I. Introduction
In recent years, courts in EU countries have repeatedly refused to enforce arbitration clauses when judges deemed that arbitrating the disputes between the parties would allow them to evade fundamental market regulations of the forum. Crucially in these cases, parties had not only agreed on arbitration outside the EU but had also agreed on non-EU law to govern their transactions. Similar scenarios have arisen with choice-of-forum clauses.Footnote 1 This court practice has drawn significant criticism from both practice and academia, the primary argument being that the lack of respect for party autonomy undermines legal certainty. Indeed, the efficiency of international business transactions is considered fundamentally dependent on legal certainty; however, legal certainty is also crucial for any regulated markets, especially in the context of economic integration.Footnote 2 Businesses seek legal certainty for their respective private economic interests; regulators seek legal certainty to enhance public welfare within their respective jurisdictions. Whereas arbitration practitioners may worry that overriding mandatory provisions could act as an impediment to arbitration, regulators may wonder to what degree arbitration constitutes an impediment to the effectiveness of market regulation, especially in the EU Internal Market. This dialectical relationship between the private and the public interests in legal certainty has not yet sufficiently informed the debate on what is one of the most pervasive issues in the field.Footnote 3
This article revisits the topic by examining these controversial court practices from the perspective of the public law dimension of overriding mandatory laws, which obliges the judiciary to uphold the public policies which it enshrines (Section II). The situation becomes more comprehensible when the transnational challenge for both courts and arbitrators in confronting the combined effect of arbitration and choice-of-law agreements is considered (Section III). The question of how to cope with the resulting potential for regulatory arbitrage is, in the EU, intimately linked with the principle of procedural autonomy of the Member States, which in turn is subject to the general requirement of effectiveness under EU law (Section IV). This analysis shows why courts, in principle, are right to be reluctant to enforce arbitration or choice-of-forum agreements where the application of EU overriding mandatory laws is at stake. However, it is necessary to push the public law dimension of the problem even further. This article argues that the prima facie legitimate refusal by the courts to enforce the parties’ agreement to arbitrate is itself subject to the principle of proportionality, by which EU law (and constitutional laws in all Member States) fetters public power when it interferes with private rights (Section V). In particular, the courts’ central problem does not really arise as a consequence of the parties’ choice of arbitration. Instead, the clash with the applicability of overriding mandatory laws results from the parties’ choice-of-law agreement, which is what purportedly requires arbitrators to apply foreign laws irrespective of the laws of the affected market. Accordingly, striking down arbitration clauses is unnecessary—and thus disproportionate—when courts could remedy the problematic choice-of-law agreement instead. This is possible by seeking the parties’ consent to the application of the overriding mandatory laws of the forum in the arbitration. If denied, the non-enforcement of the arbitration agreement is inevitable. If accepted, such consent modifies the choice-of-law agreement and allows—and, indeed, obliges—the courts to refer the parties to arbitration. This pragmatic rather than theoretical solution allows the article to conclude by reflecting on the dialectical relationship of the private and the public interests in legal certainty (Section VI).
II. The Setting of the Problem
A. Overriding Mandatory Laws as Legal Irritants
As much as the foundations of party autonomy remain ‘elusive’ and ‘theoretically unresolved’Footnote 4 and public policy is perceived as ‘an unruly horse’,Footnote 5 the situation is not much better in regard to overriding mandatory laws. The name itself is an attempt to neutrally express differing conceptualizations of a phenomenon that is the bête noire of the dominant bilateral conflict-of-laws doctrine. Lois d'ordre public, lois de police, lois d'applications immédiate ou nécessaire;Footnote 6 Eingriffsnormen, Sonderanknüpfungsnormen, Wirtschaftskollisionsnormen:Footnote 7 these French and German terms each represent varying understandings of their content, mode of operation and contextualization—and, most of all, efforts to contain their effect and significance.
