The Cartier case arose from a trademark dispute in which three well-known luxury brands demanded that several British internet access providers block or at least attempt to block access to specified target websites. These websites were advertising or selling counterfeit copies of their goods. ISPs, as typical access providers, only provided networks by which internet subscribers access content, but they neither provided nor stored any of the objectionable content. Therefore, it was common ground that they do not themselves use and thus infringe the relevant trademarks.
The right holders nevertheless requested assistance in enforcement. Armed with Art 11 third sentence of the Enforcement Directive, they demanded that non-infringing (innocent) third parties be subject to a particular type of enforcement measure. Although UK case-law in the area of copyright law accepted such measures a while ago, there is no comparable legal basis in the trademark statute. However, as previously ruled by the House of Lords, equitable jurisdiction of the English courts is unfettered by the statute. Or as Sir George Jessel once explained this: “I have unlimited power to grant an injunction in any case where it would be right or just to do so: and what is right or just must be decided, not by the caprice of the Judge, but according to sufficient legal reasons or on settled legal principles.” (Beddow v. Beddow  9 Ch. D. 89 at 93.)
The High Court
Against this background, Arnold J, hearing Cartier at the High Court ( EWHC 3354 (Ch)), accepted that the court has general equitable jurisdiction to order an innocent third party to implement website blocking. Since the UK legislator did not implement Union law in the field of industrial rights, the trademark law-related blocks had to be considered within the framework of equity. Arnold J accepted the jurisdiction and exercised his discretion by granting the order. He argued (para 106) that both domestic and the Union-conform interpretation warrant such a result:
“An analogy may be drawn with the equitable protective duty described by Buckley LJ in Norwich Pharmacal Co v. Customs & Excise Commissioners  AC 133 ( . . . ) Although this principle is inapplicable to the circumstances of the present case, it is not a long step from this to conclude that, once an ISP becomes aware that its services are being used by third parties to infringe an intellectual property right, then it becomes subject to a duty to take proportionate measures to prevent or reduce such infringements even though it is not itself liable for infringement. ( . . . ) I conclude that, even if the Court would not have power to grant a website blocking injunction in a trade mark case upon a purely domestic interpretation of section 37(1), section 37(1) can and should be interpreted in compliance with the third sentence of Article 11 by virtue of the Marleasing principle. If it were otherwise, the UK would be in breach of its obligations under the Directive.”
As I argued elsewhere (Husovec, Injunctions Against Intermediaries in the European Union: Accountable But Not Liable? (CUP 2017)), it is questionable whether EU law really forces this outcome. Though one has to admit that many national courts understand the obligation as imperative in terms of players (e.g. BGH; OGH), but modular in terms of the measures and their terms. Despite this, the holding is seminal. It is “Norwich no. 2” because Cartier essentially generalizes equitable duties of innocent third parties beyond mere provision of information. This means that even non-wrongdoers have to come to the rescue of IP right holders when requested by the court (following their petition).
For the costs of compliance, however, Arnold J did not follow the traditional approach. The practice since Twentieth Century Fox Film Corpn v British Telecommunications plc (No 2)  1 All ER 869 has been to order the right holders to bear their costs of the unopposed proceedings to obtain website-blocking orders but to leave the ISPs to bear the costs of implementing the orders. Arnold J’s argument in Cartier was essentially the same as in the copyright cases, namely that (1) the defendant is a commercial enterprise which indirectly benefits from infringing use of its subscribers and (2) it seems implicit in the Union law that the cost is imposed on the intermediary. It has to be said, however, he did not completely “rule out the possibility of ordering the right holder to pay some or all of the implementation costs in an appropriate case” ( EWHC2714(Ch) ). In Cartier, Arnold J only extended this to trademarks. He thus proposed abandoning an inherent principle which has been that innocent third parties are compensated for the ‘inconvenience’. In Ashworth (Ashworth Hospital Authority v. MGN Ltd  HRLR 41 House of Lords [35–36].), the House of Lords made it clear that reimbursement of the costs is one of the important reasons why third parties can be burdened at all.
The Court of Appeal
On the appeal ( EWCA Civ 658), Kitchin LJ delivered a judgment for the court. He fully agreed with Arnold J’s judgment. On the jurisdiction side, none of the judges on the appellate bench disagreed. In relation to costs, however, one dissenting voice opened the way to the Supreme Court.
Kitchin LJ made three additional points to support Arnold J’s costs-splits: (1) the Union law requires that injunctions against intermediaries are possible, (2) these injunctions are the price for immunities that intermediaries enjoy under the E-Commerce Directive and (3) Norwich case-law is not a source of jurisdiction, but only an analogy. Jackson LJ joined him. Briggs LJ, however, dissented on this point arguing that: (1) injunctions against intermediaries are nothing that the equitable jurisdiction couldn’t and wouldn’t achieve on its own – the Union origin is thus irrelevant, (2) the Union law does not prevent the reimbursement of the costs, (3) Norwich and Bankers Trust case-law embodies a more general principle constraining the exercise of equitable jurisdiction applicable to those who owe equitable obligations and (4) the “cost of doing business” argument is hardly convincing because the logic equally works under the Norwich disclosures. In his view, this points to the following principle: “compliance by an innocent party with an equitable duty to assist the victim of a wrongdoing should generally be at the victim’s expense.” (para 210).
