rules in response to the growing role of digital technologies. Since its adoption, several EU countries have already started preparatory steps towards putting the PLD into practice, while others are at different stages of the transposition process, with final implementation expected by December 2026.[1] This chapter presents the main features of the PLD and explains how they will affect EU and non-EU countries.

2.1 New Categories of Covered Products

The new PLD drastically extends product liability to a wider range of products compared to the old regime. Most notably, the Directive now includes software, whether integrated into movable products or installed in immovable property (Article 4(1)). Moreover, the PLD is applicable regardless of the delivery method, covering software stored on devices, accessed via communication networks, or provided through the cloud and software-as-a-service (SaaS) (Recital 13). This inclusion is directly linked to a broader definition of liability: because software is now a product, compensable harm has expanded to include not only physical injury but also non-material damage, such as the loss or corruption of personal data (Recitals 20–21; Article 6).

The Directive also extends liability to the digital ecosystem. Integrated or interconnected digital services, such as IoT environments, are treated as components of the product if they influence safety and remain under the manufacturer’s control (Recital 17). Consequently, a hardware manufacturer can now be held liable for defects in a connected service, even if that service is supplied by a third party.

Finally, the Directive clarifies the status of open-source software and manufacturing files. While generally exempt, open-source software falls within the scope of the PLD if it is supplied commercially (Recital 15). In the case of digital manufacturing files, the Directive expressly classifies them as products. These files contain the functional data needed to produce tangible items using automated machinery, such as 3D printers. Their inclusion ensures that liability arises where defects in the digital file itself result in physical damage (Recital 16).

2.2 Expanded Scope of Liable Economic Actors

As the PLD extends its product coverage, it simultaneously expands the net of liable economic actors. This includes software developers, AI system providers, authorised representatives of manufacturers, distributors, fulfilment service providers,[2] and providers of online platforms. It also encompasses third parties who substantially modify a product after it has been placed on the market, treating them as manufacturers for liability purposes (as detailed further in Section 2.5).

A specific expansion of liability also applies to online platforms (Recital 38, Article 8). Online platforms in certain circumstances will be exposed to liability, especially if they present products in a way that makes consumers think the platform is the seller, or if they fail to quickly identify an EU-based responsible operator when a claimant requests it. Platforms may also be treated as ‘economic operators’ where they act as the manufacturer, authorised representative, fulfilment provider, or distributor of the defective product.[3]

Crucially, many of these actors – from platforms to component manufacturers – are both EU and non-EU manufacturers deeply embedded in the global supply chains delivering goods and services into the EU Single Market. The allocation of liability in these supply chains is inherently complex. In modern manufacturing and digital services, the final product often relies on upstream components that are outside the direct control of the business selling to the consumer. For EU-based actors, this creates a precarious position. When a defect arises from an upstream non-EU supplier, the EU importer or distributor (Article 8(3)) absorbs the full liability through a fall-back mechanism. In cases where multiple responsible parties can be identified, the PLD provides that where two or more economic operators are liable for the same damage, they may be held jointly and severally liable (Article 12).[4]

To mitigate this exposure, EU-based intermediaries are expected to shift liability back onto non-EU manufacturers through defensive commercial measures. We anticipate a rise in demands for broad indemnities, stronger warranties, and higher insurance coverage. In extreme cases, EU intermediaries may simply refuse to carry products from non-EU jurisdictions unless margins are increased to cover the risk. Consequently, both EU manufacturers reliant on global supply chains and non-EU exporters will face higher costs and barriers to entry, even if their products are perfectly safe.

At the same time, these pressures can produce liability gaps[5]. Liability may be attributed to the incorrect actor, dispersed across multiple parties such that the most relevant actor cannot be identified or pursued, due to inaccessible technical evidence or cross-border barriers. These challenges are pronounced for digital products where firms may lack oversight over all providers involved in the design, development and supply of components. In this context, PLD’s evidentiary presumptions are likely to amplify uncertainty by facilitating claims in situations where causation and fault are technically difficult to disentangle.

2.3 Prolonged Exposure and Uncapped Liability Levels

The Directive extends liability periods for specific categories of damage, replacing the previous ten-year limit with a maximum period of up to twenty-five years for latent personal injuries (such as slow-developing diseases). This prolonged exposure – a substantial increase from the old framework – affects domestic and foreign suppliers alike, forcing them to maintain records, evidence, and insurance coverage for at least a quarter of a century before they can consider their exposure to claims fully closed.

This extension is particularly risky given the rapid transformation of modern products. Due to the continuous evolution of emerging technologies, it is increasingly difficult to predict how a product will interact with the human body or the environment two decades from now. By extending liability to twenty-five years, the PLD forces manufacturers to insure against unknown future interactions and technological shifts that are impossible to model today.

Amplifying this temporal risk is the prohibition on financial limits. Article 11 of the PLD explicitly bans any contractual attempt to limit or exclude liability for defective products. Consequently, suppliers cannot use standard risk-management tools, such as liability caps or waivers, to define their maximum exposure.

2.4 New Procedural Thresholds and Presumptions of Liability

The PLD strengthens disclosure obligations (Article 9) beyond those in the 1985 Directive by allowing courts to order defendants to produce relevant evidence – including design documents, data sets, and product samples – once a claimant demonstrates that their claim is merely plausible. While courts may apply confidentiality safeguards, the disclosure process still increases the risk that defendants will have to disclose commercially sensitive information especially if the guardrails under national law are weak or even non-existent as is the case in several Member States.

Alongside these disclosure obligations, Article 10 of the PLD introduces mandatory rebuttable presumptions of defectiveness. PLD primarily targets the so-called David v. Goliath situation involving an individual claimant and ignores the growth of mass claims supported by specialised lawyers and litigation funders with deep pockets that make the David v. Goliath analogy no longer apt. With the David v. Goliath situation still in mind, PLD addresses the challenges raised by the complexity of modern products in liability cases by substantially lowering the evidentiary burden for the claimant: a court must presume a product is defective if the claimant merely shows it is likely that a defect exists. This effectively reverses the burden of proof. It is no longer up to the claimant to meet its burden of proving a fault; rather, the onus is on economic operators to actively disprove liability based on broad, vaguely defined and easily applied presumptions. Under the revised PLD defendants must prove a negative, an expensive and technically difficult undertaking, making the defence of even meritless claims significantly riskier, likely forcing defendants into abusive settlements.

