Party autonomy and the applicable law to the merits in intellectual property arbitration

Abstract

The freedom of the parties to choose the applicable law to the merits is often presented as an important benefit of arbitrating disputes involving intellectual property (IP). Yet, the reality is more uncertain and controversial than is commonly assumed. Is party autonomy really permitted in IP arbitration? Should it be? This article answers these questions with regard to patents, trademarks, and copyrights, using recent examples drawn from arbitral practice. It first examines the situation where the parties only made a choice of contract law, and considers in this regard whether that law can and/or should be extended to infringement claims and/or the remedies to infringement, either directly or through the technique of characterization. After discussing the impact of overriding mandatory rules in contractual IP cases, the article then examines the situation where the parties chose a law to govern IP questions, including by way of a broad choice-of-law agreement covering non-contractual problems. It seeks to provide as much guidance as possible to arbitral tribunals, using mainly the distinction among infringement, ownership, and validity issues, and by distinguishing among different scenarios in which party autonomy is more or less acceptable.

Over the last 20 years, arbitration has become a popular mechanism to resolve intellectual property (IP) disputes. Although precise evidence is lacking, the caseload of both generalist and specialized arbitral institutions suggests that IP-related disputes are routinely arbitrated around the globe.¹ In order to cater to the rise of IP arbitration, new arbitral institutions dedicated to IP disputes have been set up,² while some of the world’s leading institutions have recently created a pool of specialized IP arbitrators.³

Part of the reason for this increased popularity is that the benefits of arbitration are particularly appealing in the IP context. In addition to the ease of enforcement of arbitral awards and the possibility to select arbitrators with relevant expertise, arbitration allegedly provides special benefits where the dispute involves parallel IP rights, ie, IP rights arising in different jurisdictions over the same subject matter.⁴ As states increasingly recognize the arbitrability of IP issues (including the infringement, ownership, and validity of IP rights), arbitration may allow the consolidation of disputes over parallel IP rights in a single forum, unlike in IP litigation, where the parties may have to duplicate proceedings in each country where an IP right arises.⁵ While there remains some uncertainty over arbitrability, particularly with regards to the validity of IP rights and the applicable law to arbitrability,⁶ it is clear that arbitration increasingly provides a way to avoid the costs of litigating in multiple jurisdictions.

This article focuses on a separate, though related benefit of IP arbitration—namely, the parties’ freedom to choose the law applicable to the merits of the case. According to many writers, arbitration allows the parties to choose the applicable law to any IP issue, even if that law is not that of the country in which the relevant IP right arises. To quote a recent practitioners’ guide on IP arbitration,

In principle, disputes arising from [registered] IP rights […] can be resolved [by international arbitrators] according to a law that is not the law of the jurisdiction where the IP right is registered. Choosing the substantive law of a single jurisdiction could simplify and streamline the proceedings, and avoid costly piecemeal application of multiple national IP laws.⁷

Comparable statements can be found in other publications, including a recent WIPO report on private international law and IP.⁸ The precise formulation varies, but many IP arbitration specialists agree that choice-of-law agreements are permitted for all IP rights, regardless of the issue at stake.⁹ Thus, it is believed that parties to arbitration proceedings can not only choose the applicable law to contractual issues, such as the performance or interpretation of a licence agreement. Rather, they can also resolve IP issues such as the infringement, ownership, and/or validity of parallel IP rights under the law of their choice. Some writers reserve the application of ‘mandatory rules’,¹⁰ but few venture to provide examples. And if they do, it is often in little detail and/or to emphasize that party autonomy is permitted ‘in most cases’.¹¹

To anyone familiar with private international IP law, the freedom of the parties to choose the applicable law is likely to be surprising. Before domestic courts, IP issues are typically governed by the law of the country for which protection is sought (the lex protectionis), which generally coincides with the law of the country where infringement is alleged to have taken place (the lex loci delicti), and—in the case of registered IP rights—the law of the registration country.¹² Applying the lex protectionis is generally considered as mandatory: in the EU, for example, Article 8(3) of the Rome II Regulation prohibits choice-of-law agreements in infringement cases.¹³ A handful of legal systems permits the parties to choose—after the dispute has arisen—the lex fori to IP infringement,¹⁴ or at least the remedies to infringement.¹⁵ But virtually no jurisdiction authorizes choice of law agreements with regards to the ownership and/or validity of IP rights, whether registered or unregistered.

The reason for this hostility against party autonomy is not always clearly stated. But it is often linked to the idea of ‘territoriality’,¹⁶ which is so often described as being pervasive in IP. As IP statutes only provide protection against infringing conduct located in the enacting country’s territory, party autonomy would allow the law of the chosen country to regulate infringement acts located in a foreign country. Considering that IP rights are closely tied to the public interest, insofar as they are generally viewed as an instrument to encourage innovation, creativity, and economic growth,¹⁷ respect for the territorial limitation of IP rights and the lex protectionis rule is generally considered necessary to safeguard the policy choices of sovereign countries.

Much has already been said about the problems created by this approach in our interconnected world, in which disputes frequently involve parallel IP rights, and thus the application of numerous laws. Various solutions were put forward to resolve this ‘multiplicity’¹⁸ problem, including letting parties choose the law governing the remedies for infringement in cross-border IP litigation.¹⁹ But as progress has remained limited so far, right holders must often narrow down the scope of their claims so as to only cover one or few countries,²⁰ even where claims based on parallel IP rights could have been consolidated in a single forum. These difficulties contribute to explaining the increasing attraction of arbitration, where tribunals can allegedly resolve multijurisdictional IP disputes under a single or limited number of laws.

But are the parties to an IP arbitration really free to choose the applicable substantive law? Arguably, the reality is more uncertain and controversial than is commonly assumed. There is in fact almost no evidence, whether in legislation, arbitral awards, or court decisions, that party autonomy is permitted for IP issues. Nor is there any agreement in the literature, as some writers firmly reject party autonomy, often because of the alleged mandatory character of IP statutes,²¹ the public interests safeguarded therein,²² or the ‘territoriality principle’.²³ This disagreement makes it difficult for practitioners to advise clients in drafting choice-of-law clauses, and on whether the alleged benefits of IP arbitration reflect reality. Similarly, arbitral tribunals may struggle to determine the limits of party autonomy. Although some writers have recently attempted to clarify the applicable substantive law in IP arbitration,²⁴ including the role of party autonomy, previous studies are either insufficiently comprehensive, too general, or they consider applicable law matters in a way that is almost entirely disconnected from real cases.

This article will therefore add to the existing literature in further exploring whether the parties can and/or should be able to choose the applicable law in arbitrations involving IP. It focuses on the three main IP rights, ie, patents, trademarks, and copyright.²⁵ This analysis is not only relevant for arbitral practice: it also has broader significance for the future of arbitration as a method of resolving cross-border IP disputes. Given the link between IP and the public interest, states may not tolerate arbitral tribunals disregarding their IP statutes if the dispute falls within their scope of application. Should tribunals blindly enforce choice-of-law agreements, states may be tempted to re-enact arbitrability restrictions and/or mandate courts to scrutinize the merits of arbitral awards, just like German courts must now do for awards relating to competition law.²⁶ Although the trend in favour of arbitrability is unlikely to be reversed, the question of party autonomy is thus closely tied to the reputation and future development of the arbitral process, which is likely to depend on whether arbitrators can strike a satisfactory balance between the parties’ desire to simplify the proceedings and respect for domestic IP policies.

To get a clear view of the problem, two situations should be distinguished. The first is where the parties’ choice of law agreement is narrowly drafted, eg, to provide that ‘the contract shall be governed by the laws of state X’. This situation does not directly concern choices of IP law but must be examined carefully because it is the most frequent in arbitral practice and gave rise to a number of arbitral awards which, due to their recent publication, have not been included in previous studies.²⁷ These awards notably reveal that tribunals were sometimes invited to extend the law governing the parties’ contract (the lex contractus) to infringement claims and related remedies, whether directly or through the technique of characterization. After further setting the scene, Part I will examine this question insofar as it relates to the scope of party autonomy. It will also unravel examples of overriding mandatory rules in contractual IP cases, in order to fill the gap left by previous studies.

The second situation is where the parties have chosen a law to govern IP issues, whether by making a direct choice or by way of a broad choice-of-law agreement covering extra-contractual problems. While rare, this situation is not unheard of and is the most controversial. Hence, it is necessary to consider whether arbitral tribunals should uphold or disregard choice-of-law agreements in this scenario. Although reaching definitive conclusions is often made impossible by the lack of publicly available practice, Part II will attempt to provide as much guidance as possible to arbitral tribunals, using mainly the distinction among infringement, ownership, and validity issues, and by distinguishing among different scenarios in which party autonomy is more or less acceptable.

I. THE SCOPE OF PARTY AUTONOMY IN THE EVENT OF A CHOICE OF CONTRACT LAW

Empirical evidence suggests that parties to arbitration proceedings usually select an applicable law to govern their contract,²⁸ including in IP-related agreements such as licences and assignments.²⁹ And under virtually all arbitration laws and rules, there is no doubt that parties can do so,³⁰ whether expressly or tacitly, by formulating arguments under the same law.³¹ Parties may choose one—or indeed several³²—domestic law(s), and they may even choose transnational principles, even if they seldom do so.³³ Choice-of-law agreements may be concluded before or after the dispute arises, and the choice may be registered in a Procedural Order or, in ICC Arbitrations, the Terms of Reference.³⁴

Most of the time, choosing the lex contractus is sufficient to avoid any applicable law problems. As arbitrations involving IP rights typically stem from a pre-existing contractual relationship, disputes often revolve around the interpretation, performance, termination, or validity of the parties’ agreement. In such cases, arbitral tribunals can—and indeed must—apply the law or rules of law chosen by the parties. Thus, in practice, the applicable law is often uncontested.³⁵ The choice-of-law reasoning of arbitral tribunals tends to be very short; sometimes, it is literally a footnote.³⁶

However, applicable law matters can become more complex where a case requires the application of both contract and IP law. Licensees or assignees may allege, for example, that a breach of contract claim should fail because the underlying IP right is invalid, or not owned by the party who licensed and/or assigned it. Interpreting patent claims may be necessary to determine whether the sale of certain products triggers the payment of royalties under an agreement; the validity of a copyright assignment may depend on whether the right at issue is assignable;³⁷ and—among other examples³⁸—trademark law may inform whether a trademark is used ‘in commerce’ in the context of a breach of contract allegation.

Depending on the manner in which the claim is framed, the need to refer to IP law may be avoided. Indeed, right holders may be able to rely on a contractual clause prohibiting a licensee or assignee from using a right outside a defined territory, in a certain way, after a limited period, or post-termination of the agreement. For instance, in an ad hoc award, a licensor was alleging that a licensee had assisted a third party to use US trademarks, copyrights, and trade secrets that were the object of a licence agreement governed by the law of New South Wales, Australia.³⁹ While the licensee referred to the need to examine US IP law regarding any infringement allegation, the licensor had framed its claim based on a contractual clause prohibiting the conduct in question.⁴⁰ Consequently, the tribunal was able to resolve the dispute based on contract law and the wording of the agreement, without the need to delve into US IP law.⁴¹

Nevertheless, right holders may prefer to allege (non-contractual) IP infringement in order to bypass a contractual and/or statutory limitation on liability, or to claim punitive damages or an account of profits, two remedies that are not usually available under contract law.⁴² In fact, the WIPO Arbitration and Mediation Center recently reported an increase in non-contractual claims,⁴³ an increase which may be linked to the fact that more and more cases are referred to WIPO Arbitration after the dispute has arisen, instead of a pre-existing arbitration agreement.⁴⁴ Thus, it is entirely possible and ever more likely for an arbitration to involve elements of both contract and IP law.