The Supreme Court
At the UK Supreme Court (Cartier International AG and others (Respondents) v British Telecommunications Plc and another (Appellants)  UKSC 28), only the issue of costs was directly considered. However, the principle that the court eventually spells out is inextricably linked to equitable jurisdiction over innocent third parties. Lord Sumption, with whom all other Lords agreed, concluded that ISPs should indeed be compensated for the inconvenience of helping out in the enforcement of rights which they do not infringe themselves. While doing so, he also confirms that injunctions against intermediaries are nothing that the equitable jurisdiction couldn’t and wouldn’t achieve on its own.
Lord Sumption argued that, as a matter of principle, “[a]s a matter of English law, the ordinary principle is that unless there are good reasons for a different order an innocent intermediary is entitled to be indemnified by the rights-holder against the costs of complying with a website-blocking order” (para 31). He emphasizes that “[i]t is critical to these conclusions that the intermediary in question is legally innocent. The appellants in this case are legally innocent because they are “mere conduits”. Different considerations may apply to intermediaries engaging in caching or hosting governed by articles 13 and 14 of the E-Commerce Directive, because these operations involve a greater degree of participation in the infringement, which is more likely to infringe national laws protecting intellectual property rights if the conditions of immunity are not satisfied” (para 37).
He bases his entire decision on the domestic doctrine. Like the Cour de Cassation (Société Française du Radiotéléphone v Union des Producteurs de Cinema, 1e Civ, 6 July 2017), he arrives at the conclusion that the provisions of the E-Commerce Directive “do not preclude the cost of the measures strictly necessary for preserving the rights in question … from being borne by the technical intermediaries”. While the French court would shift the cost onto intermediaries in the absence of specific entitlement, Lord Sumption observed that English law would require such circumstances to burden intermediaries. Although he attributes the outcome to the difference ‘in the starting point’, I think this might be convincing for the procedural and legal costs, but not for compliance costs (i.e. costs of technological implementation).
Lord Sumption finds that the Union law does not limit allocation of the costs, and convincingly dismantles the argument that such allocation is anyhow implicit in the EU law or case-law (as previously argued by Arnold J and Kitchin LJ). He points out that the CJEU only limits excessive burdening of intermediaries, and provides Art 3 as a general limit on such compensation. Since the principle requires reimbursement of the ‘reasonable costs’, he finds it not to be problematic from the EU law perspective (paras 30, 31, 36). Moreover, he strongly opposed the argument that the costs should constitute compensation for the presence of safe harbours (para 29). As he notes, access providers would not be liable even in their absence (para 30).
As readers might know from my previous work, I find Cartier’s approach correct. Not only because it follows clearly from the history of equitable jurisdiction, but also because it makes more economic sense and makes the litigation more effective. It lets IP owners decide with their feet (lawsuits) if the proposed measures are worth the costs. Moreover, it prevents instances of strategic behavior, where remedies against innocent third parties are used instead of remedies against infringers due to more favorable conditions. For IP owners, Cartier means that they might want to carry out more cost-benefit analysis before launching cases. However, since they are uniquely positioned to assess the benefits of such blocks, this is desirable. The copyright case-law might be revisited too, as the Cartier decision takes a general approach (para 7).
Arnold J’s review of my book, where I explained this position, rightly points out that the above is based on the assumption that the enforcement of IP rights does not at the same time advance public interest. This is entirely right. If the private litigation also advances a public cause (e.g. one can think of fake medicine), this makes a case for cost-sharing or discounts. However, to the extent that other public bodies have the power to stop such cases, it might be questionable whether to take it into account in the private litigation which can hardly be relied on systematically. Interestingly enough, Lord Sumption dedicates an entire paragraph of his decision to a debate about public/private benefits of such enforcement. He argues: “Website-blocking injunctions are sought by rights-holders in their own commercial interest. They are wholly directed to the protection of the claimant’s legal rights, and the entire benefit of compliance with the order inures to the rights-holder. The protection of intellectual property rights is ordinarily and naturally a cost of the business which owns those rights and has the relevant interest in asserting them. (..) Intellectual property rights are created by law as a reward for innovation and enterprise which confer wider public benefits. But I cannot see that this makes any difference to the analysis. It supplies the reason why the rights exist, but the public interest in their enforcement is not wider or different from the private interest of the rights-holders.”
To understand this sentiment, one has to come back to the nature of the duty. As Lord Sumption notes, “I suggested in Singularis Holdings Ltd v PricewaterhouseCoopers  AC 1675, para 22, that the duty to assist identified by Lord Reid was not a legal duty in the ordinary sense of the term. As Lord Reid himself put it in Norwich Pharmacal, the intermediary came under the duty without incurring personal liability. This is really only another way of saying that the court had an equitable jurisdiction to intervene.” (par 11) As a consequence, “[t]here is no legal basis for requiring a party to shoulder the burden of remedying an injustice if he has no legal responsibility for the infringement and is not a volunteer but is acting under the compulsion of an order of the court.” (para 33)
What this entire debate shows is that the costs are inextricably linked to the very jurisdiction to grant the measures. The allocation creates an impression of the conceptual context of the jurisdiction, and its daily use is influenced by it. Therefore, to get the injunctions against innocent third parties right, we need to get the costs right too. Other European countries might take Cartier as an opportunity to reflect on their own systems, which are often the result of a formalist inertia, rather than intentional design.