2.5 Expanded Post-Market Obligations

The PLD expands the definition of defectiveness to account for the dynamic nature of digital products. Liability is no longer assessed solely on the product’s condition at the time of sale. Instead, a product may be deemed defective based on its post-market evolution, specifically its ability to learn (AI capabilities), acquire new features via software updates, or interact with other devices (Article 7).

A product that is substantially modified and put back on the market is treated as a new product. And if a product is modified outside the original manufacturer’s control, the person who made the change may be treated as the manufacturer and held liable (Recital 39). A modification is ‘substantial’ where it changes the product’s function, affects compliance with safety requirements, or increases its risks. Where the original manufacturer (or someone acting under its control) makes the modification and this leads to a defect, the manufacturer remains liable and cannot argue that the defect arose only after the product was placed on the market.

Furthermore, for complex technologies, a product may be treated as defective not because of a technical malfunction or design failure, but merely because its risks were not properly explained. Therefore, liability may arise simply because the product’s operation or risks were insufficiently disclosed to the user. This lowers the burden significantly for a finding of liability and creates unprecedented legal uncertainty for economic actors deploying novel technologies, especially where ‘foreseeable misuse’ is difficult to define.

To mitigate these risks, producers will be compelled to implement extensive logging, traceability, and post-market monitoring systems. They must maintain a continuous digital trail to prove the state of the software at the moment the harm was allegedly caused. This is because failing to meet mandatory safety rules, or refusing to disclose evidence may lead courts to presume a product is defective under Article 10(2) (see Section 2.4).

2.6 The Increased Risk of Mass Litigation

The PLD reforms must be viewed in tandem with the EU’s Representative Actions Directive (RAD) and the push to make it easier to launch collective actions within the EU. While the PLD lowers the substantive threshold for winning a case (through new presumptions and disclosure rules as described earlier), the RAD provides the procedural vehicle to scale up these claims.

Two specific changes in the PLD fuel this trend. Firstly, the Directive removes the previous €500 threshold for property damage. This creates a new liability market for low-value, high-volume claims (e.g., a minor data loss or software glitch affecting millions of users), making the aggregation of such claims in a mass litigation suit an attractive proposition. Secondly, the PLD broadens the definition of who can sue. Claims can now be brought not only by the injured individual but by parties ‘subrogated’ to their rights such as Qualified Entities (QEs). This facilitates professionalised litigation where well-funded entities drive claims in principle (but, according to many, not always) on behalf of consumers.

These new PLD mechanisms will act as a force multiplier for an already expanding collective redress ecosystem in Europe. The combination of the RAD and other EU laws such as the General Data Protection Regulation (GDPR) and the Digital Markets Act (DMA), has already contributed to an increase in the number of collective actions across EU members.[6] This growth has been fuelled by the availability of funding for this type of litigation from third-party investors in exchange for a profitable share of litigation proceeds. Although a lack of EU regulatory oversight makes the number of third-party litigation funders is difficult to quantify, evidence from some EU member states shows that their numbers are significant and increasing.[7]

Finally, and related to the previously discussed expanded post-market obligations (see Section 2.5), claimants may attempt to frame post-market software updates as evidence of prior defectiveness. Even where an update is a responsible safety improvement, it risks being invoked in court to argue that the original version was unsafe. Although the PLD attempts to clarify that a better version does not imply an earlier defect, the practical distinction between responsible improvement and proof of defect remains uncertain. This creates a perverse incentive where responsible product evolution becomes a potential litigation trigger.

Ultimately, the rise of mass litigation delivers an economic blow because private enforcement translates into higher operating costs. As the risk of collective redress grows, companies face intense pressure to secure broader insurance cover and enforce stricter risk allocation throughout their supply chains. This dynamic sets the stage for a correction in the insurance market where premiums will increase in response to the new liability environment.

[1] CMS. (2025, November 17). Transposition Time – Update on the EU Member States’ adoption of the new Product Liability Directive. Available at: https://cms-lawnow.com/en/ealerts/2025/11/transposition-time-update-on-the-eu-member-states-adoption-of-the-new-product-liability-directive

[2] Their inclusion reflects the Directive’s aim of ensuring claimants have access to an EU-based defendant, widening the range of economic actors exposed to claims.

[3] The PLD preserves liability protections for neutral online platforms under Article 6 DSA. However, this safeguard is conditional, and the risk of reclassification and litigation remains where a platform’s conduct goes beyond mere intermediation.

[4] In practice, contribution claims may fail where upstream developers are located outside the EU, or where harm involves open-source components exempt from liability. This may leave EU-based firms carrying the full liability.

[5]A liability gap arises where harm occurs but legal responsibility cannot be clearly assigned. Bertolini, A. (2013). Robots as products: The case for a realistic analysis of robotic applications and liability rules. Law, Innovation and Technology, 5(2), 214–247. European Commission. (2019). Liability for artificial intelligence and other emerging digital technologies: Report from the Expert Group on Liability and New Technologies. European Commission. Johnson, D. G. (2015). Technology with no human responsibility? Journal of Business Ethics, 127(4), 707–715.

[6] Erixon, F., Guinea, O., Pandya, D., Sharma, V., Sisto, E., du Roy, O., Zilli, R., & Lamprecht, P. (2025). The Impact of Increased Mass Litigation in Europe. ECIPE, Brussels, occ. paper 3/2025.

[7] Guinea, O., Pandya, D., & Sharma, V. (2025). Collective Action in the Netherlands: Why It Matters for the Transposition of the Product Liability Directive. ECIPE, Policy Brief, 11/2025; and Guinea, O., Pandya, D., & Sharma, V. (2025). The Economic Risks of Transposing the EU Product Liability Directive in Germany. ECIPE, Brussels, occ. paper 16/2025; also see: Lein, E., Bonzé, C., Salim, R., Alleweldt, F., & Kara, S. (2025). Mapping third party litigation funding in the European Union. European Commission. https://doi.org/10.2838/37951892

3. The Insurance Cost of Expanded Liability

Insurance companies translate legal risks into concrete financial costs. Where regulatory frameworks expand liability or lower evidentiary thresholds, particularly in consumer-facing sectors, the expected frequency, duration and severity of claims increase, forcing insurers to revise their risk calculations. Moreover, product-specific risk factors are also considered during pricing coverage, including the product’s complexity, intended use, and likelihood of defects or misuse. To protect themselves, insurers will charge higher premiums, narrow the scope of coverage, or require businesses to bear a larger share of initial costs.