(A)  Scope of the parties’ choice of law agreement

In such cases, the first point to consider is the scope of the parties’ choice of law agreement, if any. This generally depends on the wording of the agreement and the intention of the parties.⁴⁵ Two scenarios must be distinguished:

• The choice-of-law agreement is broadly drafted, eg to cover any issue ‘relating to’ or ‘arising in connection with’ the contract (mirroring the wording often used in arbitration agreements),⁴⁶ or to clarify that any dispute between the parties ‘shall be resolved by the law of state X’ (as the Terms of Reference sometimes provide in ICC arbitrations). In this scenario, the chosen law would normally cover both contractual and IP claims,⁴⁷ and the question becomes whether party autonomy is allowed for the latter. I will return to this question in Part II.

• The choice-of-law agreement is drafted narrowly, eg ‘the contract shall be governed by the laws of state X’. In this situation, which appears to be the most common,⁴⁸ the choice-of-law agreement would presumably not encompass IP issues as these issues are non-contractual.⁴⁹

(B)  Extending the lex contractus to IP issues?

Assuming a narrow choice of law agreement, the problem becomes one of characterization: which issues are contractual, and which are non-contractual? Depending on the tribunal’s characterization, the scope of the law chosen by the parties—and accordingly of party autonomy—would differ. This is because, while contractual problems are inevitably governed by the lex contractus, there is broad agreement that IP questions should be governed by the lex protectionis, at least as far as registered IP rights are concerned.⁵⁰ And with some exceptions,⁵¹ most tribunals apply that law to IP issues, without much discussion.⁵²

For example, in ICC Case n° 18892 (Bayer Cropscience v Dow Agrosciences), an arbitral tribunal sitting in Indianopolis denied that French law, the law governing the parties’ licence agreement, could influence the infringement of US patents,⁵³ or indeed estop the licensee from challenging patent validity.⁵⁴ Even assuming that the parties could somehow ‘contract out of [US law]’, the choice-of-law agreement was ‘far too narrow to achieve this purpose’ as it only provided that the contract were to be ‘governed by and construed in accordance with the laws of France’.⁵⁵ Surprisingly, the tribunal examined whether the challenge to the validity of US patents was time-barred under French law,⁵⁶ but this was not the case and in any event, the tribunal confirmed that French law could not ‘prevent the raising of a defense to a patent infringement claim under US law’.⁵⁷

Strictly speaking, arbitral tribunals are not under an obligation to apply the lex protectionis. Absent agreement of the parties, arbitration laws and rules generally allow tribunals to apply the law or the conflict rules they deem most appropriate,⁵⁸ or sometimes the law with the closest connection to the dispute.⁵⁹ Relying on this freedom, the claimant in the Bayer arbitration attempted to extend the lex contractus to the remedies for patent infringement.⁶⁰ Although the tribunal applied US law to the infringement of US patents, the claimant sought an order for the cessation and destruction of the respondent’s products based on French tort law (in addition to unsuccessfully claiming the same remedy under French contract law and US patent law⁶¹). This was because, for the claimant, France was the ‘center of gravity of the dispute’ and tortious acts had to ‘be dealt with under the law with the closest nexus to the pre-existing relationship between the parties’.⁶² The argument was wrongly based on the Rome II Regulation,⁶³ but it would have received stronger support in several academic proposals advocating, albeit with some nuances, the application of the lex contractus to the remedies for infringement in cases arising from a pre-existing contractual relationship.⁶⁴ Should this approach receive acceptance in IP arbitration, the scope of party autonomy would be significantly enlarged, since many IP arbitrations arise from a pre-existing contractual relationship between the parties.

To the extent that the tribunal applied US law to the infringement of US patents, applying a different law to the remedies would have seemed illogical.⁶⁵ Unsurprisingly, the arbitral tribunal declined to apply French law as (i) the choice-of-law agreement was narrowly drafted, (ii) the Rome II Regulation led to the lex protectionis, not the lex contractus, and (iii) in the absence of party agreement, it was appropriate to apply US patent law to all non-contractual issues, not least because the dispute had no connection to France apart from the fact that French law was the lex contractus.⁶⁶

The tribunal’s reasoning makes sense: not only would extending the lex contractus to the remedies for infringement (or indeed any IP issue) be inconsistent with the parties’ intention in the event of a narrow choice-of-law agreement, but it would also be unjustified where the country chosen to regulate the contract has few or no geographical connections with the dispute.⁶⁷ Applying the lex protectionis to IP issues is generally preferrable, as it preserves the ability of states to tailor their IP regimes to their local needs and philosophy as much as possible. In turn, as will be further explained in Part II, the rule diminishes the risk of non-enforcement and/or annulment of the award on grounds of public policy.⁶⁸ Indeed, should a tribunal grant an injunction in favour of the IP right holder, enforcement of the award would need to be requested in the country of IP protection. Applying the lex protectionis, as the tribunal did in Bayer, facilitates enforcement since the tribunal would apply the law of the country under which compliance of the award with public policy is most likely to be examined.

To be sure, extending the lex contractus to the remedies to infringement may be tempting if the case concerns parallel IP rights. Indeed, this would allow remedies to be governed by a single law. But the Bayer arbitration only involved US patents, so extending the lex contractus was unnecessary. While doing so would have allowed the same law to govern both contractual and infringement remedies, this benefit was limited because the tribunal already applied US law to patent infringement (and validity), so it was equally efficient and far more appropriate to apply US law to the remedies for infringement.⁶⁹

Extending the lex contractus to infringement remedies would have been a more attractive solution if that law could have also governed other IP issues such as patent infringement. However, as with remedies, this would have run counter to the parties’ expectations in the event of a narrow choice of law agreement, while being difficult to justify if the legal system chosen to regulate the contract had few or no connections to the dispute. It would also have created public policy concerns, at least in some scenarios detailed in Part II.⁷⁰ Overall, arbitral tribunals should carefully separate contractual from IP issues, including infringement and related remedies, except where the lex contractus and the lex protectionis lead to the same substantive outcome.⁷¹

(C)  Extending the lex contractus through the technique of characterization?

The question of characterization in arbitration differs from litigation, where domestic courts generally apply the lex fori to characterize the dispute. As arbitrators have neither a forum nor a lex fori, they do not have to characterize the issue according to the requirements of the law of the seat of the arbitration (the lex loci arbitri).⁷² They can instead rely on any domestic law and/or characterize the dispute in an autonomous way, should they wish to do so.⁷³ Indeed, characterization is part of the tribunal’s decision on the applicable law, which domestic courts do not review, subject to compliance of the award with public policy.

In theory, arbitral tribunals are therefore relatively free to characterize an issue as contractual, even though domestic courts would have classified the same problem within IP law. In Bayer v Dow, this indirect extension of the lex contractus did not materialize, as the tribunal rigorously distinguished contractual issues from questions of patent validity and infringement.

However, the award rendered in ICC Case n° 21397 (Benihana of Tokyo LLC v EHP Pty Ltd⁷⁴) suggests that tribunals do not always characterize issues very wisely, and in doing so may wrongly extend the scope of party autonomy. In that case, the claimant had licensed IP rights—including Australian trademarks—to operate a ‘Benihana of Tokyo’ restaurant in Australia.⁷⁵ As the respondent allegedly failed to pay royalties, among other breaches, the claimant terminated the agreement and brought arbitration proceedings to obtain compensation. The seat of the arbitration was New York,⁷⁶ and the licence agreement was governed by New York law.⁷⁷ During the proceedings, the claimant alleged that the respondent continued to use the trademarks post-termination, and framed his claim on two grounds: breach of contract, and trademark infringement. The alleged facts supporting the two claims were the same: not only did the respondent fail to cancel registration of the ‘Benihana of Tokyo’ business name for some time, but the trademarks also appeared on webpages operated by the respondent and third parties,⁷⁸ and the respondent’s Facebook and TripAdvisor pages contained a number of reviews from the period before termination.⁷⁹

With regard to the trademark infringement claim, the claimant argued that because the claim ‘sound[ed] in’ or ar[ose] under’ the licence agreement, it was ‘properly governed by the contract’s choice of law clause calling for the application of New York substantive law, which includes New York’s common law on trademarks and the US federal Lanham Act’.⁸⁰ The respondent, on the other hand, contended ‘that the proper law of a claim for statutory compensation for breach of trademark infringement [was] the law of Australia, given that both marks licensed [under the agreement were] registered in Australia and neither [was] registered in the United States’.⁸¹ The respondent also denied that US law could have any extraterritorial application, and suggested that US law could only prevent the use of a US trademark, as opposed to an Australian trademark.⁸²

The respondent clearly had a point. Indeed, it is almost axiomatic that the infringement of an Australian trademark is governed by Australian law. Yet, the tribunal sided with the claimant and held—without citing any authority—that ‘where […] a claim involves rights that arise out of a contract and the parties’ entire business relationship stems from that contract, […] such claim “sounds in” contract and is properly subject to a choice-of-law provision contained therein’.⁸³

Arguably, the tribunal’s reasoning is flawed: as noted, the claimant had distinguished the breach of contract and infringement claims. The latter sounded in tort, even though infringement arose in the context of a pre-existing contractual relationship. Thus, the tribunal mischaracterized the claim and, in doing so, wrongly extended the lex contractus (New York law) despite the narrow wording of the choice-of-law agreement.⁸⁴ While this approach allowed the tribunal to apply a single law to the whole dispute, extending the lex contractus to IP issues may lead to extraterritoriality, as the law chosen by the parties could potentially apply to infringement acts committed outside the enacting state’s territory. As will be further explained in Part II, such extraterritoriality is unlikely to be problematic per se but would create a risk of non-enforcement and/or annulment of the award for breach of public policy.⁸⁵

This risk may explain why the tribunal continued its reasoning as follows. After noting that ‘New York substantive law on trademark infringement include[d] both New York common law and the US federal Lanham Act’, which created the same infringement standard,⁸⁶ the tribunal considered whether New York law could be applied to ‘conduct carried out in a foreign jurisdiction by a foreign defendant’.⁸⁷ The answer was negative based on the traditional test on the extraterritorial application of US trademark law, which focuses on (i) whether there is a substantial effect on US commerce, (ii) the nationality of the defendant, and (iii) the existence of a conflict with foreign trademark rights.⁸⁸ The first factor was not met as there was no evidence that the wrongful activity caused or could cause confusion among American customers, or that the alleged infringement would have diminished the value of the Benihana of Tokyo franchise, whether in the USA or worldwide.⁸⁹ Moreover, the defendant was not American, such that considering the potential conflict between New York and Australian law was unnecessary.⁹⁰ Accordingly, the tribunal found that infringement was not established under New York law while observing—in a footnote—that the claimant ‘did not plead an alternative claim for trademark infringement under Australian trademark law’.⁹¹

Ultimately, the award was therefore consistent with the territorial limitation of trademark law. But the result was paradoxical. On the one hand, the tribunal sided with the claimant with regard to the applicable law. On the other hand, it dismissed the claim on the substantive level because the geographical scope of the applicable law did not cover infringement acts committed abroad. Of course, it was the claimant’s choice to only formulate its trademark action under New York law. Had it also relied on Australian law, the outcome may have been different. In a way, the tribunal’s approach was consistent with the logic of the lex protectionis, which requires the tribunal to apply the law of the country for which the claimant chooses to seek protection.⁹² However, the tribunal could and should have invited the parties to make submissions under Australian law (not least because the respondent was advocating its application). By way of comparison, the US courts would not have done so, as they have so far refused to apply foreign trademark law.⁹³ But contrary to what the claimant’s counsel seemed to believe,⁹⁴ arbitrators sitting in New York are not akin to local judges and can apply any law, nor do they lose jurisdiction if New York or US law do not apply or do not confer substantive protection due to their territorial limitation.