Overriding mandatory provisions are, indeed, both elusive and unruly. They are quintessentially non-cosmopolitan, non-libertarian, and unilateral; some would say they are parochial, authoritarian and exorbitant if not chauvinist.Footnote 8 They are the flexed muscle of the long arm of interest-driven legislatures reaching beyond their jurisdictions’ domestic realm. They are d'application immediate in their local forum because their application is not mediated by neutral and gently civilizing ‘traffic rules’ and follows only from their self-determined scope of application. Moreover, they blur the neat private–public divide by injecting a messy political element into the structured clean private law sphere. It thus comes as no surprise that conflict-of-laws scholars rarely have much sympathy for these unilateralist creations and prefer to circumscribe their role to limited exceptions. The arbitration community typically reacts with irritation to the resulting interference with the parties’ contract and autonomy.Footnote 9 In light of conceptualizations of arbitration as part of international justice that creates its own legal order,Footnote 10 it should not be surprising to see the rise of rhetorical claims that these overriding mandatory laws constitute an impediment to some ‘arbitral justice’.Footnote 11
B. Overriding Mandatory Rules as Impediments to Arbitration
The fear that EU law in particular could be an impediment to the effectiveness of arbitration stems in part from a series of court cases in EU Member States.Footnote 12 In brief, German and Belgian courts refused to give effect to arbitration agreements owing to fear of the effective application of their respective overriding mandatory provisions, where the parties had also agreed to the application of foreign law. This line of case law has now been confirmed and consolidated,Footnote 13 and the situation can be summarized as follows:
The analysis of Belgian courts starts with a rule that is more the product of path dependency than conscious conceptualization: in principle, claims involving overriding mandatory provisions of the forum cannot be taken out of the forum by virtue of arbitration or choice-of-court clauses. Belgian courts, however, exceptionally agree to enforce such clauses if it can be demonstrated that Belgian overriding mandatory laws will be applied abroad or if the laws to be applied abroad provide for equivalent protection.Footnote 14
German courts reach essentially the same result in practice, though by means of a reverse—and arguably conceptually sounder—rule-exception relationship. Claims involving German overriding mandatory rules are, in principle, capable of settlement by arbitration or in foreign courts; however, arbitration and other jurisdictional agreements are unenforceable if it is unlikely that the arbitral tribunal or court abroad will apply the overriding mandatory laws of the (German) forum and the applicable foreign laws provide no equivalent solution.Footnote 15ost of these cases concerned mandatory provisions in the European Commercial Agency Directive of 1986Footnote 16 and are thus spin-offs of the seminal Ingmar decision of the EU Court of Justice (CJEU, at that time still the ECJ), which was not then concerned with jurisdictional issues.Footnote 17 In the cases at hand, the parties had chosen laws and jurisdictional seats outside the EUFootnote 18 that provided neither any equivalent substantive protection nor any assurance that the arbitrators or the foreign courts would take the EU overriding mandatory provisions into consideration, let alone apply them. In all cases, the courts therefore refused to give effect to the arbitration or choice-of-court agreements. Similar cases can be found in other European countries, such as Italy, England and, most recently and in the same vein, Austria.Footnote 19
C. Overriding Mandatory Laws as Public Law
The unpopular effects of overriding mandatory laws on party autonomy must be placed in the context of the public law nature of market regulation. As illustrated by the Ingmar case, they are exceptional key vehicles for implementing (only) the market's most important policies; that is, those deemed so fundamental that a materially diverging outcome cannot be tolerated, irrespective of the international implications. If one wishes to conceptualize party autonomy as the product of fundamental or even human rights of the individual against the State,Footnote 20 then it is equally necessary to recognize that overriding mandatory provisions are constitutionally allowed—and, indeed, warranted—legal limitations upon those underlying fundamental rights, enacted in the public interest. The State has a duty to regulate in order to ensure that all stakeholders in the market, and those affected by it, can equally enjoy and develop (at least the core of) their fundamental rights. The law regulating private relations in particular hinges on the State ensuring a level playing field that is ‘so regulated that it reflects the objective order incorporated in the fundamental rights’.Footnote 21 The arbitrages that become necessary in the policy = making process, the balancing of interests, and the drawing of lines between them, through the (re-)definition of rights, are primarily entrusted to the political legislative process.Footnote 22In the EU, this legislative process is subject to the checks of constitutional safeguards. Whenever the legislature regulates in the public interest it still needs to preserve the essence of the rights of the individuals affected by the regulatory measures; likewise, the constitutional safeguards mandate the judiciary to protect these individual rights and their effectiveness.Footnote 23 The protection of the minority (which unsuccessfully resisted the regulatory change in the political process and is thus adversely affected) is what legitimizes the rule of the majority, upon which democracy is based. The ultimate safeguard limiting government interference in individual rights is the protection against arbitrariness and excessive individual burdens, operationalized through the principle of proportionality, a cornerstone of most national constitutional laws and of EU law itself.Footnote 24 This system of freedoms and limitation with checks and balances is designed to produce legality and legitimacy in the public interference with, and limitation of, individual rights. If the legislature has the competence to legislate on a given matter, and if it has respected required procedures and the principles of subsidiarity and proportionality, the enacted provisions must have full effect. This is the implementation of the underlying policy adopted in the democratic process and cast into law. That is the reason why, if a provision's legality has been affirmed and its scope interpreted accordingly by the CJEU, its international and overriding mandatory application must be respected—at least by all judges in EU Member States.