Insurers price expected litigation risk rather than past claims alone, meaning that regulatory changes such as the PLD which increase legal uncertainty or anticipated claim frequency result in disproportionate premium increases. Where these assessments result in elevated premiums or constraints on available coverage, participation by overseas manufacturers and suppliers in the EU Single Market may be discouraged. This may sound like an exaggeration, but there is a growing number of examples of non-EU companies postponing the introduction of a product due to regulatory uncertainty and risk.[1]

Higher liability premiums will also affect EU importers. Insurance policies from outside the EU often have limited coverage or are difficult to enforce. When this is the case, liability claims arising from defective or damaged products can be directed towards the importer or distributor.[2] As a result, insurers incorporate the risk of unrecoverable upstream liability into their pricing models, leading to higher product liability premiums for importers and distributors.

In these circumstances, authorised representatives will have a strong incentive to obtain product liability insurance where none previously existed, or to expand existing coverage to reflect the expanded scope of potential claims. The European Commission estimates that, as a result of the new PLD, authorised representatives could see their product liability insurance costs and other insurance-related costs increase by 15 per cent. These costs are likely to be partially or completely passed on to the non-EU producers they represent.[3]

As described in Section 2.2, in the case of complex products, such as software, fault attribution is often difficult to determine. This creates a significant challenge for insurance companies that may be unable to identify where product liability risk is most likely to materialise. To manage this uncertainty, insurers are likely to apply broad, conservative price increases across entire sectors rather than targeting specific risks. This will inevitably lead to higher insurance costs for entire categories of software-enabled products.

The uncertainty surrounding AI-related liabilities is already causing a retreat in the insurance market. Major insurers are actively seeking to introduce new policy exclusions for claims arising from the use of AI, viewing the technology as too much of a ‘black box’ to underwrite reliably. The insurance industry fears that a single flaw in a widely used AI model could trigger systemic, multi-billion-dollar losses, a risk they are unwilling to bear. It has been reported that businesses wanting to insure AI-enabled products will face significantly higher premiums and narrower coverage terms.[4]

Finally, the new PLD increases both the frequency and the expected value of claims, attracting third-party capital and supporting the growth of third party profit-driven mass litigation (see Section 2.6). The increase in claim activity raises insurers’ expected losses beyond what would be implied by the merits of individual claims alone. Moreover, insurers’ defence costs, as a share of incurred losses, are relatively high in product liability lines, reflecting the substantial legal and technical resources required to contest such claims, especially in the discovery phase that is expanded by PLD.[5] As a result, insurance premiums are likely to rise still further.

This is not a theoretical argument but a pattern that has been observed in other markets where product liability has interacted with mass litigation. In the US, 28 per cent of product liability losses are driven by legal system abuse and social inflation, where rising costs reflect aggressive litigation tactics rather than underlying economic factors.[6] These escalating costs are increasingly fuelled by third-party litigation funding (TPLF), where hedge funds and private equity firms finance claimants’ legal costs in exchange for a share of the settlement. This creates a self-reinforcing cycle: increased funding supports more litigation, which drives up claim costs and larger verdicts, which in turn attracts further speculative funding and places severe financial strain on insurers and policyholders.

This dynamic has already crossed the Atlantic. In the UK, social inflation contributed more than 10 per cent to the growth of liability claims in 2023. While continental Europe has historically been shielded from these costs, Swiss Re warns that the expansion of collective action and broader liability definitions introduced by the PLD could drive a similar surge in social inflation across the EU over the coming years.[7]

Ultimately, the cost of higher insurance, fuelled by the new PLD and its drive towards mass litigation, will lead to higher premiums that are passed on throughout the supply chain, raising the expected cost for EU and non-EU manufacturers and innovators selling products in the EU Single Market. By contrast, producers located in jurisdictions characterised by lower regulatory burdens or lower litigation intensity may benefit from comparatively lower liability-related costs, including reduced insurance expenses. Inevitably, this will increase the cost base of EU firms, large and small, that may export outside the EU but also sell their products within it, reducing their competitiveness in both domestic and international markets.

[1] In July 2024, Meta announced it would withhold its advanced multimodal AI models from the EU market, explicitly citing the ‘unpredictable nature of the European regulatory environment. Source: https://www.theguardian.com/technology/article/2024/jul/18/meta-release-advanced-ai-multimodal-llama-model-eu-facebook-owner; Apple postponed the EU release of its Apple Intelligence features –  including iPhone Mirroring and SharePlay due to ‘regulatory uncertainties’ brought about by the Digital Markets Act (DMA). Source: https://www.reuters.com/technology/artificial-intelligence/apple-delay-launch-ai-powered-features-europe-blames-eu-tech-rules-2024-06-21/

[2] Coyle, G. (2025, December 23). Product Liability Insurance Cost 2025. The Coyle Group. Available at: https://thecoylegroup.com/product-liability-insurance-cost-2025/#:~:text=Does%20business%20size%20influence%20product,calculating%20premiums%20for%20small%20businesses.