To be fair, the tribunal can hardly be criticized for only addressing the parties’ arguments. Perhaps the tribunal did not want to rule ultra petita. But in the end, the award might not have resolved the whole dispute between the parties, for the claimant could still bring a claim for infringement under Australian trademark law.⁹⁵ Indeed, that claim would most likely not be res judicata as the tribunal only ruled on the infringement of an unregistered US trademark, as opposed to a registered Australian trademark. To diminish the risk of subsequent proceedings, the tribunal should therefore have applied Australian law, either by characterizing the dispute as non-contractual, or by denying the applicability of New York trademark law due to its territorial limitation.⁹⁶

The Benihana of Tokyo award illustrates that the freedom of arbitral tribunals with regard to characterization can lead to unexpected outcomes. The better approach would be to apply the lex contractus to breach of contract claims, and the lex protectionis to determine the question of infringement, as the tribunal did in Bayer. If contrary to this recommendation, a tribunal were to extend the lex contractus to infringement matters (whether directly or through the technique of characterization), it would be preferable to apply the lex contractus extraterritorially, as the claim would otherwise be dismissed on the merits even though acts of infringement may have been committed somewhere. However, as will be further explained in Part II, it should be noted that the enforcement of the award could not be guaranteed if a tribunal decided to do so.

Apart from this example, one could easily find other potential scenarios creating characterization difficulties. In IP litigation, there is little doubt that the assignability or waivable character of copyright should be characterized as a copyright question, differing from the contractual question of the validity of an assignment.⁹⁷ Arbitral tribunals could follow this approach by analogy.⁹⁸ But there is more controversy with regards to termination rights, eg under section 203 of the US Code, whereby an author is entitled under certain conditions to terminate a licence or an assignment of copyright thirty-five years after its date of execution. Given that section 203 clarifies that termination rights under this section only affect copyrights arising under US law,⁹⁹ termination rights best seem characterized as a matter of copyright, governed as such by the lex protectionis. Thus, a worldwide licence or assignment would remain ‘subject to the US termination right with respect to exploitations in the US’,¹⁰⁰ even if the contract were governed by foreign law.¹⁰¹ However, some commentators have argued that termination rights affect the parties’ contractual relationship without affecting the copyright’s transmissibility.¹⁰² As a result, characterization—and accordingly the scope of party autonomy—would not be completely certain if the question arose in an arbitration.

Characterizing the formal requirements of transfer and/or licence agreements is likely to create similar difficulties.¹⁰³ Under French law, for instance, copyright transfer agreements must define the scope of the transfer regarding its extent, destination, geographical scope, and duration.¹⁰⁴ Any total transfers of future works are null and void.¹⁰⁵ While some consider that these requirements aim at protecting authors and therefore belong to copyright law, one could just as well argue that formality requirements concern the validity of contracts, and should therefore be governed by the lex contractus.¹⁰⁶ The latter view would be most attractive if the case concerned a worldwide licence and/or assignment, as arbitral tribunals could examine the formal requirements of a single law. However, doing so is not without risks in terms of public policy, for it is not guaranteed that the French courts would tolerate derogations from the French requirements,¹⁰⁷ even in arbitration proceedings. This issue relates to the impact of overriding mandatory rules and is addressed in the next section.

(D)  The impact of overriding mandatory rules in contractual cases

The preceding section explained that characterizing an issue as contractual leads to the application of the lex contractus, which is most often chosen by the parties. But even if an arbitral tribunal opts for a contractual characterization, the parties’ freedom to select the applicable law remains subject to overriding mandatory rules, ie, to quote a well-known definition, rules ‘whose respect is necessary for safeguarding a state’s political, social, and economic organization’.¹⁰⁸ Under virtually all arbitration laws and rules, arbitral tribunals must generally respect the parties’ choice, failing which they may be said to exceed the scope of their authority.¹⁰⁹ But tribunals must also consider overriding mandatory rules, as not doing so would create a strong risk that the award is refused enforcement or set aside for breach of public policy.¹¹⁰ There is some uncertainty as to which overriding mandatory rules tribunals should apply, as some commentators still believe that arbitrators derive their powers from the legal system of the seat of arbitration, and must therefore apply the overriding mandatory rules of the lex loci arbitri.¹¹¹ However, it is now widely accepted that whether an overriding mandatory rule is triggered depends on whether the dispute falls within its geographical scope of application.¹¹² That condition is not simply met because the arbitration is seated in a particular jurisdiction: independent geographical conditions are required, depending on the rule in question.

In the case of contractual disputes involving IP rights, it has been argued that overriding mandatory rules are unlikely to play a significant role.¹¹³ While this is true, several scenarios nevertheless illustrate that arbitral tribunals may have to apply overriding mandatory rules belonging to the country in which the relevant IP right arises. As previous studies do not examine this question in sufficient detail, the section below examines these examples in order to clarify the limits of party autonomy.

(1)  Potential overriding mandatory rules in contractual copyright cases

(a)  Termination rights

As noted above, there is some debate as to whether termination rights (eg, under US copyright law) should be characterized as a contractual or IP issue. Assuming a contractual characterization, leading to the application of the lex contractus, one could argue that such rights constitute overriding mandatory rules insofar as they reflect an important policy choice to protect authors.¹¹⁴ To be sure, domestic statutes and case law appear to be silent on the internationally mandatory character of termination rights. However, it has been argued that arbitral tribunals have a discretionary power to identify overriding mandatory rules based on the rationale and importance of the relevant provision, wherever domestic law is ‘completely silent’ on whether a rule can be derogated from by way of a choice-of-law agreement.¹¹⁵

This discretion could be used in the case of termination rights. As others have explained, authors are generally the weaker party in contractual negotiations and may licence and/or assign their copyright before the full commercial value of their work is known.¹¹⁶ Termination rights ensure that authors can regain copyright ownership in case compensation and/or the commercialization of their work is insufficient. Admittedly, termination rights seek to protect authors, ie, private parties, whereas overriding mandatory rules are traditionally meant to safeguard public interests. But it has long been acknowledged that rules protecting categories of individuals also safeguard the interests of society.¹¹⁷ In the case of termination rights, allowing authors to get their copyright back ensures that they can regain control of their financial future,¹¹⁸ and thus continue to create for the benefit of society.¹¹⁹ Moreover, by increasing an author’s bargaining power, termination rights encourage contracting parties to conclude balanced agreements in the first place—an aspect that may also be said to benefit society.¹²⁰ In any event, courts in some countries did identify overriding mandatory rules aiming to protect weaker parties.¹²¹ While characterizing termination rights within copyright law would provide a simpler way to protect authors, there is therefore a good argument that termination rights cannot be derogated by way of a choice-of-law agreement. Arbitral tribunals should ensure, for example, that authors can rely on US termination rights whenever a licence or assignment concerns acts of exploitation in the USA, even if termination rights are characterized as a contractual problem and the lex contractus does not authorize authors to terminate the agreement.

(b)  Rules on the non-waivable character of moral rights

By contrast, French authorities suggest that the French provisions on the formal requirements for the transfer of copyright are unlikely to constitute overriding mandatory rules.¹²² While it would be more prudent to comply with these requirements if a transfer agreement covered the French territory, parties could most likely avoid their application by choosing the applicable law.

However, there is a good argument that substantive rules on the non-waivable character of authors’ moral rights (including, for example, the right to be named as an author or to safeguard the integrity of a work¹²³) are so important that they should trump the lex contractus wherever the dispute falls within their geographical scope of application. In a case where the author of a book had waived her moral rights in a contract governed by US law, the Paris Court of Appeal found the waiver to violate public policy with regard to the commercialization of the book in France.¹²⁴ Arguably, a public policy violation would similarly arise if an arbitral tribunal applied a law upholding a moral rights waiver in circumstances where French copyright law is geographically applicable, ie, where a contract concerns the exploitation of work within the French territory. The same argument could be made for moral rights in other civil law countries, as these countries tend to emphasize the importance of moral rights and the link between a work and the author’s personality.¹²⁵ But this problem is unlikely to occur as moral rights disputes are seldom the subject of arbitration proceedings, most likely because they are generally of small financial value, and some civilian commentators have expressed (albeit unconvincingly) doubts over their arbitrability.¹²⁶

(c)  Rules on the equitable remuneration of authors

In the future, perhaps a more important example of overriding mandatory rules will concern the remuneration of authors for the use of their works,¹²⁷ an area where alternative dispute resolution has received official encouragement,¹²⁸ and is thus likely to grow. As authors are often the weaker party when licensing or transferring their rights, some legal systems grant authors the right to obtain ‘equitable remuneration’ or, to use the wording of the European Union’s Digital Single Market (DSM) Directive, ‘appropriate and proportionate remuneration’.¹²⁹ Under the Directive, authors also benefit from a contractual adjustment mechanism whereby authors ‘can claim additional, appropriate and fair remuneration […] when the remuneration originally agreed turns out to be disproportionately low compared to all the subsequent relevant revenues derived from the exploitation of the works […]’.¹³⁰ Thus, authors can sometimes obtain higher remuneration than the amount contractually agreed between the parties, whether by way of a lump sum or royalty fees.

Some legal systems such as the USA are not as protective of authors.¹³¹ Contractual protection is organized through other means, including—as noted above—by allowing authors to terminate a licence or an assignment 35 years after its date of execution. As a result, companies in the creative industries may be tempted to select, for instance, the law of an American state to govern a licence agreement. Insofar as the remuneration of authors is usually characterized as a contractual question,¹³² governed as such by the lex contractus, authors would then be deprived of their entitlements under the law of the country where the work is used under the licence or assignment agreement.

