By Royal Appointment: No Closer to an EU Private International Law Settlement?

Members of the British Royal Family and aristocracy have long contributed to the development of the law in England governing matters of personal privacy. As long ago as 1849, Prince Albert, the prince consort of Queen Victoria, resorted to the courts to prevent the publication of etchings and drawings by the Royal couple, including of their children (Prince Albert v Strange (1849) 2 De G & Sm 652). In a 1964 case, the Duchess of Argyll sued her formal husband, the 11th Duke, to prevent disclosure of the secrets of their marriage to national newspapers (Argyll v Argyll [1967] Ch. 302). In recent years, both Her Majesty the Queen and Prince Charles, Prince of Wales, have taken legal action in the English courts following the disclosure, or threatened disclosure, of personal information.

The recent flurry of judicial activity following the unwarranted invasion of the privacy of Her Royal Highness Princess William, Duchess of Cambridge, Countess of Strathearn and Baroness Carrickfergus (a.k.a. Mrs Mountbatten-Windsor) highlights the potential advantages for claimants of French privacy laws, both civil and criminal. No doubt, the Duchess and her husband wished to be seen to have taken prompt and effective action to protect their private lives in this high profile case pour encourager les autres. Their chosen avenues of recourse through the French courts would appear to have been designed to serve both as a swift, effective and public assertion of their rights (the civil injunction) and as a deterrent (the nascent criminal complaint).

As yet, the incident and its aftermath do not seem momentous from a private international law perspective. The prosecution by English nationals of a civil claim in France against a French publisher, requiring the delivery up of photographs in the publisher’s possession which are said to have resulted from an invasion of the claimant’s privacy on French territory, would not appear to raise significant or complex issues of jurisdiction or applicable law.

Nevertheless, the case encourages reflection as to how well EU private international law deals with situations involving (alleged) violations of personal privacy, and other contributors to this symposium have raised a variety of issues.

Two introductory points may be noted before embarking on further discussion of this topic. First, and putting to one side the need to provide an autonomous definition in an EU context (see below), one must accept that the notion of a “violation of privacy” may in common usage cover a wide variety of fact situations, which are not necessarily to be treated alike. Taking the facts of the Duchess of Cambridge case as an example, the essence of any judicial complaint could rest upon the unauthorised (i) taking, (ii) transmission, (iii) receipt or (iv) publication of photographs or other media, with any transfer or publication occurring either (a) electronically (including via the internet) or (b) by other means. In other circumstances, a violation of personal privacy may be tantamount to a physical assault, as in the case of stalking, or to theft, as in the case of the removal of papers (the Pontiff’s butler) or computer hacking. The matter may also have a commercial background, in particular if the claimant intended himself to exploit the disclosed information, as in the Douglas-Zeta Jones wedding case (Douglas v Hello! Limited [2007] UKHL 21).

Secondly, if it is determined that any or all of these situations do require special treatment within EU private international law instruments, one must recognise that that this will inevitably create problems of classification, which may be thought to compromise the underlying objectives of promoting legal certainty, and harmonious decision making, that these instruments outwardly pursue.

EU law has already shown itself to be adept in creating difficulties of this kind. In the Rome II Regulation, non-contractual obligations arising out of violations of privacy (and of personality rights) are presently excluded altogether (Art. 1(2)(g)), but the task of elaborating what wrongful conduct amounts or does not amount to a “violation of privacy” for this purpose has been left to the courts, and remains incomplete. Following criticism levelled at this exception, there have been (as Professor von Hein explains) various proposals for a new, special rule covering the same ground as the current exclusion. If adopted, however, the new rule would not remove the classification problem, but merely transfer it from being one of the material scope of the Regulation to one of the material scope of a rule within the Regulation, and its separation from other rules (in particular, the general rule for tort/delict in Art. 4).

In relation to online activities, the eCommerce Directive raises many (as yet unresolved) issues as to the scope of its “country of origin” regulation, and the various exceptions and qualifications to that regime. The European Court’s eDate Advertising / Martinez decision, rather than clearing the air, has only heightened the challenges that this Directive presents in the area of civil liability.

Last but not least, the eDate decision also has a separate jurisdictional aspect, on which the remainder of this comment will focus. The effect of this part of the Court’s judgment is that a distinction must now be drawn for jurisdiction purposes between “an infringement of a personality right by means of the internet” (which the CJEU has told us merits a special, claimant-friendly interpretation of Art. 5(3)) and other cases (which remain subject to well-established principles governing the operation of that Article).