[3] Impact assessment study on the possible revision of the Product Liability Directive (PLD) 85/374/EEC Author: Centre for Strategy & Evaluation Services (CSES) Date: June 2022 Contract No: 887/PP/GRO/IMA/20/1133/11700. Source: https://op.europa.eu/en/publication-detail/-/publication/e4854516-728b-11ed-9887-01aa75ed71a1

[4] Harris, L., & Criddle, C. (2025, November 23). Insurers retreat from AI cover as risk of multibillion-dollar claims mounts. Financial Times. https://www.ft.com/content/insurers-retreat-from-ai-cover

[5] In 2024, defence and cost containment expenses (DCCE) accounted for 33.6 per cent of incurred losses in product liability insurance in the US. This means that roughly one-third of insurers’ total claim costs in this line arise from legal defence and claim-handling expenses, not from compensation payments alone. The relatively high DCCE ratio reflects the complexity and cost of defending product liability claims, and implies that increases in claim frequency can raise insurers’ costs even where individual claims do not result in large payouts. See: Facts + Statistics: Product liability. Available at: https://www.iii.org/fact-statistic/facts-statistics-product-liability

[6] Insurance Information Institute, & Casualty Actuarial Society. (2025, October 30). Increasing inflation on liability insurance: Impact as of year-end 2024 [White paper]. Insurance Information Institute. https://www.iii.org/white-paper/increasing-inflation-on-liability-insurance-impact-as-of-year-end-2024-103025

[7] Boerlin, M., Finucane, J., Holzheu, T., Lechner, R., Vanolli, A., Wu, S., & Zhu, J. (2024). Social inflation: Litigation costs drive claims inflation (sigma No. 4/2024). Swiss Re Institute. https://www.swissre.com/institute/research/sigma-research/sigma-2024-04-social-inflation.html

4. From Ex-Ante Verification to Ex-Post Litigation

The new EU PLD introduces several features that fundamentally conflict with the established regulatory framework of conformity assessment and diminish the commercial and legal value of the services provided by the Testing, Inspection, and Certification (TIC) sector. This paradigm shift – from a system anchored in ex-ante verification to one where ex-post litigation – represents a critical departure from the principles that have underpinned the EU’s regulatory approach for decades.

The established framework relies on conformity assessment to verify compliance before products reach the market. Conformity assessment in the EU was based on the so-called ‘New Approach’ adopted by the EU in 1985. The New Approach established a system of checks to ensure products meet the required level of safety. For many lower-risk products, businesses can issue a Supplier’s Declaration of Conformity (SDoC) and affix a CE Marking to declare the product complies with all essential requirements.

However, private enterprises do not have to undertake the conformity assessment themselves. They can buy the services of third-party conformity assessment firms in the TIC sector. These firms will determine the characteristics of a product (testing); examine the product to ensure it meets the specified criteria (inspection); and provide a formal written assurance that a product, service, or system meets specific requirements (certification).

A certificate of compliance served as a proof of compliance. However, under the PLD, this protection is substantially weakened because the formal status of approval that businesses have invested significantly to obtain can be invalidated during litigation, rendering the compliance verification process non-definitive. This new uncertainty directly undermines the commercial incentives for suppliers to buy the services provided by the TIC sector.

Even though certification was never a ‘perfect shield’ against a lawsuit, the PLD signals a fundamental shift in approach by introducing procedural mechanisms that systematically favour claimants (see Chapter 2). This de facto reversal of the burden of proof fundamentally alters the value of third-party certification. While TIC documentation remains relevant evidence, it is no longer as determinative as it once was.

Finally, it is important to acknowledge that the conformity assessment sector is distinctly European. Major global leaders such as SGS, Bureau Veritas, Intertek, and TÜV SÜD are all European-based organisations. This European dominance is not accidental but a consequence of the EU’s market size and stringent regulatory standards. Exporters hire European-accredited labs to certify their products before shipping, creating a demand that has established Europe as the global centre of conformity assessment.

5. Estimating the Economic Impact: Methodology and Scenario Design

5.1 Transmission Channels for EU and Non-EU Producers

In its PLD Impact Assessment (IA), the European Commission estimates that liability insurance costs account for between 0.1 and 0.5 per cent of a product’s value, depending on the product category and risk profile.[1] At the same time, the IA estimates that liability insurance costs for complex products could rise by 10 to 15 per cent. Moreover, in the IA accompanying the withdrawn AI Liability Directive, the European Commission estimated an increase in liability insurance premiums of around 25 per cent due to AI-related liability risks, rising to 35 per cent when combined with a mandatory insurance obligation.[2] While these estimates relate to a different Directive, they provide a useful benchmark for the likely scale of expanded insurance liability costs for stand-alone software, including cases where software is embedded in a product. This approach is consistent with the PLD’s recognition that certain product categories beyond AI, such as life-sustaining medical devices, entail particularly high risks and, therefore, stricter safety and liability expectations.[3]

The figures put forward by the European Commission assume that insurers can accurately price liability risks. As explained in Chapter 3, in practice this is very challenging and insurers are likely to reflect this uncertainty in higher premiums. These cost effects are reinforced by the fact that the revised liability framework changes not only how risk is assessed but also how easily claims can succeed. In particular, as liability becomes easier to establish, for example through lowering of the burden of proof, insurers may raise the cost of insurance, especially where risks remain difficult to quantify.

While the new PLD fundamentally transforms liability for digital products, its regulatory reach also extends to traditional goods already governed by the 1985 Directive. In established sectors, the risk profile for non-complex products remains unchanged; however, the reform introduces procedural changes that amplify the cost and likelihood of litigation. In particular, the introduction of rebuttable presumptions of defectiveness and strengthened disclosure obligations significantly lower the evidentiary bar for claimants. Furthermore, extending the liability period to twenty-five years for latent injuries and removing the €500 threshold for property damage will heighten long-tail exposure and facilitate mass litigation across all economic sectors. Consequently, although the economic impact on traditional sectors is incremental compared to newly covered digital goods, these procedural modifications introduce a new layer of legal uncertainty and administrative friction that insurers will inevitably price into their premiums.

Finally, the PLD will have effects on non-EU companies that go beyond insurance costs. The PLD reduces the value of certification by shifting the regulatory approach from ex ante verification to US-style ex post litigation. Certification should serve as a clear signal of product safety and compliance with rigorous EU standards; however, the PLD will allow courts to give less weight to such third-party certification. As a result, exporters are likely to face a more demanding contracting environment, with importers seeking stronger warranties, driving up costs and making trade with the EU more expensive and riskier.

Taken together, these mechanisms form a compounding regulatory structure in which higher insurance costs, lower evidentiary thresholds, and devalued certification reinforce one another. Table 1 summarises the principal channels through which the revised PLD translates into economic costs for EU and non-EU producers.

Table 1: Summary of Economic Effects on EU and Non-EU ProducersSource: ECIPE.

5.2 Defining the Affected Economic Sectors

5.2.1 Newly Covered Sectors

The legal and procedural shifts detailed in the preceding chapters do not apply uniformly across all product categories. As explained, the risks are most pronounced in sectors characterised by high technological complexity, deep integration of digital components, and long international supply chains.