This is where overriding mandatory rules have a role to play. For example, under section 32(b) of the German Copyright Act, the right of equitable remuneration and contractual adjustment mechanism that authors enjoy under that legislation are ‘compulsory’ and apply

• 1) if German law would be applicable to the contract of use in the absence of a choice of law; or

• 2) to the extent that the agreement covers significant acts of use within [Germany].¹³³

Insofar as section 32(b) is applicable even where the parties agreed on the applicable law, German courts and commentators rightly treat the German author-protective provisions as overriding mandatory rules.¹³⁴ Arbitral tribunals should therefore ensure that authors can benefit from these provisions even if the contract were governed, eg, by New York law. Unfortunately, the territorial scope of the German provisions is unclear: in the second alternative of section 32(b), the German legislator did not define the term ‘significant’, and the original German term (‘maßgeblich’) could also be translated as ‘relevant’.¹³⁵ While some commentators suggested that ‘Germany has to be one of the economically important places of business with respect to the contract that is at stake’,¹³⁶ the Higher Regional Court of Stuttgart recently stated that the ‘decisive factor is (only) whether domestic acts of exploitation take place [in Germany]’.¹³⁷ As this statement does not appear to require acts of use to be significant, arbitral tribunals should ideally apply the German provisions to any acts of use located in Germany. However, subject to the first alternative of section 32(b),¹³⁸ remuneration claims based on German law would need to be limited to acts of use located therein.¹³⁹

Article 25h of the Dutch Copyright Contract Act similarly provides that the Dutch provisions on equitable remuneration can apply regardless of the lex contractus. Unlike under German law, this is notably the case if ‘the acts of exploitation take place or should take place wholly or predominantly in the Netherlands’.¹⁴⁰ Any acts of use in the Netherlands would thus not be sufficient. However, since the term ‘predominantly’ is not defined, it is unclear whether the Dutch provisions apply if over half of the exploitation occurs in the Netherlands, or if most of the exploitation occurs in the Netherlands, even if it is less than half.¹⁴¹ The latter approach would be the most prudent to diminish the risk of public policy issues, but tribunals have a certain margin of discretion here.

Finally, in implementing the DSM Directive, the French legislator recently clarified in Article L-132-24 of the IP Code that authors of musical compositions for audiovisual works cannot be deprived of the benefit of various author-protective provisions, including fair and appropriate remuneration, regardless of the law chosen by the parties.¹⁴² To reinforce these provisions’ overriding mandatory character, Article L-132-24 even allows authors to bring proceedings before the French courts despite any choice-of-court agreement in favour of a foreign court¹⁴³—an aspect which made a French commentator speak of ‘super loi de police’ (super overriding mandatory rule).¹⁴⁴ While Article L-132-24 does not suggest that authors can seize the French courts despite an arbitration agreement, the importance given to the French author-protective provisions suggests that arbitral tribunals should ensure that authors can benefit from them where the parties’ agreement concerns the exploitation of a work in France.¹⁴⁵

Surprisingly, the wording of Article L.132-24 is restricted to authors of musical compositions for audiovisual works, who were apparently in special need of protection due to the development of globalized platforms where the protection of French law might not be available.¹⁴⁶ This may suggest, a contrario, that the French provisions on equitable remuneration do not constitute overriding mandatory rules regarding other authors. However, some commentators have convincingly argued that these provisions could and should be treated as overriding mandatory rules regarding any author, based on an analogy with moral rights (which the French Court of Cassation famously recognized as constituting overriding mandatory rules).¹⁴⁷ Arbitral tribunals would be free to accept this argument, despite the silence of French legislation and case law.¹⁴⁸

Beyond France, Germany, and the Netherlands, one could argue that the provisions of the DSM Directive on equitable remuneration constitute overriding mandatory rules as well.¹⁴⁹ As such, arbitral tribunals should apply them whenever a copyright licence or assignment covers the European territory.¹⁵⁰ Certainly, doing so may complicate arbitral proceedings if the lex contractus is from a non-European state and does not contain similar author-protective provisions. However, this complication is arguably necessary to ensure the compliance of the award with European public policy.

Whether inside or outside the EU, domestic copyright systems may still diverge, particularly on whether remuneration may take the form of a flat fee or must be proportional to the revenues derived from the work’s exploitation.¹⁵¹ However, based on existing case law, there would be a good argument that domestic rules favouring proportional remuneration are not important enough to force arbitral tribunals to apply them if the law selected by the parties grants authors a flat fee.¹⁵² Acknowledging the equivalence between the two forms of remuneration would sometimes allow tribunals to calculate an author’s remuneration under a single law, and arguably achieve a satisfactory compromise between efficiency and respect for the policy choices of sovereign countries.

(2)  Overriding mandatory rules on competition law

Apart from the above examples, contractual cases relating to IP rights may require the application of overriding mandatory rules on competition law. In a trademark licensing dispute, the Court of Justice of the EU held that a domestic court that is requested to set aside an arbitral award must do so if the award violates European rules on anticompetitive practices.¹⁵³ Consequently, arbitral tribunals must apply—most likely ex officio—such rules whenever the dispute falls within their scope of application. This would be the case whenever the contract impacts competition in the European territory,¹⁵⁴ ie, given the territorially limited effect of IP rights, whenever the dispute involves an IP right arising under the law of a European country (or a unitary European right).

Surprisingly, the Swiss Federal Tribunal denied that competition law belongs to public policy,¹⁵⁵ so it is unclear that the Swiss courts would set aside or refuse enforcing an award for breach of Swiss and/or foreign competition law. But as the link between public policy and competition law is ‘almost universally acknowledged’,¹⁵⁶ arbitral tribunals are most certainly under a general duty to apply any overriding mandatory rules on competition law. While some jurisdictions only sanction ‘flagrant’ (ie, obvious) breaches of public policy,¹⁵⁷ recent developments suggest that domestic courts increasingly favour a ‘maximalist’ approach whereby courts must fully review the consistency of the award with public policy.¹⁵⁸ In Germany, the Federal Court of Justice went even further and ordered German courts to fully review the merits of awards on competition law.¹⁵⁹ This confirms that arbitral tribunals should not turn a blind eye to overriding mandatory rules on competition law, despite some uncertainty on the public policy standard and/or the link between competition law and public policy in some jurisdictions.

Of course, whether a licence or assignment agreement violates competition law depends on the competition rules in question. In the EU, the Court of Justice ruled that European competition law does not preclude a licence agreement from imposing payment of a royalty in the event of the revocation or non-infringement of the underlying patent, so long as the licensee can freely terminate the licence agreement upon reasonable notice.¹⁶⁰ By contrast, the US Supreme Court traditionally bars patent licensors from collecting royalties for the use of an invention beyond the expiry date of a patent.¹⁶¹ Because the rationale of this rule is to safeguard competition within the US territory, the US rule would need to be applied whenever the case involves a US patent,¹⁶² even if the parties chose the law of a European country to regulate their agreement.

There may be other scenarios in which party autonomy is and/or should be restricted in contractual disputes involving IP rights.¹⁶³ As will be explored in Part II, the example of competition law is particularly noteworthy, for competition law can sometimes provide limits to party autonomy where the case involves IP aspects and the parties chose the applicable IP law.

II.  THE SCOPE OF PARTY AUTONOMY IN THE EVENT OF A CHOICE OF IP LAW

As a practitioner observed, most parties to an IP arbitration only choose the law governing their contract.¹⁶⁴ This is perhaps because parties do not think about choosing the applicable IP law. If they do envisage it, right holders might fear losing rights under the normally applicable IP legislation, whereas IP users such as licensees might be worried about exposing themselves to a highly protective IP regime. Even if an agreement could be reached, parties might fear that choosing the applicable law is not allowed for IP issues and that problems could ensue at the stage of enforcement and/or annulment of the award.

However, as noted in Part 1, choice-of-law agreements are sometimes broad enough to cover non-contractual questions, eg, if they provide that any issue ‘relating to’ or ‘arising in connection with’ the contract shall be resolved under the law of state X. Considering that tribunals appear to have a general tendency to interpret choice-of-law agreements broadly,¹⁶⁵ such agreements may well be understood as covering IP questions.

Furthermore, there is anecdotal evidence that parties sometimes make direct choices of IP law. For example, in a dispute involving patents registered in twelve EU countries, the USA, Canada, and Taiwan, the parties were anecdotally reported to have chosen to apply ‘the laws of the UK and Germany that presented a common position regarding all patent-related claims’.¹⁶⁶ Likewise, in a dispute concerning supplementary protection certificates granted in eight European countries, the parties agreed on ‘the application of the principles as developed by the ECJ and the interpretation of these principles by state courts in certain European jurisdictions’.¹⁶⁷ While these disputes are confidential, they provide evidence that express choice-of-law agreements relating to IP are not unheard of. In such a situation, must and/or should arbitral tribunals uphold or disregard the parties’ agreement?

Unfortunately, this question does not have a straightforward answer. Although the possibility to choose the applicable IP law is often presented as an important benefit of IP arbitration, there is apparently no published case where a court confirmed the possibility of derogating from the lex protectionis. The judgment of the US Court of Appeals for the Federal Circuit in Deprenyl Animal Health v. Univ. of Toronto¹⁶⁸ was once cited as suggesting that ‘parties may require arbitrators to apply foreign law to questions involving US patents’,¹⁶⁹ but the judgment does not actually support this proposition.¹⁷⁰ On the contrary, section 294(b) of the US Code provides in essence that US invalidity defences may be raised by any party to the proceedings in arbitrations involving US patents.¹⁷¹ Other US statutes and statutes in other jurisdictions are silent, but virtually none of them allows the parties to an arbitration to choose the applicable law to IP issues.

For the most part, the publicly available arbitral case law is similarly uninformative. Where the parties express views on the applicable law to IP issues, it is generally to confirm that the lex protectionis applies.¹⁷² In Bayer v. Dow, the tribunal was invited to opine on the controversy: to challenge the application of the lex contractus to non-monetary remedies for the infringement of US patents, the respondent argued that US patent law ‘is a mandatory regime as a matter of US public policy, prevailing over state law, including parties’ contractual choice of law’.¹⁷³ But as the wording of the choice of law agreement only covered contractual questions, the tribunal concluded that US law governed patent issues ‘without the need to invoke [its] peremptory or mandatory operation’.¹⁷⁴ In passing, the tribunal noted that applying the law of the registration country accorded with ‘the public considerations in favour of a unified application of patent and related laws’.¹⁷⁵ And it notably cited, to support this statement, Article 8(3) of the Rome II Regulation—on the prohibition of party autonomy.¹⁷⁶ While acknowledging that arbitrators are not bound by the Regulation,¹⁷⁷ the tribunal thus seemed sceptical about the possibility of derogating from the law of the country of patent registration.

In the literature, commentators are divided. Some reject party autonomy, invoking the mandatory character of IP statutes,¹⁷⁸ the public interests safeguarded therein,¹⁷⁹ or the ‘territoriality principle’.¹⁸⁰ Others defend party autonomy, usually by emphasizing that arbitral awards, unlike court judgments, only bind the parties.¹⁸¹ In the middle, Hosna Sheikhattar and Alexander Odle have recently suggested that the enforceability of the award should be the main factor in determining the potential limits to party autonomy. In their view, the parties’ choice should only be disregarded if it creates a risk of annulment or non-enforcement.¹⁸² This argument goes in the right direction: indeed, it is too simplistic to suggest that party autonomy should always or never be accepted.

However, predicting whether the parties’ choice creates a risk of non-enforcement and/or annulment is not easy, nor is it clear that the enforceability of the award should be the only consideration. As others have argued outside the IP context, arbitral tribunals should also consider the need to safeguard the reputation of the arbitral process, as blindly enforcing the parties’ choice may encourage states to enact restrictions on arbitrability.¹⁸³ At the core of the debate on party autonomy lies an essential tension between efficiency and territorial sovereignty: while parties may prefer to limit the number of applicable IP laws, states have an interest in safeguarding their IP policies. In addressing this tension, the enforceability of the award in a given case is an important consideration, but beyond this arbitral tribunals should consider more generally whether any relevant public interests may limit party autonomy.