III. The Transnational Challenge
A. Substantive Party Autonomy and Regulatory Arbitrage
Drilling into the public law dimension of overriding mandatory laws illustrates the dialectic between party autonomy and the public policies that these laws seek to enforce. The issue is not the denial of party autonomy as such; on the contrary, it is about ensuring that party autonomy can be exercised without infringing upon third parties’ rights and public interests and that contractual freedom is protected for all in the long run. This is the very old quest for an appropriate balance between private and public interests, the complexity of which increases exponentially in an international context. Cross-border transactions pose a particular challenge for the conceptual relationship between freedom of contract and regulatory powers. The latter find their limitations in the (physical) boundaries of sovereignty and territoriality and thus struggle to maintain control over the former when facing the discourse and the reality of party autonomy both in its substantive and procedural dimension.
Freedom of contract, or private autonomy, allows parties to determine the content of their agreements: parties determine and make enforceable at law, and thus themselves regulate, the rights and obligations they wish to assume towards each other. Party autonomy in its substantive dimension, in turn, enables parties to determine the meaning, and thus the substance, of these rights and obligations through their choice of the applicable law. The acceptance of party autonomy allows parties to enhance legal certainty in an uncertain, non-uniform transnational legal setting and implies accepting strategic regulatory arbitrage. Regulatory arbitrage has been defined in the context of financial transactions as ‘those … transactions designed specifically to reduce costs or capture profit opportunities created by different regulations or laws’.Footnote 25
Regulatory arbitrage, such as forum shopping, is not illegal per se; arguably, it is very much a part of party autonomy.Footnote 26 The legitimate purpose is legal certainty: the free choice of the applicable law allows the parties to choose the regulatory framework for their transaction. Indeed, the entire idea of normative competition is based on regulatory arbitrage.Footnote 27 In the absence of international institutions with regulatory powers, the result is the well-known cat-and-mouse game between regulators and international economic operators.Footnote 28 This is where overriding mandatory laws kick in. Just as the parties seek legal certainty when operating at the international level,Footnote 29 their individual needs cannot result in regulators abandoning control over the effects of such transactions within their own jurisdiction. If party autonomy is contractual freedom with transnational ‘superpowers’, then internationally or overriding mandatory laws are those (actually rather few) mandatory regulatory provisions that are equally equipped with certain ‘superpowers’ to reach beyond the market's geographical borders. They neutralize—because they override—party autonomy (only) where the most important, fundamental policies are at stake. Doubts and reservations may well exist regarding the content and scope of certain regulatory interventions, as illustrated by the controversial Ingmar case. However, that is—or, rather, should be—a very different debate about specific content and normative choices. In the absence of institutionalized global regulation of the globalized economy, which unfolds through party autonomy, the need for national regulation to have international reach, and thus for overriding mandatory laws of the forum, can hardly be questioned.
B. Substantive and Procedural Party Autonomy
The balance between private and public interests in terms of choice of law, however, becomes fragile, if not a delusion, when the procedural dimension of party autonomy is added to the equation. As already mentioned, certain national and EU provisions are ‘overriding mandatory’ so long as they are part of the lex fori: they are binding for judges in the EU or in the Member State whose fundamental policy is at stake. That overriding mandatory effect, however, largely vanishes in other jurisdictions, as seen in Article 9(3) of the Rome I Regulation. This is arguably even more so in arbitration, which leads to the heart of the problem: can parties avoid l'application necessaire, and disarm the overriding mandatory effect of such provisions, simply by reserving the resolution of their disputes for courts or arbitral tribunals which are outside the reach of the regulator? The prima facie finding is that they can to a rather extensive degree.
The combined effect of choice-of-law agreements (opting for the laws of a different legal order) and of jurisdictional agreements (providing for a forum or an arbitral seat in a foreign jurisdiction) has led Luca Radicati di Brozolo to conclude famously that parties can actually reduce overriding mandatory laws to mere ‘lois d'application semi-nécessaire’, pointing at the potential benefits of the enhanced normative competition.Footnote 30 The background to his observation is the overall pro-enforcement bias flowing from the New York Convention, its prohibition of substantive review of arbitral awards and the presumption of the validity of arbitral awards. Both these privileges of arbitration have been developed further in the name of a ‘strong policy in favour of arbitration’,Footnote 31 culminating in the French doctrines of the autonomous validity of arbitration agreements independent of any national contract law consideration (Uni-Kod),Footnote 32 the immunity of arbitral awards from the effects of annulment in their foreign country of origin (Putrabali),Footnote 33 and the reduction of court review under the public policy exception (both in annulment and recognition and in enforcement proceedings) to only ‘flagrant, effective and specific’ violations of (French) international public policy (Thalès).Footnote 34 Except in outrageous cases, arbitration agreements and awards can be expected to ‘fly’ (to dwell on Robert Wai's theme of the ‘transnational lift-off’),Footnote 35 at least in France, thus seeming to fulfil the prediction (allegedly dating back to Domat in 1702) that ‘l'arbitre est … le juge naturel du commerce international’.Footnote 36