Our analysis identifies five sectors principally affected by the new PLD: Automotive and Vehicle Systems, Medical Devices, Information Technology, Consumer Electronics and Household Products, and Industrial and Construction Equipment. Many products within these industries rely on continuous connectivity, regular updates, and adaptive learning processes, meaning their behaviour and functionality can evolve after being placed on the market. Consequently, potential harm arising from software modifications, algorithmic decision-making, or interactions between interconnected systems creates uncertainty over where responsibility lies and when it arises.

However, exposure to these new liability rules varies within each sector, depending on the extent of digital integration. The study narrows its focus to specific product categories that incorporate – or have the capacity to incorporate – software, AI, Internet of Things (IoT) technologies, and digital manufacturing files. In total, this paper identifies 73 product categories across the five sectors that are capable of integrating software and digital services (the full list is provided in Annex 1). Furthermore, the new PLD applies directly to standalone software as a product, our analysis also includes information technology (IT) services.

Table 2 illustrates the proportion of each sector captured by the expanded scope of the new PLD. Automotive and Vehicle Systems and Information Technology are fully included because virtually all modern vehicles depend on embedded software and connectivity, and IT products are inherently digital by definition. In contrast, the remaining three sectors exhibit partial exposure. Within Medical Devices (29.9 per cent), Consumer Electronics and Household Products (36.1 per cent), and Industrial and Construction Equipment (12.1 per cent).

Table 2: Share of Sectoral Trade Flows Newly Covered by the Directive (%)Source: ECIPE.

5.2.2 Traditional Sectors (Covered by the 1985 Directive)

For the wider economy already governed by the 1985 Directive, the impact of the revised PLD will be less transformative but remains economically significant. The new PLD therefore affects them primarily through procedural adjustments, heightened compliance requirements, and growing litigation risks.

These sectors include economic areas such as agriculture, food and primary products; mining and energy; chemicals, pharmaceuticals, basic metals; and textiles, and consumer goods. Additionally, the non-digital subsets of the five primary sectors identified previously, including Medical Devices (70.1 per cent), Consumer Electronics and Household Products (63.9 per cent), and Industrial and Construction Equipment (87.9 per cent) also fall into this secondary tier.

5.3 Modelling Assumptions and Scenario Design

Our economic modelling makes two important distinctions. First, it separates the effect of the PLD on the economic sectors newly affected by the Directive from the economic sectors that were already subjected to the 1985 Directive, but which will nonetheless see growing costs reflected in higher insurance premiums. The second distinction is between the effect of the new PLD in EU countries and non-EU countries. A description of the economic modelling and the scenarios is presented below.

5.3.1 Insurance Costs as a Baseline Percentage of Production

While the European Commission’s own analysis for the PLD provides a useful starting point, its estimates for insurance cost increases do not fully capture the commercial realities of the insurance market. The European Commission’s baseline, which assumes liability insurance costs of 0.1 to 0.5 per cent of a product’s value, is a general benchmark. However, the product categories most affected by the PLD are inherently more complex and carry a higher risk profile than the average product. Moreover, the European Commission’s IA did not account for many of the risks and economic costs described in Chapter 3 and Section 4.1. Therefore, a baseline insurance cost of 1 per cent of product value is a more realistic, albeit still conservative, assumption. For comparison, liability insurance costs in the US are generally higher than this assumption, reflecting higher litigation exposure.

5.3.2 Premium Growth for New Sectors

The European Commission’s estimated premium increase of 10 to 35 per cent underestimates the compounding effect of several risk factors explained earlier. Insurers must price not only the expanded scope of liability but also the heightened uncertainty, the difficulty of fault attribution, and the increased likelihood of mass litigation, which drives costs far beyond the merits of individual claims. A 40 per cent increase in premiums is, therefore, assumed to better reflect this new, riskier legal environment. Critically, in cross-border supply chains, this cost is incurred at least twice: once by the non-EU producer and again by the EU importer, who must also secure liability coverage. This ‘double incidence’ of insurance costs is a practical reality of international trade that amplifies the total cost of the PLD.

5.3.3 Premium Growth for Traditional Sectors

For traditional manufacturing sectors, the economic impact will be more contained, primarily driven by new compliance requirements and procedural adjustments. While the PLD formally expands liability, its practical impact on these established sectors will be incremental. These industries have long faced claims under national tort law, meaning the physical risk profile of their products remains unchanged. Consequently, this study estimates that for sectors already covered under the previous PLD, insurance premiums will see an additional increase of 5 to 10 per cent. This projection has been validated by insurance industry experts, who confirm it aligns with their expectations.

5.3.4 Simulating Domestic Shocks (EU Countries)

To measure the regulatory burden of the revised PLD on the domestic EU economy, the study employs an “all-input augmenting technological change” shock, which captures how the new PLD increases the baseline cost of doing business within the Single Market. In economic modelling, this is represented as a percentage decrease in Total Factor Productivity (TFP). Economically, this means that a sector now requires significantly more inputs to produce the exact same amount of final output.

In the real world, this shock translates the risks of the PLD into tangible economic costs. Instead of allocating capital toward product development, EU manufacturers and innovators are forced to divert critical resources to cover surging liability insurance premiums and implement extensive data logging and post-market monitoring systems to defend themselves against potential mass litigation. Consequently, they have fewer resources available for productive technology and are less incentivised to invest in new digital technologies or iterate on existing products. Over time, this increase in operational costs leads to diminished access to the technology frontier and lower economic growth.

5.3.5 Simulating Trade Shocks (Non-EU Countries)

In contrast to the domestic EU economy, to measure the regulatory burden of the revised PLD on non-EU countries, the study employs an “import-augmenting technological change” shock, which captures how the new Directive increases the baseline cost of exporting into the Single Market. In economic modelling, this is represented as a decrease in bilateral trade efficiency, often referred to as an “iceberg trade cost.” This means that a foreign manufacturer now incurs higher transaction costs simply to deliver the exact same amount of final goods to the EU. Unlike EU-based producers who face unavoidable domestic cost increases, non-EU partners incur this friction only when interacting with the EU market.