The weight of territorial sovereignty considerations depends on the IP issue at stake, ie, infringement, ownership, or validity. It also depends on the type of IP right: patents, for instance, may affect the public interest more closely than copyright, although this is not always the case. Efficiency considerations, on the other hand, mainly depend on the number of parallel IP rights, if any: the case for party autonomy is stronger if the choice-of-law agreement would allow the parties and the tribunal to avoid the application of numerous laws. Other factors, which do not squarely fall within any of the two categories, may be relevant in some cases, including the need to protect the rights of weaker parties, and to diminish the risk of conflicting decisions between arbitral tribunals and the courts of the country of IP protection. Clearly, given this variety of factors, the limits of party autonomy cannot be delineated by rigid rules. However, the following guidance may be given, using mainly the distinction among IP infringement, ownership, and validity in order to structure the analysis.

(A)  Choice-of-law for infringement issues

(1)  Irrelevance of the lex loci arbitri

As in litigation, the possibility for the parties to an arbitration to choose the law governing the infringement of IP rights is highly controversial. One thing is clear: as international arbitrators are not the judges of any country, and do not have a lex fori, they are not—contrary to what is sometimes assumed¹⁸⁴—bound by the prohibitions or restrictions on party autonomy established by the lex loci arbitri.

Arbitrators sitting in a European country, for example, are not bound by Article 8(3) of the Rome II Regulation, under which the lex protectionis rule may not be derogated by party agreement.¹⁸⁵ Nor are arbitrators sitting in Switzerland bound by Article 110(2) of the Swiss Private International Law Act,¹⁸⁶ which restricts party autonomy before Swiss courts to the remedies for infringement, provided the parties choose the lex fori after the dispute has arisen. Conversely, arbitrators sitting in China are not under an obligation to enforce choice-of-law agreements, even though the Chinese legislation allows the parties to choose the lex fori in infringement cases (not just with regards to remedies).¹⁸⁷ These provisions target domestic courts, not international arbitrators.

(2)  Party autonomy and the territorial limitation of IP

However, an essential problem is whether party autonomy is precluded by the fact that IP statutes do not provide protection against infringement acts committed outside the enacting country’s territory. Admittedly, an agreement providing, eg,  that the patent law of state A shall apply to the infringement of patents registered in states A and B would extend the geographical scope of law A. Law A would cover not only infringement acts committed in state A, but also in state B. But the parties could agree that the selected IP legislation must be considered as territorially unlimited, or that the dispute shall be decided as if law A were applicable to patents registered in state B.¹⁸⁸

Even without such an agreement, an arbitrator would be relatively free to apply the selected IP legislation extraterritorially. As the territorial limitation of IP is often understood to operate on the substantive level (ie, the claimant loses on the merits if the law for which it claims protection does not cover the alleged infringement¹⁸⁹), and domestic courts cannot generally review the merits of the case, an arbitrator’s decision on the geographical scope of an IP legislation cannot generally be second-guessed at the stage of enforcement and/or annulment. The same would be true if the territorial limitation of IP were understood to influence the applicable law,¹⁹⁰ since the applicable law is part of the merits. By way of exception, egregious legal error is a ground for annulment and/or non-enforcement in some countries such as the USA, but this ground tends to be interpreted very narrowly.¹⁹¹

Conversely, an arbitrator may refuse to apply the selected legislation beyond its scope of application. While this may create concerns as to whether the tribunal exceeded its authority,¹⁹² the tribunal could avoid this by emphasizing that it does apply the selected legislation but decides, on the substantive level, that the selected legislation does not provide extraterritorial protection (in the same way as the tribunal dismissed the claim in Benihana of Tokyo¹⁹³). By analogy, an arbitral tribunal declined to apply the French legislation on subcontracting because the scope of that legislation was implicitly limited to construction works performed in France, and the works in question had been performed abroad. Although French law was the lex contractus, the Paris Court of Appeal refused to set aside the award, holding that the choice of French law as the applicable law ‘did not have the effect of eliminating the question of the [legislation on subcontracting’s] scope of application’.¹⁹⁴ Subject to due process,¹⁹⁵ a similar reasoning could be followed if a tribunal refused to apply an IP legislation extraterritorially or otherwise dismissed an infringement claim, even if that legislation was selected by the parties. It would thus appear that, whatever the tribunal chooses to do, the territorial limitation of IP is unlikely on its own to create problems at the stage of enforcement and/or annulment of the award.

(3)  Party autonomy and public policy

One may nevertheless wonder if applying the selected IP legislation to infringement acts committed abroad would violate public policy (i) in the country whose IP legislation was applied extraterritorially, (ii) in the country of the arbitral seat, and (iii) in the country whose IP legislation was not applied even though the dispute fell within its scope of application, due to infringement being located in that country.

Arguably, the courts of the first country would be unlikely to object to the extraterritorial application of their own law, unless they are willing to consider the public policy of the foreign country whose law was not applied,¹⁹⁶ perhaps out of comity. Likewise, the public policy of the country of the seat would not be involved unless the seat has a relevant geographical connection with the dispute. In an infringement dispute, this would only be the case if the seat is also the country where infringement is alleged to have taken place, though again the courts of the seat may be willing to set aside an award violating foreign public policy.¹⁹⁷

Enforcement would be the most difficult in the country whose law was not applied despite infringement occurring there. Thus, arbitral tribunals should verify whether upholding a choice-of-law agreement would comply with the public policy of that country, not least because the award would most likely need to be enforced there, assuming the tribunal rules and grants an injunction in favour of the right holder.¹⁹⁸ Two scenarios must be clearly distinguished in this regard: one where the chosen law leads to more protection than the lex protectionis, and another where it offers less protection.

(a)  The chosen law affords more protection than the lex protectionis

The first situation to consider is where the chosen law leads to more protection than the lex protectionis: the former leads to a finding of infringement whereas the latter would not. Here, the award would create a restriction on competition in the country whose law was not applied. To avoid the risk of public policy violation in that country, it has been argued that arbitral tribunals should disregard the choice of law agreement in such circumstances.¹⁹⁹

This makes sense: if an arbitral tribunal declared that a party committed acts of infringement under the chosen law, the losing party could not use the IP right without paying royalties. As a result, that party would need to increase the price of its products or services, thereby forcing the public in the country of IP protection to pay higher prices than if there had been no choice of law.²⁰⁰ Furthermore, the losing party would compete at a disadvantage compared to other market participants, who would not need to purchase any licence for the same activity. If the losing party goes out of business, competition would be even more severely affected, and consumer prices in the country of IP protection may go up more generally. Considering that achieving low prices is the central goal of competition law,²⁰¹ and that competition law is generally acknowledged to be linked with public policy,²⁰² there is a strong argument that arbitral tribunals should disregard choice-of-law agreements designating an IP law providing for more protection than the lex protectionis. Indeed, it is unlikely that a state would allow the law of another state to create a restriction on competition within its own territory.

These competition law concerns are relevant for any IP rights, as the nature of IP rights is to create a private monopoly that may be enforced against third parties. But they seem especially strong in the area of patents, which may involve products that are of special importance for the public good, such as pharmaceuticals. How could one accept, for example, that a choice-of-law agreement, albeit in an arbitration between private parties, could increase the price that consumers or welfare systems pay for vaccination? The special nature of patents may be the reason why, as early as 1983, the US legislator appeared to prohibit choice-of-law agreements in arbitrations involving patent infringement. Under s. 294(b) of the US Code, ‘the defenses provided for section 282 [eg, non-infringement and invalidity] shall be considered by the arbitrator if raised by any party to the [patent arbitration] proceeding’. Thus, provided US law is raised by the defendant, arbitrators would be required to rule on the alleged infringement of US patents under US law, regardless of the parties’ agreement.²⁰³ Alternatively, one could read section 294(b) as requiring arbitrators to apply US law in addition to—not instead of—the selected foreign law.²⁰⁴ But this would complicate the proceedings and, if foreign law were to prevail, disregard the clear intent of the US legislator to have its policy choices respected in arbitrations involving US patents.

For other US IP rights, or indeed for IP rights in other countries, there is no legislative guidance. Regardless, competition law concerns would be a legitimate reason to invalidate a choice-of-law agreement leading to more protection than under the normally applicable law. This is the case even in copyright, for which competent courts and authorities do intervene to safeguard competition in the relevant markets, eg, in the cinematographic, book, or software industries.²⁰⁵ Similarly, the scope of trademark rights is closely linked to competition policy, regardless of the industry.²⁰⁶ In L’Oreal v. Bellure, for example, the CJEU famously held that free riding, ie, taking unfair advantage of the reputation enjoyed by an earlier mark, constitutes trademark infringement without any likelihood of confusion being necessary.²⁰⁷ Common law judges and scholars almost unanimously condemn such an expansive scope of trademark protection, including because they view the actionability of free riding as hurting the public interest in free competition.²⁰⁸ Choosing the law of a state recognizing free riding to govern the infringement of a trademark registered in a state rejecting this view would expand the scope of trademark protection beyond the latter’s wishes, and may, in turn, create issues of public policy from the perspective of the state of registration.

The link between infringement and consumer welfare can be directly invoked as a reason to limit party autonomy. But one could also argue, in a more constructed way, that infringement rules constitute overriding mandatory rules, applying regardless of the parties’ choice of law. At first sight, infringement rules hardly ever self-identify as overriding mandatory rules, and courts have seldom identified such rules in IP litigation. However, this scarcity is not due to the lack of importance of infringement rules. Rather, it is due to the fact that, as right holders generally bring proceedings in the country of alleged infringement, courts can simply apply their own law (as the lex protectionis or lex loci delicti) without triggering their overriding mandatory rules or the public policy exception.²⁰⁹ An arbitral tribunal would therefore be free to characterize an infringement rule as an overriding mandatory rule, provided they can show that the relevant rule is essential for a state’s economic and/or social organization.

In this regard, one could argue that infringement rules do not merely affect competition on the consumption level; they also embody essential policies on competition on the level of innovation (one of the main goals of IP, at least industrial property).²¹⁰ In order to facilitate improvements to existing inventions, the European Patent Convention, for example, broadly permits research ‘for experimental purposes relating to the subject-matter of the patented invention’.²¹¹ Some other jurisdictions such as the USA traditionally adopt a narrower understanding of this experimental use defence, as they consider that innovation is best encouraged by securing the interests of patentees, who are supposedly more incentivized to invent if they receive stronger protection.²¹² Whatever the best approach in terms of incentive to innovate, there is a good argument that national approaches to the experimental use defence constitute overriding mandatory rules, as they are closely tied to a state’s innovation policy.²¹³ If innovation drives economic growth,²¹⁴ then these approaches do seem essential for safeguarding a state’s economic organization.