C. Arbitrators as the Guardians of Public Policies?
Would an emerging ‘ordre arbitral international’Footnote 37 and virtually unfettered party autonomy ruin the effectiveness of EU—or any—public policies? The most avant-garde—and in the literal sense utopian—answer is given by Emmanuel Gaillard:Footnote 38 public policies would not be imperilled in practical terms because the essence of those public policies, reflected in national (and thus potentially parochial) overriding mandatory laws, would be captured by the transnational or ‘truly international’ public policy that would emerge as the flip side of the lex mercatoria coin. This transnational public policy would be distilled through the methodology of comparative law from the vast range of domestic limitations to party autonomy and the public policies underlying them. International arbitrators could then define the boundaries of party autonomy and contractual freedom at the transnational level and refuse to enforce obligations irrespective of the chosen law when enforcing them would violate transnational public policies. As a consequence, arbitrators would no longer be required to apply national overriding mandatory laws. This grand construction of the transnational arbitral legal order deserves more attention than space permits. Suffice it to say that this vision is simply irreconcilable with any notion of legitimacy, ranging from democracy to separation of powers: it can hardly be accepted that ‘arbitrators are becoming—if with some hand-wringing and reluctance—default law makers for international traders’.Footnote 39 Arbitrators cannot possibly fill the void left by the absence of a global regulator, which States are not willing or able to create. Conflict of laws remains the uncomfortable and unavoidable reality.
A slightly less utopian answer seems to be premised on Holmes’ realist view that law is nothing more pretentious than ‘the prophecies of what courts will in fact do’.Footnote 40 Luca Radicati, as well as Pierre Mayer and Natalie Voser,Footnote 41 has suggested that
[t]he application of [overriding mandatory laws] by arbitrators is in a sense the implementation of a bargain between states, on the one hand, and the arbitral community, on the other hand. … [A]rbitrators may also have in mind a broader consideration of the ‘interests of arbitration’ which requires that they confirm the perception that arbitrators apply [such laws], since a generalized refusal of arbitrators to do so could have negative repercussions on the future amenability of courts to recognise arbitrability in this field.Footnote 42
As compelling as the argument may be in terms of arbitral deontology and self-representation, it is obviously not a legal argument that could assure courts that their overriding mandatory provisions are being respected, as Radicati himself readily acknowledges.Footnote 43
However, more realist constructions prove equally elusive. The much-invoked obligation of arbitrators to render an enforceable award has traction only if the seat of arbitration is in the forum State whose public policies are at stake; in any other State, a request for an annulment of the arbitral award on grounds of public policy will usually be futile.Footnote 44 If the seat is elsewhere, the award may not be enforceable in the forum State concerned owing to the public policy exception of Article V(2)(b) of the New York Convention. However, it is unlikely to be set aside in the (foreign) seat of arbitration and, as a result, will be enforceable in any other jurisdiction whose market is not affected by the transaction and that may host assets of the award debtor.
A somewhat more attractive, though still unsatisfying, pragmatic argument is that of Jan Paulsson,Footnote 45 and also made by Gary Born,Footnote 46 which is that arbitrators are given (expressly or inherently) the mandate by the parties to resolve their dispute over the applicable law so long as the arbitration agreement is couched in the usual broad terms (‘any dispute arising out of or relating to this contract…’). It seems difficult to conceive that the safeguard of public interests should depend on the exact wording of the arbitration agreement. The approach raises further questions in cases in which the application of an overriding mandatory provision is not pleaded by one of the parties.Footnote 47 In any case, this approach primarily calls for arbitrators to be accorded broad jurisdiction, and thus maximum discretion, without providing binding rules on the basis of which the arbitrator is to resolve such a true conflict of law.
In sum, none of these constructions gives a reassuring answer to the question of whether the fundamental policies underlying the forum's overriding mandatory provisions will be given proper, or at least equivalent, effect in the arbitration.Footnote 48 Arbitrators themselves affirm ‘that international arbitrators almost invariably apply these rules, or at least consider the question of their applicability, and do so with considerable proficiency’.Footnote 49 This is certainly an expression of a desideratum—yet the reality appears less promising.