At first impression, these two points may seem to pull in different directions, the first supporting a more granular approach and the second tending towards a uniform solution. Both, however, provide reasons for caution when formulating special rules, whether of jurisdiction or applicable law, which treat violations of privacy and personality rights as a single, separate category. Further, the proliferation of different fact patterns within the realm of “violations of privacy” and analogies to other categories of wrongdoing (such as those highlighted above) may itself be thought to militate in favour of maintaining general rules such as Art. 5(3) of the Brussels I Regulation in its pre-eDate form and Art. 4 of the Rome II Regulation. The latter provision, in particular, may be argued to be sufficiently well-calibrated to deal with the range of new situations that would fall within its scope if the Art. 1(2)(g) exception were simply to be removed when the Regulation is reviewed.

In his contribution, Professor von Hein supports the adoption of a special rule for violations of privacy and personality rights. As part of his proposal, he favours giving claimants who sue in the courts of their own habitual residence or of the defendant’s domicile a right to elect to apply the law of the forum to the entire claim.

This element of Professor von Hein’s proposal seeks to build upon the jurisdictional aspect of the CJEU’s decision in eDate. This, however, is the law reform equivalent of constructing a house on swampland. The decision has strong claims to be the worst that the Court has ever delivered on the Brussels I regime, conflicting with long established principles central to the functioning of the Regulation and giving the impression either that the Court considers itself at liberty to make up new rules of jurisdiction on the spot or that there is a sacred text in its library in which the Regulation’s rules are elaborated, but to which the outside world does not yet have access.

The decision may be criticised in no less than seven respects.

First, having expressed ubiquitous remarks about the ubiquitous nature of internet publications (para, 45), the Court observed (with good reason) that this causes difficulty in applying the criterion of “damage” as a factor connecting the tort to a given legal system for the purposes of Art. 5(3) of the Regulation: “the internet reduces the usefulness of the criterion relating to distribution in so far as the scope of the distribution of content placed online is in principle universal” (para. 46). In light of these conclusions, and given that the special rules of jurisdiction are intended to secure “a close link between the court and the action” and/or “to facilitate the sound administration of justice” (Recital (12); see also para. 40 of the eDate judgment), one might have expected that the Court would conclude that the concept of “harmful event” should be given a narrow reading in cases of this kind so as to exclude the criterion of damage as a connecting factor for jurisdiction purposes (for an analogous approach in a contractual context, see Case C-256/00, Besix, paras 32 and following). That conclusion would have been consistent with the dominant approach in the case law to the interpretation of exceptions to the general rule in Art. 2 (e.g. Case C-103/05, Reisch Montage, paras 22 and 23). The Court, however, chose a different path.

Secondly, the Court asserted that the connecting factors used within Art. 5(3) “must therefore be adapted in such a way that a person who has suffered an infringement of a personality right by means of the internet may bring an action in one forum in respect of all the damage caused” (para 48). This argument, which the Court uses as its launching pad for its novel “centre of gravity approach”, is utterly devoid of merit. As the Court had acknowledged (para. 43), the claimant in such a case already has at least one, and possibly, two options available for bringing an action in respect of all the damage caused in one Member State court. Most significantly within the framework of the Regulation, he/she may always bring an action in the Courts of the defendant’s domicile (see Besix, para 50; Case C-420/97, Leathertex, para 41). Moreover, if the publication emanates from an establishment in a Member State other than that of the publisher’s domicile, the claimant may bring an action in that Member State, as the place of the event giving rise to damage, (Case C-68/93, Shevill, paras 24-25; eDate, para. 42; Case C-523/10, Wintersteiger, paras 36-39). There was no need to create a new global connecting factor.

Thirdly, having concluded that the Regulation did not present the claimant with sufficient options for pursuing his claim, the Court proposed attributing full jurisdiction to “the court of the place where the victim has his centre of interests” on the ground that the impact of material placed online might best be assessed by that court (para. 48), sitting in a place which corresponds in general to the claimant’s habitual residence (para. 49). In these two sentences, and without further explanation or justification, the Court repudiates its longstanding principle of avoiding interpretations of the rules of special jurisdiction in Art. 5 which favour the courts of the claimant’s domicile in such a way as to undermine to an unacceptable degree the protection which Art. 2 affords to the defendant (e.g. Case C-364/93, Marinari, para. 13; Case C-51/97, Réunion Européenne, para. 29).

Fourthly, the Court considered that its proposed new ground of jurisdiction has the benefit of predictability for both parties, and that the publisher of harmful conduct will, at the time content is placed online (being, apparently, the relevant time for this purpose†), be in a position to know the centres of interests of the persons who are the subject of that content (para. 50). It is, however, extremely difficult to reconcile this confident statement with the Court’s earlier recognition that “a person may also have the centre of his interests in a Member State in which he does not habitually reside, in so far as other factors, such as the pursuit of a professional activity, may establish the existence of a particularly close link with that State” (para. 49). If predictability were the objective, it is hard to see how the Court could have done more to remove it.