Besides these aspects relating to what French scholars would call ‘economic public policy’,²¹⁵ infringement rules can also be essential for safeguarding a state’s social organization. Indeed, some infringement rules are influenced by fundamental rights considerations.²¹⁶ In copyright law, for example, the CJEU interpreted the scope of the right of communication to the public by considering not only the interests of right holders but also the ‘interests and fundamental rights of users of protected objects, in particular, their freedom of expression and of information’.²¹⁷ On this basis, the court held that posting a hyperlink to unauthorized content not legally available elsewhere does not constitute copyright infringement unless the defendant knew or should have known that the content was illegal (though this knowledge is presumed if the defendant carried an activity for profit).²¹⁸ Arguably, enforcing an arbitral award that applies a more protective standard in circumstances where EU copyright law is normally applicable would create an undue restriction on EU users’ ability to post hyperlinks and retrieve information from hyperlinks posted by others, and therefore their freedom of expression and information. The CJEU may not allow it, not least because the court has prohibited European states from interpreting the right of communication to the public in a way that provides broader protection,²¹⁹ reflecting the court’s intention to create a uniform standard of protection within the European territory.

Fundamental rights may also influence exceptions to infringement. Some US courts, for example, have recognized the existence of a freedom of speech defence against trademark infringement.²²⁰ This defence could be said to constitute an overriding mandatory rule, particularly due to the general importance of free speech under US constitutional law.²²¹ In the same vein, UK copyright legislation bars the enforceability of contractual terms purporting to prevent or restrict the doing of any act which is permitted under certain statutory defences such as parody.²²² The purpose of this provision—namely, protecting freedom of expression in the UK—would be defeated if the parties chose a law that does not recognize the parody exception. To be sure, there is a widely accepted difference between mandatory rules in the domestic context and overriding mandatory rules: the latter is generally understood as a more restrictive category, which only includes provisions that are so important that they prevent the parties from choosing the applicable law.²²³ But the rationale of an exception rooted in freedom of expression, and more generally fundamental rights, should be sufficient to convince an arbitral tribunal that the relevant provision is an overriding mandatory rule, applying regardless of the chosen law.

By contrast, choice-of-law agreements could, and should be upheld if the relevant provisions of the lex protectionis are not mandatory in the domestic context, as this suggests that they cannot be overriding mandatory rules either. In Hong Kong, for example, the legislator recently declined to prohibit the contractual override of copyright exceptions when it amended its copyright legislation.²²⁴ If a state permits private arrangements to provide more protection than under its legislation, there would be no reason why the parties should not be able to achieve the same result by choosing the applicable law.

Unfortunately, the overridability of exceptions is an open question in many legal systems, and the answer is likely to vary depending on the jurisdiction and the IP right in question.²²⁵ This can make it extremely difficult for arbitral tribunals to assess whether an exception can be overridden by a choice-of-law agreement, especially in copyright law, where the overridability of exceptions is particularly controversial.²²⁶ Based on the general principle of freedom of contract, a useful starting point may be that copyright exceptions are generally overridable unless otherwise stated,²²⁷ but courts in the country of protection could still recognize the mandatory character of an exception even in the silence of statutes and/or case law.²²⁸ The UK courts, for example, have sometimes recognized the existence of a public interest defence operating, eg, where the alleged infringer sought to inform the UK population.²²⁹ Although there is no indication of the mandatory character of this exception, would the UK courts not try to protect an alleged infringer’s freedom of information if an arbitral tribunal applied a law that does not sufficiently take the interest of the UK public into account?

Regardless of the argument against party autonomy, choosing the law governing infringement is likely to create problems where the choice is made in favour of a law providing more protection than the lex protectionis. While arbitral tribunals have wide discretion to identify the limits of party autonomy, the most prudent approach is to generally disregard choice-of-law agreements raising the level of IP protection, unless it is reasonably clear that the relevant rules of the lex protectionis may be derogated from in the domestic context, which would suggest that they may also be derogated by way of a choice-of-law agreement.

(b)  The chosen law is less protective than the lex protectionis

The second situation to consider is where the selected law is less protective than the lex protectionis: the former does not lead to infringement, whereas the latter would have. Here, it has been argued that party autonomy is acceptable because IP rights can generally be waived; hence, the situation would be similar to one where the right holder decides not to enforce its right against the defendant.²³⁰

This argument is convincing: right holders can indeed waive (or license) their IP rights in most scenarios. However, it should be emphasized that the moral rights of authors are non-waivable in civil law countries. In France, the Court of Cassation has even held that the French provisions on the infringement of the moral rights of integrity and paternity (ie, the right to object to derogatory treatment of a work and to be identified as the author) constitute overriding mandatory rules.²³¹ To ensure the enforceability of the award, arbitral tribunals should therefore take these rules into account whenever the dispute falls within their scope of application, which is the case whenever the alleged infringement of moral rights is located in France. Of course, an award finding that moral rights are not infringed may not need to be enforced in France. But the losing party could always bring annulment proceedings at the seat of the arbitration, arguing that the award violates French public policy. While this argument would be more likely to succeed if France were the arbitral seat, the courts of a foreign country in which the arbitration is seated may also be willing to take French public policy into account, although this is not guaranteed. And regardless of enforcement and/or annulment issues, France would like to see its approach to moral rights applied, rather than another approach which does not grant equivalent protection to authors. If arbitral tribunals enforced a choice-of-law agreement that conflicted with the French view on moral rights, the French legislator and/or courts could take steps to limit the arbitrability of such matters in the future.

Another reason that may justify disregarding a choice-of-law agreement lowering the level of IP protection is the need to protect the rights of weaker parties. Unscrupulous publishers, for example, may include arbitration and choice-of-law agreements in standard form publishing contracts, without providing authors with a meaningful opportunity to negotiate. Employment agreements may raise the same concern, as authors and inventors may be forced to agree on the applicable law due to their weaker position. However, assuming the dispute is arbitrable (which is not always the case in employment disputes²³²), inequality of bargaining power would not be an issue if the choice-of-law agreement were concluded ex post, after the dispute has arisen.²³³ Agreements concluded before any dispute may create difficulties, but these could be resolved under general principles of contract law, such as the doctrines of unconscionability or abusive clauses, depending on the applicable law.²³⁴ In other cases, where the parties freely consented to the choice-of-law agreement, denying protection to the right holder is not unfair; on the contrary, it would be unfair if the right holder could consent to the applicable law, perhaps for a consideration, and then override a choice-of-law.

It is unclear whether any other scenarios justify disregarding a choice-of-law agreement lowering the level of IP protection. As noted above, section 294(b) of the US Code requires arbitrators to consider US patent defences if raised by any party, but the provision only refers to ‘defenses’, so it is unclear whether it can be invoked offensively by the patent holder. On the other hand, the provision specifies that US patent defences may be invoked by ‘any party to the proceeding’, which would not make sense if the US legislator had intended to prevent patent holders from invoking US law. Unfortunately, US courts have never interpreted s. 294(b), and neither arbitral case law nor the legislative history of the provision provide any guidance.²³⁵ As a result, the scope of party autonomy would be uncertain in arbitrations involving US patents.

Still, party autonomy is overall less problematic where the selected law lowers the level of protection compared to the lex protectionis than where the selected law increases it. One could therefore envisage a system where, save in exceptional circumstances (eg, regarding moral rights and weaker parties), accepting party autonomy would generally depend on comparing the outcome under the selected law and the lex protectionis. That is, choice-of-law agreements relating to infringement matters would only be enforceable if the chosen law lowers the level of IP protection compared to the normally applicable law.

At first sight, this approach would make party autonomy lose its main benefit—namely, the simplification of proceedings, since arbitral tribunals would need to determine the content of at least two bodies of laws. But one could mitigate this issue by accepting that choice-of-law agreements are generally valid, unless IP users such as licensees can prove that the selected law affords more protection than the lex protectionis. In other words, the burden of proof would fall on IP users to show that the chosen law should not apply, so that arbitral tribunals would not need to compare several laws ex officio.

Admittedly, this approach would create an asymmetry between IP right holders and users since, except in some cases,²³⁶ only the latter could disregard a choice of law. Should it become the norm in arbitral practice, right holders would be unlikely to agree to the applicable law once they realize that the choice could mostly play in the other party’s favour. But right holders may prefer having the ability to formulate their claim under a single law and let the burden shift on IP users to show that the chosen law affords more protection than the lex protectionis. Indeed, this would make it more difficult for IP users to defend themselves against infringement claims. Accepting that choice-of-law agreements are generally valid unless IP users can show that the chosen law leads to more protection than the lex protectionis would make choice-of-law agreements less useful for right holders,²³⁷ but it would not deprive such agreements of any practical utility.

(4)  Remedies for infringement

The lex protectionis is also the natural approach for the remedies to infringement, including damages and injunctions. Yet, party autonomy is easier to accept here, as there is a wide consensus that remedies primarily involve the interests of the parties, as opposed to those of the public.²³⁸

This argument is the most convincing for damages, as parties can already conclude agreements limiting or exempting liability. Hence, they could also reach the same outcome by choosing the applicable law. Party autonomy only seems excluded with regards to punitive damages, ie, damages whose purpose is ‘not to compensate the claimant […], but to express [a] disapproval of the defendant conduct’.²³⁹ In some civil law countries, courts take the view that damages only serve a compensatory function. Hence, they refuse to enforce foreign judgments granting punitive damages, at least where the amount is disproportionate to the damage suffered.²⁴⁰ In arbitration, the enforcement of awards granting punitive damages is uncertain,²⁴¹ but the same approach as foreign judgments is likely to prevail by analogy.²⁴² As a result, arbitral tribunals should consider whether granting punitive damages would comply with the public policy of the country of IP protection (and the applicable procedural rules, if any²⁴³). Contrary to the views of some arbitral tribunals,²⁴⁴ the potential policies of the country of the seat against punitive damages would not be involved unless that country has a relevant geographical connection with the dispute.²⁴⁵ In an arbitration involving IP, this would only be the case if infringement were alleged to have taken place in the country chosen as the seat of arbitration.

By contrast, the argument that remedies mainly involve the parties’ interests is less convincing when it comes to injunctions. Injunctions do impact the availability of the infringer’s products and/or services on the market, and consequently the interests of the public in the country of IP protection. While the same would be true if an arbitral tribunal issued an injunction for breach of contract,²⁴⁶ the availability of IP injunctions often differs depending on the jurisdiction, for public policy reasons. Courts in civil law countries tend to quasi-automatically issue injunctions upon a finding of infringement, whereas courts in common law countries typically consider a whole range of factors, including the adequacy of injunctive relief for the type of harm suffered by the right holder, the balance of hardships between the parties, and whether the public interest or the rights of third parties may be affected by the issuance of an injunction.²⁴⁷ Should a tribunal bypass these discretionary tests and issue an injunction even though the lex protectionis would have led to a different outcome, the award would be unlikely to comply with public policy in the country of IP protection. If raised by any party (or the arbitral tribunal sua sponte), the rules of the lex protectionis on the availability of injunctive relief should therefore be considered, particularly if they suggest that injunctive relief should be granted based on individual circumstances.

Save in these scenarios, choosing the law applicable to remedies should be generally possible. And for the avoidance of doubt, there would be no fundamental objection against the application of different laws to infringement and remedies.²⁴⁸ Although it would be more coherent if both issues were governed by the same law, parties may be afraid of derogating from the lex protectionis with regards to infringement, while still wishing to choose the applicable law to the calculation of damages, for example. However, save where the parties wish to reduce the time and cost of the arbitration in cases involving parallel IP rights, this dépeçage would be an unnecessary complication.