What arguably weighs more heavily with arbitrators are cases such as the Swiss Hilmarton case,Footnote 50 the indignation over which is at the heart of the French doctrine that the effect of the law of the seat of arbitration should be ignored.Footnote 51 Swiss courts annulled an arbitral award because the arbitrator had applied Algerian mandatory laws concerning corruption when the parties had agreed on the exclusive application of Swiss law. The same pattern underlies Northrop Corp v Triad Int'l Marketing, in which the US 9th Circuit insisted on the exclusive application of the chosen Californian law. The court rejected the argument that the arbitrators should have applied Saudi Arabian law, which prohibited ‘intermediary contracts’ for the peddling of influence similar to the one in that case.Footnote 52 Indeed, the arbitrators themselves had excluded California's conflict-of-laws rules, and thus any renvoi to Saudi law, since this would ‘interject the laws of various other countries into the resolution of these disputes, thereby causing the uncertainty and lack of uniformity which the parties sought to avoid’.Footnote 53 The same argument has much traction in arbitrations under English law: section 46(1)(a) of the Arbitration Act, instructing arbitrators to apply the law chosen by the parties, could be read in conjunction with the no-renvoi provision of section 46(2) to prohibit arbitrators from moving away from the chosen English law.Footnote 54 Another recent example is that of an arbitral tribunal sitting in New York and mandated to apply New York law to a maritime agency agreement that did not consider an Austrian agent's counter-claim for compensation under EU law despite the fact that in parallel proceedings brought by the agent in Vienna, the US principal acknowledged the agent's entitlement when arguing (ultimately unsuccessfully) that this claim should be subject to the arbitration clause.Footnote 55 Other similar cases confirm the preference for party autonomy over overriding mandatory laws.Footnote 56
A particularly telling example is the indifference of the Canadian arbitrators in Accentuate v Asigra, which later played out in the English High Court. They affirmed that ‘[t]here may be interesting academic and intriguing domestic and international policy reasons why an arbitral tribunal should or should not apply non lex contractus mandatory rules of law to certain situations’ and acknowledged the Ingmar case law of the CJEU but concluded that ‘this does not justify restricting the parties’ freedom to choose a desired governing law in Ontario’ and to exclusively apply the latter.Footnote 57 The English judge's comments in Accentuate are also noteworthy. Tugendhat J first determined that the claim brought by the English distributor could not be obstructed by the arbitration agreement—and, indeed, the award—because of the need to give effect to the English provisions transposing the Commercial Agency Directive, which were overriding mandatory provisions according to Ingmar. He then observed:
I wish to add that nothing in this judgment should be taken as a criticism by me of the conduct or reasoning of the Arbitral Tribunal … [I]t was the duty of the Tribunal to apply the law which, according to the [contract], was designated by the parties as the law applicable to the substance of the dispute: see UNCITRAL Art 33. The passages cited from the Award … above demonstrate that the Tribunal was fully conscious of the relevant considerations. They were clearly aware that the English court might approach the matter differently for reasons which do not reflect adversely upon the Tribunal.Footnote 58
Such solicitude for the arbitrators’ consideration of the matter is somewhat surprising in view of the costs that their decision not to apply the EU overriding mandatory provisions—contrary to the suggestions in the arbitration literatureFootnote 59—inflicted on the Canadian principal.Footnote 60 However, the point is that judges, even when enforcing overriding mandatory laws against the parties’ combined arbitration and choice-of-law clauses, do not actually expect arbitrators to apply these laws. It is consequently not surprising that judges in EU countries, when in doubt, refuse to give effect to the parties’ transactional constructions that risk reducing overriding mandatory provisions of the forum to lois d'application semi-nécessaire. However, to return to the original question: why do these provisions need to remain d'application nécessaire?