Fifthly, given that a person’s private life (and reputation) may have several centres, which change over time, it does not seem possible to say more than that there might be a strong link between the facts of a particular case and the place where the claimant’s centre of interests is held to lie. Equally, there might not. Take the case of a former Bundesliga footballer, with Polish nationality, who signs for an English club and moves to England. While visiting a German friend, he has rather too much to drink in a nightclub. The story is published, in German, on a German football website. Does the sound administration of justice support giving the English courts jurisdiction over the footballer’s claim against the website publisher? In the Duchess of Cambridge’s case, does the sound administration of justice support giving the English courts jurisdiction over the publication of photographs on a French, or Italian or Irish, website, particularly as the current position is that those courts would have no jurisdiction with respect to hard-copy publications by a newspaper or magazine under the same ownership? Given that the French, Italian or Irish courts would have global jurisdiction under Art. 2, it is suggested that the answer is a resounding “no”.

Sixthly, having decried the utility, in internet cases, of the criterion of damage á la Shevill, the Court inexplicably chose to retain it as a connecting factor for jurisdiction purposes, allowing an action “in each Member State in the territory of which content placed online is or has been accessible” (para. 51). This begs the following question: if the new connecting factor is not a substitute for the “damage” limb of the Bier formulation, what then is it? In para. 48 of its judgment, the Court had seemed to suggest that the claimant’s centre of interests was “the place in which the damage caused in the European Union by that infringement occurred”, but this cannot be taken literally given that the Court returns three paragraphs later to the view that damage may occur in each Member State. The eDate variant of “damage” would seem to be a derivative or indirect form, of the kind that the Court had in its earlier case rejected as being a sufficient foundation for jurisdiction (Marinari, para. 14). If a label is needed, perhaps “damage-lite” would do the job?

Finally, the Court’s assertion that its new rule corresponds to the objective of the sound administration of justice (para. 48) is also called into question by the second part of its judgment, interpreting the eCommerce Directive in a way that gives an essential role in cases falling within its scope to the law of the service provider’s (i.e. the defendant’s) country of origin. Although questions of jurisdiction and applicable law are distinct, and the Brussels I Regulation and eCommerce Directive pursue different objectives, the suitability of the courts of the claimant’s centre of interests is undermined by the need to take into account, in all cross-border cases, a foreign law. By contrast, jurisdiction and applicable law are much more likely to coincide where jurisdiction is vested in the courts of the defendant’s domicile or establishment.

Any proposed new rule in the Rome II Regulation must also face the complexity which the eCommerce Directive introduces in this area, particularly after the eDate judgment. In an ideal world, the priority between the two instruments would be reversed, with the Directive being pruned to exclude its effect upon questions of civil liability and to enable a single instrument to govern questions of the law applicable to non-contractual obligations arising out of violations of privacy and personality rights. That, however, may be too much to hope for – once embedded, an EU legislative instrument is hard to dislodge.

Professor Muir-Watt makes the important point that, in this area, choice of law rules must yield, to a greater degree than in many other areas of civil law, to considerations of public policy and to the fundamental rights to which all Member States subscribe as parties to the European Convention (we will have to agree to disagree about the significance of the Charter of Fundamental Rights even if the Rome II Regulation were extended).

In cases such as that of the Duchess of Cambridge, there is of course a tension between (at least) two rights – that of the right to a private and family life (Art. 8) and that of freedom of expression (Art. 10). As recent cases before the European Court of Human Rights demonstrate (in particular, the two decisions involving Caroline, Princess of Monaco), the balance between them is not easy to strike, and the margin of appreciation will continue to allow different solutions to be adopted in different States. It may be questioned, however, whether this perilous balance is well served by a rule of election for applicable law which, coupled with claimant friendly rules of jurisdiction, enables the subject of a publication which is alleged to be defamatory or to violate privacy to choose to apply to the whole of his claim either the law of his country of habitual residence or the law of the defendant’s domicile, whichever is the more favourable. This, unlike environmental damage (Rome II Regulation, Art. 7) is not an area where the policy factors favour an overwhelmingly pro-claimant approach.

Enough said. To offer a personal view in conclusion: the best way forward would be  (1) to amend the Brussels I Regulation to reverse the eDate decision, (2) to carve civil liability out of the eCommerce Directive, and (3) to remove the exception for violations of privacy and personality rights in Art. 1(2)(g) of the Rome II Regulation, leaving the general rule for tort/delict (Art. 4) to apply to such cases. At the same time, it seems more likely that my own daughter will marry into the Royal Family than that these three reforms will come to fruition. Princess Nell anyone?


† Straying into the detail of Professor von Hein’s rule of election, one consequence of this would appear to be that the claimant’s habitual residence and the defendant’s domicile would be tested by reference to a different point in time (the latter being identified at the date of commencement of proceedings). This is not a reason in itself to reject the rule.