(B)  Ownership

The limits of party autonomy are similarly unclear in ownership disputes, ie, in disputes where the tribunal must determine who the owner of an IP right is. However, as with the remedies for infringement, a convincing argument is that ownership mainly involves the interests of the parties:²⁴⁹ while the public in the country of IP protection may be interested in knowing whether an IP right is valid or infringed, it is less likely to be interested in knowing who owns the right in question.²⁵⁰ Moreover, an award on ownership would only bind the parties, so the impact of an award on ownership would have a limited impact on the market, regardless of the law applied by the tribunal.

Accepting party autonomy should create the least difficulties for the ownership of copyright, as copyright arises without formalities. And since copyright ownership is governed by the lex originis in some countries,²⁵¹ courts in these countries would be likely to accept a contractual derogation from the lex protectionis, though whether they would accept a derogation from the lex originis is an open question. But even in countries favouring the lex protectionis, copyright ownership is unlikely to create public policy issues, as ownership disputes tend to revolve around the factual question of identifying the person(s) who created a work.

In some cases, copyright ownership does involve public policy: the ownership of copyright in works created in the context of employment relationships, for example, is the subject of a famous divergence between civil and common law countries. While the former consider that authors have a natural right in their creations, and consequently vest copyright in the employee, the latter vest copyright in the employer as a way to encourage investments in creative individuals who will prepare a work for the benefit of the hiring party, and ultimately society.²⁵² Given the importance of these approaches in terms of economic and/or social organization, upholding a choice-of-law agreement in such circumstances would most likely create public policy issues if the chosen law led to a result that is incompatible with the lex protectionis.²⁵³ Waivers of moral rights may also involve public policy,²⁵⁴ but this issue concerns the contractual transfer of copyright, as opposed to the determination of the first copyright owner. In other scenarios, party autonomy appears to be generally acceptable, subject to any other overriding mandatory rule on, eg, the ownership of copyright in AI-generated works, which is currently the subject of a controversy on whether copyright should be vested—if at all—in the AI user or the creator of the AI tool.²⁵⁵

Disputes on the ownership of registered IP rights are similarly unlikely to raise public policy issues as they tend to revolve around the factual question of determining who developed inventive elements or, in the case of trademarks, who first used a sign on the market and/or first registered the mark. Historically, US law would consider that the first owner of a patent was the first person to invent. However, since the USA shifted its position in 2013, virtually all countries now operate under a first-to-file system whereby the first owner of a patent is the person who first registers it.²⁵⁶ Choosing the applicable law to ownership questions has therefore become less likely to create public policy issues, subject to potential overriding mandatory rules on—at least—the right of the inventor to be mentioned in his capacity of inventor (which is akin to the moral right of paternity in copyright law²⁵⁷), and perhaps also on the ownership of AI-created inventions.

It should also be noted that an award ruling on ownership, whether under the law of the country of registration or not, would be unlikely to allow the award-winner to directly modify the name of the owner on the register.²⁵⁸ This severely limits the benefits of submitting an ownership dispute to arbitration, as the award-winner could not easily enforce the IP right against third parties, let alone exploit it. Arbitral tribunals could order the losing party to accomplish the relevant transfer formalities,²⁵⁹ but—given their lack of imperium—tribunals should not do so unless the parties agree beforehand to implement the ownership determination on the relevant register. In practice, this will rarely be the case as right holders normally prefer an inter partes determination.

Yet, in the case of trademarks, there is an indirect way for arbitral awards to be implemented on the register. If an award declares (inter partes) that a trademark registered by one party belongs to the other, the registrant can no longer use the trademark. As use is generally a requirement to maintain—and sometimes, obtain—trademark registration,²⁶⁰ the award-winner can seek cancellation or revocation of the trademark, and then register the same mark in its own name.²⁶¹ Cancellation may be slowed down by the TRIPS Agreement’s requirement that a minimum period of uninterrupted non-use lapses before obtaining cancellation,²⁶² and the priority date for the new trademark would be different, but awards on trademark ownership can ultimately take effect against the public, whether or not ownership is determined under the lex protectionis.

This indirect way of achieving of achieving erga omnes effect should not raise public policy concerns, as a change of owner is unlikely to affect the interests of the public in the country of IP registration. In fact, the public is likely to prefer having clarity on the IP owner’s identity, instead of trying to understand the implications of an inter partes ownership determination. Admittedly, third parties having a direct interest in the IP right, such as licensees or assignees, would have their legal situation affected by an award ruling that an IP right is not owned by the licensor or assignor. But this problem is not without solutions and would exist even if ownership were determined under the lex protectionis.²⁶³ It should therefore not interfere with the possibility to choose the applicable law.

(C)  Validity

The possibility to choose the applicable law is perhaps the most controversial with regard to IP validity. While several writers are in favour of party autonomy,²⁶⁴ some commentators are radically opposed to it.²⁶⁵ As Thomas Halket puts it,

Does it make any sense to adjudicate the validity of IP other than under the law of the State that gave it birth? In most cases, the answer must be no. The following example demonstrates why this is so. Under U.S. law, a patentee has one year from the date it describes the invention in a printed publication to file the application for the patent on it. In the European Union there is no grace period, and any such prior publication destroys the right to patent.²⁶⁶ Would it make any sense to enforce an otherwise invalid EU patent by applying the U.S. rule or invalidating an otherwise valid U.S. patent by applying the European rule?²⁶⁶

This argument only relies on common sense as opposed to a full demonstration, but it can be expanded upon as follows.

(1)  The chosen law designates a law validating an otherwise invalid IP right

If an arbitral tribunal enforced an otherwise invalid IP right, the award would create a restriction on competition in the country of IP protection. Given the almost universally recognized link between public policy and competition law, party autonomy does seem unacceptable in this situation.²⁶⁷ Although IP validity awards generally bind the parties only,²⁶⁸ even an inter partes validity determination can have indirect consequences on the public, by forcing the losing party to acquire a licence to use the IP right, and thus raise the price of its products and/or services in the country of IP protection. By analogy, pay-for-delay agreements (whereby pharmaceutical companies agree with competitors to hold generic medicines off the market for a certain time after a patent expires) are severely criticized for being against the public interest.²⁶⁹ In the EU, these agreements have even been found to violate competition law.²⁷⁰ The public interest and competition law would be similarly impacted if the tribunal applied a law validating an IP right that is invalid under the lex protectionis. The reasoning here is exactly the same as with infringement, but the case against party autonomy is perhaps even stronger here, as the whole objective of rules on the validity of IP rights is to define the boundary between private monopolies and the public domain.

Arguments based on competition law arguments can be made for any IP right and in any industry sector, though the argument may be more compelling in some cases than others. But one could also argue, against party autonomy, that rules on the validity of IP rights constitute overriding mandatory rules, applying regardless of the parties’ choice. They have never been recognized as such because IP statutes are silent, and—in the case of registered IP rights—courts have exclusive jurisdiction to determine the validity of local rights, so courts need not suggest that validity rules constitute overriding mandatory rules to safeguard the policy objectives of forum law. But the case can be made that at least some validity rules are essential for safeguarding a country’s public interests, such as its political, economic and/or social organization.

Rules on patent validity, for example, are an essential feature of a state’s innovation policy, ie, a state’s economic organization.²⁷¹ Changes to patent subject matter eligibility, inventive step or novelty requirements directly influence incentives to invent: if protection is easily obtained, incentives to invent are strong, but patent rights may block future innovation; if obtaining patent protection is difficult, incentives to invent are weak, but future innovation is less constrained. Striking the right balance involves sensitive policy decisions which states have kept for themselves despite international IP treaties such as the Berne Convention, the Paris Convention and the TRIPS Agreement. Courts are likely to insist on the respect of these choices, even in arbitration proceedings.

The same argument can be made for rules on copyright validity, insofar as domestic originality requirements may be guided by the need to strike a balance between the protection of authors and ‘society’s interest in maintaining a robust public domain that could help foster future creative innovation’.²⁷² While originality requirements around the world have overall converged in recent years,²⁷³ some copyright subsistence issues are particularly controversial, such as whether copyright may subsist in AI-generated works. The Chinese courts, for example, have recently accepted that AI-generated works are protectable,²⁷⁴ while the US Copyright Office has repeatedly refused to do so.²⁷⁵ Choosing Chinese law for the existence of copyright in AI-generated works in circumstances where US law would normally have applied would jeopardize the policy choices of the USA, and most likely compromise the enforcement of the award for reasons of public policy.

Validity rules can also be essential for safeguarding a state’s social organization. Under most patent systems, inventions whose commercial exploitation would violate ordre public or morality are not patentable.²⁷⁶ Applying this rule raises difficult ethical and societal questions on the patentability of certain biotechnological inventions, eg, stem cells from human embryos,²⁷⁷ or rodents genetically programmed to develop cancer.²⁷⁸ And, since the concepts of ordre public and morality are deeply rooted in local culture, there is no common understanding of these concepts across different jurisdictions. In fact, some legal systems do not even have an ordre public and morality exception.²⁷⁹ Applying a law validating a patent despite the law of the registration country invalidating it on ordre public and/or morality grounds would thus create a serious risk of public policy violation, not least because of the identity between the terms ‘ordre public’ in the patent validity context and ‘public policy’ at the stage of enforcement and/or annulment.²⁸⁰ A risk of public policy violation would similarly exist in trademark law, where some states prohibit the registration of trademarks that violate public policy or morality, and may disagree on the balance between the right of the public to not come across outrageous trademarks and freedom of expression.²⁸¹

Thus, for several reasons, validity rules can be sensitive from a public policy perspective; in fact, this sensitivity was historically an important reason justifying non-arbitrability.²⁸² One may argue that, given the policy goals protected by validity rules, states did not broaden the scope of arbitrability without expecting arbitrators to apply their laws whenever the dispute falls within their scope of application. In the USA, this view was accepted in section 294(b) of the US Code, which mandates arbitral tribunals to consider US invalidity defences in disputes involving US patents. While other statutes are silent, a similar approach could prevail for any IP right and in any jurisdiction, at least where the parties’ agreement designates a law creating an IP right that would otherwise be invalid.

(2)  The chosen law designates a law invalidating an otherwise valid IP right

As with infringement, there is nevertheless a stronger argument in favour of party autonomy if the choice-of-law agreement designates a law invalidating an otherwise valid IP right. As right holders would also be free not to enforce their IP right, the choice-of-law agreement would effectively operate as a waiver of the IP right between the parties.²⁸³ But while this argument may be convincing at first sight, there are at least three counter-arguments.²⁸⁴

First, choice-of-law agreements would distort competition even in this scenario. If an arbitral invalidated—inter partes—a normally valid IP right, the award-winner could permanently use the IP right for free, while the public in the country of IP protection would have to purchase a licence to do so. To some extent, differences of treatment are part of commercial reality since the terms of licence agreements can be freely negotiated, though licences of standard-essential patents must be fair, reasonable, and non-discriminatory. But one party being free to use an IP right while others have to pay is a more significant inequality than mere differences on the amount of royalty fees.²⁸⁵ Normally, if an arbitral tribunal invalidated an IP right under the lex protectionis, third parties could restore market equality by bringing invalidity proceedings before the competent court, using the arguments of the arbitral tribunal as a road map.²⁸⁶ Third parties could also defend themselves against enforcement of the IP right by invoking, depending on the applicable law, the doctrine of abuse of right,²⁸⁷ misuse,²⁸⁸ and/or the doctrine of unclean hands.²⁸⁹ In some jurisdictions, third parties may even attempt to bring a counterclaim for damages,²⁹⁰ and/or an award of attorney fees²⁹¹ on the basis that the right holder attempts to enforce an IP right which it knows or should have known to be invalid. But if an arbitral tribunal invalidated an IP right under another law than the lex protectionis, third parties could not restore equality, and competitors may ultimately be pushed out of the market.