IV. Procedural Autonomy and the Principle of Effectiveness
A. The Principle of Procedural Autonomy of the Member States
The fundamental question is: does EU law compel judges to deny the enforcement of arbitration agreements when they cannot expect that the overriding mandatory laws of the forum will be applied in the arbitration or in courts abroad, and the foreign laws to be applied provide no equivalent solution?Footnote 61 Some argue that it does not. When discussing a decision of the German Bundesgerichtshof that rejected the need for a preliminary ruling from the CJEU concerning whether Ingmar could justify not enforcing a choice-of-forum agreement, Jürgen Basedow pointed to the ‘principle of procedural autonomy of the EU Member States’ and concluded:
Thus, the effectiveness of enforcement mechanisms for the post-contractual compensation of commercial agents is not ensured by the law of the Union, instead being entrusted to the Member States, and the obligation of the Member States is expressed with some reserve: they ‘must not render practically impossible or excessively difficult’ the exercise of individual rights. This negative obligation is clearly less than a duty to ensure the greatest possible effect to exercise the rights in question. Thus, the effectiveness of the right to such payments is a matter for national law, with some broad limits drawn by the law of the Union.Footnote 62
The ‘principle of procedural autonomy’ is also the main argument relating to the extent to which national arbitration laws in EU Member States may have a pro-arbitration bias, despite the requirements of EU law. The CJEU confirmed in EcoSwiss v Benetton that the prohibition of anti-competitive agreements in (what is now) Article 101 TFEU constituted an overriding mandatory provision of EU law.Footnote 63 Questions of EU law raised in an arbitration would need to remain under the control of ordinary courts at the different stages of review of the award, ‘which may be more or less extensive depending on the circumstances’.Footnote 64 Moreover:
[i]t follows that where its domestic rules of procedure require a national court to grant an application for annulment of an arbitration award where such an application is founded on failure to observe national rules of public policy, it must also grant such an application where it is founded on failure to comply with the prohibition laid down in [now] Article [101] of the Treaty.Footnote 65
This decision has been hailed as implicitly recognizing that competition law matters are, in principle, capable of settlement by arbitration in Europe.Footnote 66 Much attention has also been paid to the other part of EcoSwiss, in which the Court clarified:
that [EU] law does not require a national court to refrain from applying domestic rules of procedure according to which an … award … in respect of which no application for annulment has been made within the prescribed time-limit acquires the force of res judicata and may no longer be called in question by a subsequent arbitration award, even if this is necessary in order to examine … whether an agreement which the interim award held to be valid in law is nevertheless void under Article [101] of the Treaty.Footnote 67
This is, indeed, an expression of the ‘principle of procedural autonomy of the Member States’. Some see this as giving carte blanche to the French doctrine of minimalist court review of awards for violations of public policy, as found in the Thalès case.Footnote 68 It is also reflected in Basedow's argument that procedural autonomy grants the courts of Member States the freedom to enforce jurisdictional agreements despite the expectation that an overriding mandatory provision of the forum will neither be applied nor compensated for by an equivalent foreign provision. In both instances, it is questionable whether the CJEU would understand the ‘principle of procedural autonomy’ in this way, as suggested by the opinion of AG Wathelet in Genentech v Hoechst and Sanofi-Aventis.Footnote 69 On the contrary: national procedural autonomy does not, itself, limit the principle of effectiveness of EU law but is subordinate to it.
B. The Context of the Principle of Effectiveness
The CJEU in Nordsee and, building on this, in EcoSwiss insisted on the importance of references for preliminary rulings to the CJEU in order to ensure the uniform application of EU lawFootnote 70—a path not open to arbitral tribunals since they are mere creatures of contract and thus not ‘courts or tribunals of a Member State’ in the sense of (now) Article 267 TFEU. The orthodoxy of the European legal construction would be at stake if the implementation of EU law were to diverge materially from one Member State to another.Footnote 71 Accordingly, (now) Article 5(3) of the Treaty on European Union (TEU) spells out the requirement of effectiveness as a benchmark for uniformity, which in the EU requires the principle of sincere cooperation and is a specific obligation of all Member States.
This requirement of effectiveness has been elaborated by the CJEU in a long line of cases:
In such circumstances [of lack of EU rules], it is for the domestic legal system of each Member State to designate the courts and tribunals having jurisdiction and to lay down the detailed procedural rules governing actions for safeguarding rights which individuals derive from [EU] law, provided, first, that such rules are not less favourable than those governing similar domestic actions (principle of equivalence) and, second, that they do not render virtually impossible or excessively difficult the exercise of rights conferred by [EU] law (principle of effectiveness).Footnote 72
The first part of this sentence led the CJEU in 2006 to affirm the ‘principle of procedural autonomy of the Member States’.Footnote 73 However, as its genesis shows, this principle applies only for the purpose of filling a void left by EU law.Footnote 74 In filling this gap with domestic procedural law, Member States are acting as trustees of the interest of the EU, which necessarily requires allowing for some procedural variance—but not variance in substance. National procedural provisions cannot, of their own right, take precedence over substantive EU law. They merely remain applicable to the degree to which they accord with the general principles on which EU law itself is based,Footnote 75 such as the protection of legitimate expectations or, such as the time-limitation issue in EcoSwiss, legal certainty and res iudicata.Footnote 76 Expressed differently, if national procedural law circumscribes the application of substantive EU law, it is not because of a higher principle of ‘national procedural autonomy’ but because EU law relies on national procedural law for its own purposes.Footnote 77
National procedural provisions, even if ‘autonomous’, must still comply with the requirement of effectiveness, which means that, at the very minimum, they should ‘not make the exercise of the rights conferred … [in practice] virtually impossible or excessively difficult’.Footnote 78 If taken in isolation, one could conclude—as Basedow does—that ‘this negative obligation [would be] clearly less than a duty to ensure the greatest possible effect to exercise the rights in question’.Footnote 79 This misses the point, however, as clarified by the President of the Court, Koen Lenaerts:
There is a fundamental difference between the national procedural autonomy to provide adequate remedies, which may indeed result in differences in enforcement of EU law, and the prior question whether a certain EU norm can be invoked before the national judge. The former is as a matter of subsidiarity usually left to the Member States and rightly so. The latter should be a matter of EU law, applying equally across all the Member States, thus giving everyone as much or as little chance to rely on EU law using whatever procedural format the Member State provides for this.Footnote 80
Compelling a party to bring a claim on the basis of an overriding mandatory provision in a non-EU forum, when it is likely that a choice-of-law clause will mandate the application of non-EU law, is making the exercise of that right ‘virtually impossible or excessively difficult’, given the predictability of failure as well as the time and costs involve.