In practice, this shock translates the extraterritorial risks of the PLD into tangible non-tariff trade barriers. It reflects the reality of the ‘contractual squeeze’ where EU importers and distributors demand stricter warranties, extensive technical documentation, and higher liability insurance from non-EU suppliers to offset their own legal exposure. In the long-term, this increase in market entry costs disincentivises non-EU entities from selling products in Europe, leading to diminished access to global innovation for EU businesses and consumers.

5.3.6 Scenario Outlines (Lower and Upper Bounds)

Based on these assumptions, the study sets out two scenarios, presented in Table 3, which underpin the computable general equilibrium (CGE) model used to estimate the costs of the PLD (see Annex 2 for a full description of the model).

The lower-bound Scenario 1 assumes that liability insurance costs for newly impacted sectors increase by 40 per cent, and by 5 per cent for economic sectors previously affected by the 1985 PLD. Importantly, under this Scenario the impact of the PLD through higher trade costs only affects exporters rather than both exporters and importers. The same increase in liability insurance costs is applied to the EU economy (see Section 6.1) and to non-EU countries (see Section 6.2).

Scenario 2 represents the upper bound. It assumes that higher insurance costs for the products and IT services newly affected by the PLD will be borne by both non-EU producers exporting to the EU and EU enterprises importing these goods and services (see Section 4.3.5). This combined effect is captured by doubling insurance costs as a share of product costs – from 1 per cent in Scenario 1 to 2 per cent in Scenario 2 – while the assumed 40 per cent increase in liability insurance costs remains unchanged.[4] For economic sectors that were already in the scope of the 1985 PLD, Scenario 2 assumes that the increase in liability insurance costs is equal to 10 per cent, and this increase applies to both EU importers and non-EU exporters, again doubling the effect. In the case of the EU economy, because the effects of the PLD affect the domestic economy directly (see Section 4.3.4), the increase in liability insurance is equal to 40 per cent for economic sectors newly brought within the scope of the PLD and 10 per cent for economic sectors that were already in the scope of the PLD.

The total increase in product costs is calculated by multiplying insurance costs (as a share of product costs) by the assumed increase in liability insurance costs under each scenario. The actual coefficients entered into the economic model are adjusted by the share of affected products within each sector. For instance, while the Automotive and IT sectors have a 100 per cent share and receive the full 0.4 (Scenario 1) or 0.8 (Scenario 2) shock, Medical Devices has a share of 29.9 per cent, resulting in a lower shock coefficient (0.12 in Scenario 1) (see Annex 2 for a full list of the shocks per individual sector).

Table 3: Modelling Scenarios and Shock CalibrationsSource: ECIPE.

[1] European Commission (2022). Impact assessment study on the possible revision of the Product Liability Directive (PLD) 85/374/EEC – No. 887/PP/GRO/IMA/20/1133/11700, p. 152. Based on the Commission’s assessment, the benchmark data on product liability insurance costs could not be obtained from Europe, and was based on US data where commercial providers provided information about general product liability insurance costs.

[2] Commission Impact Assessment. (2022). Proposal for a Directive of the European Parliament and of the Council on adapting non-contractual civil liability rules to artificial intelligence. Available at: https://commission.europa.eu/system/files/2022-09/1_4_197608_impact_asse_dir_ai_en.pdf

[3] Ibid.

[4] This is, however, an imperfect assumption. In practice, exporters and importers form part of the same value chain. The final cost burden would be shared according to the relative bargaining power of the parties and the availability of alternative suppliers. The cost is not automatically duplicated; rather, it is negotiated and shared. However, it is not uncommon for both sides of a transaction to take out insurance, much as all drivers purchase car insurance regardless of who is more likely to cause an accident. Taking these factors into account, Scenario 2 serves as the upper-bound estimate.

6. Simulation Results: The Economic Impact of the PLD

The analysis that follows focuses exclusively on the cost side of the regulation. The new PLD is expected to deliver benefits to consumers, including easier access to compensation for harm allegedly caused by defective software and AI systems. The quantitative analysis presented here does not seek to minimise these benefits, but rather to estimate the economic costs that are the necessary counterpart of those benefits, and which will be borne by EU and non-EU firms selling their products in the EU Single Market.

6.1 Economic Costs on EU Countries

The PLD is expected to have a significant negative impact on the EU economy. EU GDP is estimated to be 0.17 and 0.24 per cent lower in Scenarios 1 and 2, respectively, than it would be in the absence of the PLD (as shown in Figure 1). To put this in context, the estimated GDP cost of the PLD is larger than the projected gains from several of the EU’s most significant recent trade agreements. The European Commission’s own impact assessment for the EU-Japan Economic Partnership Agreement estimated a long-run GDP gain of 0.14 per cent for the EU.[1] The projected gains from the EU-Mercosur agreement, which took over two decades to negotiate, are estimated at around 0.05 per cent of EU GDP.[2]

Figure 1: Estimated Decline in EU GDP (% Change)Source: ECIPE, authors’ calculations.

The fall in GDP is the result of the combined effect of the two shocks described earlier. On the production side, PLD raises the cost of domestic output: EU firms that must allocate more resources to insurance, compliance, and legal risk management produce less per unit of input, lowering TFP across the affected sectors. On the trade side, the import-augmenting technological change shock increases the effective cost of trading with the EU. The net result is a contraction in real output: production becomes less efficient, imported inputs become more expensive, and the economy’s overall capacity to generate value diminishes.

Figure 2 presents the estimated reduction in EU imports by sector under the new PLD. Pharmaceuticals and life sciences experiences the largest contraction across both scenarios, falling by 0.50 per cent under the lower bound and 0.73 per cent under the upper bound. Information and Communication Technology (ICT) follows a similar pattern, with import reductions ranging from 0.34 per cent in the lower bound to 0.71 per cent in the upper bound. The automotive and transport manufacturing sector and electronics, optical products, and consumer durables occupy the middle range with a fall between scenarios of 0.47 to 0.58 per cent and 0.36 and 0.56 per cent respectively, while industrial machinery and electrical equipment reports falls ranging between 0.32 and 0.42. Sectors already covered by the 1985 PLD face a smaller increase in liability insurance costs and, consequently, a more limited impact, ranging from 0.06 to 0.13 per cent.

Figure 2: Estimated Decline in EU Imports by Sector (% Change)Source: ECIPE, authors’ calculations.