Second, there may be other reasons than competition law to disregard the choice-of-law agreement. As noted above, patent systems sometimes diverge when it comes to the patentability of ethically problematic inventions. While courts would be more likely to refuse to enforce an award validating an IP right that the law of the registration country would declare contrary to ordre public or morality, the same may also be true if an award invalidated an IP right on this ground against the wishes of the registration country. Indeed, weakening the patentee’s position by creating an exception to patentability could attract concerns that the award would hinder innovation in that country’s territory.²⁹²

Third, and perhaps most importantly, choice-of-law agreements on IP validity would increase the likelihood of conflicting decisions between national courts and arbitral tribunals. As arbitral awards on validity generally only bind the parties, an arbitral tribunal might invalidate an IP right inter partes before a court validates the right in question. Conversely, an award might uphold validity inter partes and award damages before a court invalidates the right erga omnes. Normally, such conflicting decisions are unlikely to materialize, mainly because the parties have an incentive to comply with inter partes validity determinations, except where an arbitral tribunal upholds an IP right and awards compensation.²⁹³ Moreover, arbitral tribunals would most likely reach the same outcome as the courts of IP protection if both apply the same law. But the risk of conflicting decisions would suddenly become very real if an arbitral tribunal applied a law other than the lex protectionis to IP validity. While there may be ways to resolve conflicting decisions, eg, by finding guidance in the way that courts resolve conflicting judgments belonging to the same legal system,²⁹⁴ this situation is still highly problematic and should be avoided as much as possible.

In sum, arbitral tribunals would have good reasons to generally disregard choice-of-law agreements on IP validity, regardless of the content of the chosen law. If, contrary to this argument, a tribunal were to enforce a choice-of-law agreement, the award could in any case only bind the parties: although Swiss and Belgian law accept that awards invalidating registered IP rights can have erga omnes effect,²⁹⁵ the Swiss and Belgian IP offices would be unlikely to agree to remove an IP right from the register if an arbitral tribunal invalidated Swiss or Belgian IP rights according to foreign law.²⁹⁶ The award-winner would have no incentive to request cancellation anyway, considering that an inter partes invalidity determination provides a competitive advantage over third parties.²⁹⁷

(D)  Cases involving parallel IP rights

It should be clear by now that choosing the applicable law to IP issues is likely to create public policy issues in various scenarios, particularly with regard to the infringement and validity of IP rights, at least where the chosen law expands the scope of protection compared to the lex protectionis. But notwithstanding this problem, there would be a good reason to accept party autonomy in one situation: where the case concerns a large number of parallel IP rights, and applying multiple laws would be unreasonable and too cumbersome. Depending on the number of jurisdictions involved, arbitral tribunals may be forced to apply a single law anyway, eg, the law with the closest connection to the dispute. Indeed, how could one expect tribunals to apply 15, 20, or even 100 laws? Unless the parties and/or the tribunal are exceptionally well versed in comparative IP law, legal expertise would be required to prove the content of each law, which would make it virtually impossible to resolve the dispute by arbitration. There is therefore a good argument that, in exceptional cases where applying multiple laws would be unreasonable and too cumbersome, arbitral tribunals should be free to enforce choice-of-law agreements regardless of the IP right and/or IP issue at stake. As the most that tribunals could do is to identify a single law to govern the whole dispute, or perhaps form different groups of compatible laws, allowing the parties to select the applicable law(s) themselves would not create an additional erosion of territorial sovereignty.²⁹⁸

To diminish public policy concerns, the parties’ choice should ideally be in favour of a country with a close geographical connection to the dispute.²⁹⁹ But the parties may have good reasons to choose a certain law beyond geographical connections, eg, if the chosen law is clear, easily accessible, or perhaps if the parties’ lawyers are trained in that law. The law chosen by the parties may in fact have more substantive similarities than another with a stronger connection: for instance, since copyright in common law countries shares the common heritage of UK law, it could make sense for two British parties to choose UK law to govern the infringement of copyrights arising under the laws of numerous common law jurisdictions. Of course, for all the reasons discussed above, the enforcement of the award could not be guaranteed. But the parties should arguably assume the risk of adjudicating their dispute in a single forum, and under a single law.

Unfortunately, it is impossible to fix a number of parallel IP rights starting from which arbitrators should be willing to tolerate party autonomy, as there would be an element of arbitrariness in any attempt to do so. Consistent with the flexibility of arbitration, arbitrators should have a discretion to decide whether to uphold choice-of-law agreements. They could do so, for example, if the dispute involves a global licensing agreement or an infringement claim occurring via a ubiquitous medium, but they could also enforce choice-of-law agreements where invalidity is raised in relation to a sizeable number of parallel IP rights, where appropriate.

Indeed, the key factor to determine whether party autonomy is likely to be acceptable is not necessarily the number of parallel IP rights, but the compatibility of all potentially applicable laws. For example, in the case of disputes involving European patents, it has been argued that parties could agree to the application of ‘Article 138 of the [European Patent Convention] and the practice and case law of the [European Patent Office] and its Boards of Appeal, including the Enlarged Board of Appeal regarding patentability, disclosure of the invention and added subject matter’.³⁰⁰ This makes sense: although European patents are a bundle of national patents, they are granted by the European Patent Office applying the rules of the European Patent Convention. Hence, party autonomy could be accepted as a matter of principle, regardless of whether the chosen rules of law validate or invalidate the patents. However, it should be recalled that divergences between domestic law and the case law of the EPO are not unheard of, as national courts are responsible for invalidating the national branches of a European patent.³⁰¹ If raised by any party and/or the tribunal sua sponte, differences between domestic regimes should therefore be taken into account as much as possible to preserve the policy choices of the registration country, and thus ensure that the award complies with public policy.

Similarly, in cases concerned with copyright originality, party autonomy should not create major difficulties in intra-European contexts, as the requirement of originality was harmonized by the CJEU.³⁰² Although courts in Member States have implemented this approach more or less quickly, parties could choose European law to determine the existence of copyrights arising under the laws of different European countries.³⁰³ But even if the case involved copyrights arising under US law, for example, applying European law to copyright originality would not necessarily create public policy issues, as both US and EU law adopt a test based on creative choices,³⁰⁴ which is arguably compatible with the philosophy of both civil and common law jurisdictions.³⁰⁵ Whether a US court would enforce an award derogating from US copyright law is unclear, but again, the risk of non-enforcement and/or annulment should be anticipated by the parties.

There may be other examples where party autonomy could be acceptable, eg, if the dispute concerns IP rights arising under the laws of the member states of the African Regional Intellectual Property Organization, where a multilateral treaty unifies industrial property while still maintaining that registered IP rights are independent and granted by national law.³⁰⁶ Parties can be creative, and should ideally seek advice from counsels with expertise in comparative IP law. They should try to identify, in particular, whether any common principles could assist the arbitral tribunal in reaching a decision. If there are none, parties should be aware that alternative approaches are conceivable, including requesting the tribunal to identify a reasonable middle ground between the IP laws and policies interested in the dispute, consistent with the ‘substantive law method’ notably advocated by Graeme Dinwoodie in international copyright litigation.³⁰⁷ But if the parties decide to go this route, they would move away from arbitration in law, and most likely enter the ground of arbitration in equity (which could be the subject of a separate study).

III.  CONCLUSION

Choosing the applicable law in IP arbitration is more complex and controversial than is commonly assumed. This article has clarified the limits of party autonomy by examining the impact of (i) choices of contract law and (ii) choices of IP law. In the first scenario, this article argued that tribunals should generally avoid extending the lex contractus to IP issues, including infringement and related remedies. Based on examples drawn from arbitral practice, it also suggested that tribunals should not abuse their discretion to characterize issues as contractual, even if doing so would simplify applicable law matters compared to an IP classification. Finally, this article identified a number of overriding mandatory rules which, in practice, may limit party autonomy in contractual IP arbitrations.

In the second scenario, confidentiality often makes it impossible to reach any definitive conclusions. All that can be done is to assess whether choosing the applicable IP law is likely to create issues at the stage of enforcement and/or annulment of the award. In this regard, this article has demonstrated that unsupported claims that party autonomy is always permitted are unconvincing and lack nuance. These claims arguably cause more harm than good: not only may states view them with suspicion, but arbitration users will be greatly disappointed once they realize that arbitration does not meet their expectations. Instead of portraying IP arbitration as an unregulated land where party autonomy is unrestricted, one should acknowledge that party autonomy does have limits, depending on the content and sensitivity of the substantive rules in question.

At the same time, it is equally exaggerated to suggest that party autonomy is never acceptable. Arbitral tribunals do have a larger freedom than domestic courts, as the former are not bound by domestic conflict rules against party autonomy and, to some extent, the territorial limitation of IP. In practice, the only true limit to party autonomy is public policy, but arbitral tribunals have wide discretion to decide whether to enforce or disregard choice-of-law agreements.

As explained above, derogating from the lex protectionis is unlikely to create public policy issues with regard to the ownership of IP rights and the remedies for infringement, subject to overriding mandatory rules. By contrast, choosing the law applicable to infringement matters would create public policy concerns if the chosen law increases the level of protection compared to the lex protectionis.³⁰⁸ Choosing a law leading to a less expansive scope of protection should be acceptable in most cases, subject to overriding mandatory rules and general doctrines on the rights of weaker parties. Thus, arbitral tribunals could, and should, accept that choice-of-law agreements relating to infringement claims are generally valid unless IP users such as licensees can prove that the selected law affords more protection than the lex protectionis (in which case the latter should apply). Choice-of-law agreements relating to the validity of IP rights, however, should be generally disregarded regardless of whether the chosen law validates or invalidates the right in question.

Notwithstanding public policy concerns, tribunals should have the discretion to enforce choice-of-law agreements in exceptional scenarios where the dispute involves a large number of parallel IP rights, and applying multiple laws would be too cumbersome and unreasonable. However, parties should be warned that party autonomy does not come without risks even in this scenario. Arguably, parties cannot consolidate disputes relating to parallel IP rights and always expect to resolve their dispute under a single or limited number of laws.³⁰⁹

These findings provide a flexible starting point to determine the limits of party autonomy in IP arbitration. States should, on their end, trust arbitral tribunals to take the public interest into account and disregard choice-of-law agreements where appropriate. Unfortunately, due to confidentiality, nobody really knows whether tribunals correctly do so. While this problem exists in any area affecting public policy,³¹⁰ IP arbitration mostly differs from other areas in that publicly available awards are particularly scarce, not least because some arbitral institutions have so far been reluctant to lift obstacles to publication.³¹¹ To strengthen legal certainty and public confidence in the arbitral process, arbitration stakeholders should continue their efforts towards achieving greater transparency, for scholarly efforts to clarify the limits of party autonomy or otherwise contribute to the development of IP arbitration will remain in vain if they do not do so.

Footnotes

Author notes