Moreover, effectiveness is required not only in negative terms but also in positive terms,Footnote 81 as clarified by the CJEU:
For that purpose [of guaranteeing the application and effectiveness of EU law], whilst the choice of penalties remains within [the Member States’] discretion, they must ensure in particular that infringements of [EU] law are penalized under conditions, both procedural and substantive, which are analogous to those applicable to infringements of national law of a similar nature and importance and which, in any event, make the penalty effective, proportionate and dissuasive.Footnote 82
The much-criticized refusal of courts to enforce choice-of-forum or arbitration agreements in tandem with choice-of-law clauses where EU public policies are at stake is precisely such a sanction mandated by EU law. The importance of the principle of effectiveness becomes even more tangible when considering the consequences of courts’ failure to ensure such effectiveness: under the Francovich I case law, this would arguably give rise to the liability of the Member State to the individuals protected by the EU substantive provisions in question.Footnote 83
C. The Requirement of Effectiveness and Its Impact on Arbitration
The limits of national procedural autonomy have been clearly captured by AG Wathelet in Genentech: ‘No system can accept infringement of its most fundamental rules making up its public policy, irrespective of whether or not those infringements are flagrant or obvious … Put another way, one or more parties to agreements which might be regarded as anticompetitive cannot put these agreements beyond the reach of review under Articles 101 TFEU and 102 TFEU by resorting to arbitration.’Footnote 84 Put yet another way, it is simply unacceptable, because it is unconstitutional, for judges to allow deference to arbitration to water down ‘the most fundamental rules making up [a country's] public policy’ so that they become merely ‘semi-mandatory rules’.Footnote 85
This reasoning mutatis mutandis also shows the insufficiency of the suggestion that courts should refuse to enforce arbitration or choice-of-forum agreements only in the case of ‘fraude à la loi’Footnote 86 or ‘in clearly abusive circumstances’.Footnote 87 Courts have rightly refused to accept this very restrictive criterion.Footnote 88 To do so would be the equivalent of adopting the problematic French Thalès approach when scrutinizing the arbitration agreement for its compatibility with the public policy of the forum. Nor is it compatible with the requirement of substantive effectiveness of EU law. Indeed, it would make little sense, including from a costs perspective, to first pursue pointless arbitration or court proceedings, in which the EU overriding mandatory provisions are not expected to be applied, and then litigate anew the same dispute.
The concept of lois d'application semi-nécessaire may accurately describe the enforcement deficit in a transnational globalized economy, but it cannot justify accepting such a regulatory deficit. The refusal of courts in the EU to give effect to arbitration agreements that would undermine the effectiveness of EU overriding mandatory provisions is simply the application of the principle of effectiveness in its negative dimension. It is regrettable that this frustrates the parties’ (not the Member States’) procedural autonomy—but such a sanction, at least as a starting point, may be required even from Member States in order to ensure the full effectiveness of EU law.Footnote 89 Concerns that such an approach would be merely speculative and should therefore be left to the ex post facto control at the enforcement stage, as suggested by the US Supreme Court in Mitsubishi,Footnote 90 cannot take priority over ensuring the effectiveness of core public regulation. Advocating a solution such as that in Mitsubishi is unrealistic not only where courts adopt a problematic minimalist control of public policy.Footnote 91 As already noted, requiring parties to engage in a pointless arbitration is incompatible with the court's obligation to respect procedural efficiency as well as the EU requirement of substantive effectiveness. The reactions of European courts to the potential circumvention of their overriding mandatory laws, especially those originating in EU law, can hardly be condemned as a matter of principle.Footnote 92 The question that remains is whether breaking the arbitration agreements is the only and necessary consequence.