In terms of public procurement, even though contracting authorities tend to source disproportionately from domestic suppliers, with only 3 per cent of public sector consumption coming from foreign imports, the EU public sector as a whole remains a large purchaser of foreign products. In 2017, the EU’s public sector imported goods and services worth €122 billion, amounting to 6.9 per cent of total goods and services imported by the EU.[3] Similar to other economic actors, rising import prices following the implementation of the PLD will lead public authorities to reduce import demand and pass part of these higher costs on to EU households and economic actors that rely on public services. As presented in Figure 3, the EU public sector’s imports are projected to fall between €1.45 billion and €2.5 billion as a result of the PLD.

Figure 3: Estimated Decline in EU Public-Sector Imports (€ millions)Source: ECIPE, authors’ calculations.

Ultimately, the increase in import prices and the fall in productivity will be borne by European households. For products with limited substitutes, where few domestic suppliers exist, households may be forced to cut spending elsewhere to absorb the higher costs induced by the new PLD. For other products, EU consumers may shift demand towards domestically produced alternatives, provided these are available at lower relative prices. However, the productivity shock weakens this substitution channel: because EU producers themselves face higher insurance and compliance costs, domestic alternatives are unlikely to offer meaningful price relief. The overall effect is likely to be a decline in European households’ standard of living, driven by the combined pressure of costlier imports and less efficient domestic production.

These effects are captured in our economic modelling results in the form of a variable that measures the equivalent variation.[4] This economic concept measures the monetary amount that households should receive to make them as well-off as they were before the implementation of the PLD. Figure 4 measures this amount as a cost to EU households, estimated to be between €24 billion (lower bound Scenario) and €33 billion (upper bound Scenario) in total, translating to €119 and €163 per EU household, respectively.

Figure 4: Estimated Cost Burden on EU Households (€ billions and € per Household)Source: ECIPE, authors’ calculations. Note: results are expressed in euros, converted using the 2025 exchange rate of 1 USD = 0.885 EUR. Source: European Central Bank.

The divergence between the cost estimates produced in this study and those of official bodies is not a matter of degree, but of methodology. Both the European Commission’s Impact Assessment[5] and the German government’s Ministerial Draft explanatory memorandum[6] restrict their analyses to the direct administrative and compliance mechanics of the Directive. These are real costs, but they represent only the visible surface of the PLD’s economic impact. What these assessments do not capture, and what this study explicitly quantifies, are the behavioural and market-level responses that the Directive will trigger, as described in Section 5.1. As a result, the analyses conducted by the European Commission and the German government severely underestimate the economic cost of implementing the revised PLD.

6.2 Economic Costs on non-EU Countries

6.2.1 Baseline Trade Values for Affected Non-EU Countries

As a Directive that applies to all products placed on the EU market, the new PLD will have a significant impact on the largest EU suppliers. Looking only at EU imports of the 73 product categories and IT services described in Section 4.2.1 and Annex 1, we find that the new PLD will impact 15.3 per cent of the EU’s total imports, accounting for €586 billion. To put this into perspective, the €586 billion worth of imports affected by the Directive exceeds the EU’s total imports from China, its largest supplier, which amounted to €522 billion in 2024.[7]

Across the five sectors (see Figure 5), consumer electronics and household products account for the largest import flows affected by the new PLD, at €222.7 billion, followed by automotive and vehicle systems (€137 billion) and the IT sector (€125 billion). Industrial and construction equipment, together with medical devices, account for a combined €101 billion in imports affected by the PLD.

Figure 5: Value of EU Imports in Newly Covered Sectors (2024, € billions)Source: Eurostat, author’s calculations.

These EU imports of goods and services worth €586 billion were supplied by 213 countries and territories.[8] Figure 6 shows the ten largest exporters to the EU of the 73 affected goods in 2024, while Figure 7 presents the corresponding figures for IT services imports.

China was the largest supplier to the EU of the 73 goods directly affected by the new PLD, with exports worth €183 billion. This was followed by the US, the UK, Turkey, and Japan, with combined exports of affected goods to the EU worth €125 billion. Beyond the EU’s traditional trade partners, newer and smaller partners will also see a significant share of their exports to the EU affected by the PLD. For example, Vietnam, Taiwan, and Thailand rank among the top ten countries most affected by the new PLD. Together, these countries export around €50 billion worth of goods to the EU that will fall within the scope of the new PLD.

Figure 6: Top 10 Non-EU Suppliers of Newly Covered Goods (2024, € billions)Source: Eurostat, author’s calculations.

Looking at the import of IT services, Figure 7 shows the top 10 exporters to the EU impacted by the new PLD. The US and the UK were the leading suppliers of IT services to the EU with combined exports affected by the new PLD amounting to €77 billion. India ranked as the third-largest exporter of IT services to the EU in 2024, highlighting the broad and significant impact of the new PLD on both established and emerging trade partners. Even smaller economies, such as Singapore, would see services trade flows of around €2 billion affected by the new PLD.

Figure 7: Top 10 Non-EU Suppliers of Newly Covered IT Services (2024, € billions)Source: Eurostat, author’s calculations.

6.2.2 Projected Export Losses

As previously explained, as a result of the PLD, the EU will experience a drop in imports from non-EU countries. This fall in imports is estimated to be €21 billion under the lower bound scenario and €36.5 billion in the upper bound scenario. Figure 8 shows the percentage decrease in exports to the EU in the lower and upper bound scenarios for a selection of countries. India and the US face the largest relative decreases, with their exports to the EU falling by 1.1 per cent in the lower bound scenario and 1.8 per cent in the upper bound. Switzerland experiences the next largest reduction, with exports falling by 0.7 per cent under the lower bound and 1.3 per cent under the upper bound. Canada’s exports to the EU decline by 0.9 per cent and 1.5 per cent in the lower and upper bound scenarios respectively. The Rest of the World (RoW) aggregate follows a similar pattern, with a 0.9 per cent reduction in the lower bound rising to 1.5 per cent in the upper bound. The UK records falling exports to the EU between 0.4 per cent in the lower bound and 0.9 per cent in the upper bound.

This reduction in exports is driven by the two compounding forces described previously. First, exporting to the EU becomes structurally more expensive as non-EU producers are forced to internalise higher liability insurance premiums and increased compliance costs. Second, as the EU economy absorbs its own negative shock, which makes domestic production less efficient, the EU becomes poorer in aggregate, leading to a contraction in overall demand for goods from abroad.

Figure 8: Estimated Decline in EU Imports from Key Trading Partners (% Change)Source: ECIPE, authors’ calculations.

Because these reductions are expressed in percentage terms, countries with larger baseline trade volumes will experience significantly greater absolute losses. Figure 9 illustrates this clearly. Among the selected countries, the US faces by far the largest absolute contraction, with exports to the EU falling by €3.6 billion in the lower bound scenario and €6.2 billion in the upper bound. Switzerland and India follow, with estimated export losses of €884 million and €766 million respectively in the lower bound, rising to €1.7 billion and €1.3 billion in the upper bound. The UK records a loss of €721 million under the lower bound and €1.5 billion under the upper bound. Canada, despite facing one of the steeper percentage declines shown in Figure 8, records the smallest absolute loss among the five countries at €253 million in the lower bound and €432 million in the upper bound, a direct consequence of its comparatively smaller trade volumes with the EU in the affected product categories.

Figure 9: Estimated Decline in EU Imports from Key Trading Partners (€ millions)Source: ECIPE, authors’ calculations.

[1] European Commission (2012). Trade sustainability impact assessment of the free trade agreement between the European Union and Japan (Final Report). European Commission, Directorate-General for Trade. https://policy.trade.ec.europa.eu/eu-trade-relationships-country-and-region/countries-and-regions/japan/eu-japan-agreement_en

[2] European Commission (2025). Economic analysis of the negotiated outcome of the EU-Mercosur partnership agreement (EMPA). Publications Office of the European Union. https://op.europa.eu/ en/publication-detail/-/publication/6f1a741f-677e-11f0-bf4e-01aa75ed71a1/language-en

[3] Prometeia SpA, BIP Business Integration Partners – Spa, Economics for Policy a knowledge Center of Nova School of Business and Economics Lisboa, (2021), Study on the measurement of cross-border penetration in the EU public procurement market. European Commission

[4] Hertel, T. W. (Ed.). (1997). Global Trade Analysis: Modeling and Applications. Cambridge University Press.

[5] European Commission (2022). Impact assessment study on the possible revision of the Product Liability Directive (PLD) 85/374/EEC – No. 887/PP/GRO/IMA/20/1133/11700. Table 4.70. Under the preferred policy option, the study estimates incremental annual insurance costs for producers equal to €28.1 million and incremental enforcement costs equal to €1.59 million.

[6] Federal Ministry of Justice and Consumer Protection, Ministerial Draft (Referentenentwurf) of the new German Product Liability Act, 11 September 2025, Explanatory Memorandum. The memorandum estimates total national costs at approximately €72,000, of which the majority falls on the public sector (state and federal courts) for evidence disclosure processing and publication of decisions. For the private sector, the memorandum assumes “no change in compliance costs for the economy”; the only firm-level cost acknowledged is the time required to respond to court disclosure orders, estimated at 40 hours per case across 1,000 annual cases, yielding a total private sector cost of €26,000.

[7] Author’s calculations based on Eurostat data

[8] Eurostat data shows trade flows to partner countries as well as territories. Territories are usually offshore areas that might belong to a certain country but have their own trade and customs rules. Examples of territories included in the Eurostat database include French Southern Territories, British Indian Ocean Territories, Virgin Islands (British) and Virgin Islands (United States), etc.

7. Conclusion

The revised Product Liability Directive (PLD) represents a fundamental shift in how the EU regulates the digital economy. While its stated objective is to modernise consumer protection, this study demonstrates that the Directive imposes a severe dual economic burden: it acts as a domestic drag on European economic output while simultaneously erecting new barriers for international trade.

This dual burden is driven by a structural shift in the rules that govern product liability in the EU. By expanding the definition of a product and of the types of damages that may be claimed, extending liability to new actors, and lowering the evidentiary bar for claimants, the PLD forces businesses to internalise unquantifiable legal risks. The inevitable result is a surge in liability insurance premiums and compliance costs. These costs impact both non-EU exporters and EU domestic industries selling their goods and services into the EU Single Market.

Our economic modelling quantifies the magnitude of this regulatory shift. The key empirical findings are:

EU Domestic Impact

• Contraction of the EU economy: The PLD is projected to reduce EU GDP by between 0.17 and 0.24 per cent. This domestic economic loss exceeds the projected long-term gains from several of the EU’s most significant recent free trade agreements.
• Cost to European households: The Directive will erode European household purchasing power by an estimated €24 billion to €33 billion annually, equivalent to a cost of €119 to €163 per household.
• Strain on public services: The EU public sector faces a €1.45 billion to €2.5 billion reduction in the imports it relies upon to deliver public services to EU citizens.

Trade and Non-EU Impact

• Substantial volumes of affected trade: The new PLD brings 15.3 per cent of the EU’s total imports within its scope, representing €586 billion in annual trade – a figure that exceeds the EU’s total annual imports from China.
• Reduction in EU imports: Total EU imports from non-EU countries are projected to fall by between €21 billion and €36.5 billion as foreign producers absorb higher insurance and compliance costs.
• Export losses across trading partners: Every country exporting to the EU will see a reduction in its exports as a result of the PLD. Under the upper bound scenario, India and the US face the largest relative declines, with exports to the EU projected to fall by 1.8 per cent. Canada follows closely with an estimated reduction of 1.5 per cent, while Switzerland and the UK record projected declines of 1.3 per cent and 0.9 per cent respectively.

Sectoral Vulnerability

• Disproportionate impact on technology and health: Imports of pharmaceuticals and life sciences are projected to fall by up to 0.73 per cent, and ICT by up to 0.71 per cent. These sectors are central to Europe’s long-term economic ambitions.

The revised PLD illustrates the hidden costs of aggressive unilateral regulation. While the intention to protect consumers in the digital age is legitimate, the chosen mechanism – relying on expanded liability and expensive litigation – creates profound inefficiencies. By raising the cost of both domestic production and international trade, the Directive risks undermining Europe’s economic competitiveness, slowing its adoption of new technologies, and straining relationships with global trading partners.