V. Proportionality and Pragmatism
There is a solution to the tension between ensuring the effectiveness of EU overriding mandatory provisions and the efficiency of international arbitration. It lies in the public law dimension of the problem. As described above,Footnote 93 overriding mandatory provisions limit the fundamental rights of the parties upon which party autonomy is constructed. The refusal to enforce the parties’ choice-of-law agreement and their arbitration or choice-of-forum agreement is a legitimate yet harsh sanction. Such interference with the freedom of the parties can be justified only to the degree to which it is necessary. This follows directly from the principle of proportionality, which is the essential constitutional limitation on the exercise of sovereign power affecting the fundamental rights of individuals. This principle binds all public authorities, at both EU and Member State levels,Footnote 94 most of whose constitutions recognize the principle.Footnote 95 Although it is generally accepted that legislatures have a particular broad margin of appreciation regarding economic matters,Footnote 96 it is also clearly recognized that freedom of contract is covered by EU fundamental rights—and thus subject to the principles of proportionality.Footnote 97 This means, in the CJEU's own words:
In accordance with the principle of proportionality, which is one of the general principles of [EU] law, the lawfulness of the prohibition of an economic activity is subject to the condition that the prohibitory measures are appropriate [or suitable] and necessary in order to achieve the objectives legitimately pursued by the legislation in question, it being understood that when there is a choice between several appropriate measures recourse must be had to the least onerous, and the disadvantages caused must not be disproportionate to the aims pursued.Footnote 98
This is the benchmark of legality that the legislator must respect when devising overriding mandatory laws and that courts must respect when interpreting provisions as constituting overriding mandatory rules.
A. Suitability
Proportionality is, perhaps, what Basedow had in mind when he argues that refusing to enforce a jurisdiction agreement and entertaining the claim by the protected European party is ‘far from being effective; in terms of its broader impact, it is almost completely ineffective by its very nature’.Footnote 99 This is because an EU judgment that gives effect to overriding mandatory provisions and ignores an arbitration agreement is enforceable only within the EU. Anywhere else, and especially in the foreign party's home jurisdiction, enforcement is likely to fail. Indeed, courts outside the EU, which are not bound by EU public policy considerations, will usually recognize the validity of the choice-of-forum or arbitration agreement, which then becomes an obstacle to recognizing a European court's jurisdiction and its judgments. With this in mind, Basedow suggests:
Well-advised principals from third states will react to this case law by cautiously avoiding the placement of any assets in the EU which might serve as the basis for the subsequent enforcement of an agent's claims. For the everyday payment operations undertaken in Europe, a bank account in the Principality of Liechtenstein (or other similar places) will usually be sufficient; in Liechtenstein, a Contracting State of the European Economic Area, judgments from EU Member States will be not be enforced under Brussels I or the Lugano Convention. Moreover, well-informed principals, when negotiating an agreement with an EU agent, will use the uncertainty of a duty to pay post-contractual compensation as an argument to squeeze the agent's remuneration that is due during the term of the contract. Thus, instead of protecting the agent, the courts might pave the way for the principal's higher profits.Footnote 100
This point sits uncomfortably with Basedow's subsequent conclusion that arbitration agreements should be disregarded in order to protect against abuses when justified on the basis of the facts of the case.Footnote 101 If he were right,Footnote 102 courts in the EU would be all the more entitled to start from a presumption that such ‘off-shore’ arbitration or choice-of-forum clauses flanked by choice-of-law clauses should not be given effect. Moreover, focusing on whether there is abuse on the basis of the facts of a particular case does not properly reflect the nature of overriding mandatory laws as understood by Basedow himself: they are not merely domestic statuta interventionalia that protect special group interests but statuta institutionalia that implement fundamental economic, social and political policies.Footnote 103 The protection they provide to individuals is secondary to this goal.
Basedow's deeper point seems to be that the principle of effectiveness is not served by the courts’ disregard of the parties’ procedural autonomy when the application of overriding mandatory laws is at stake because of the parties’ possibility of escaping that application. Re-contextualized in light of the public law dimension of overriding mandatory laws, Basedow's argument could be that the courts’ repressive approach would fail the first hurdle of the test of proportionality, which is that the measure must be suitable for attaining the pursued policy goal. However, this would ignore the already mentioned deference owed to the regulator's margin of appreciation and prerogative of choice of means, especially in economic matters.Footnote 104 More importantly, the argument relies on the same fallacy that animates the construct of lois d'application semi-nécessaire:Footnote 105 a sanction enforcing a regulatory measure designed to remedy a market failure cannot be unsuitable because the regulated actors simply resort to ever more inventive transactional engineering and regulatory arbitrage to undermine the regulatory effort. Otherwise, most financial regulations (if not all regulation in the face of globalization) would likely fail the test of proportionality and therefore be declared illegal.
The argument is also questionable factually. It may be possible to access the EU market in circumstances where there are no assets whatsoever, merchandise or payments, within the reach of the execution measures of European courts—albeit at potentially significantly increased transaction costs. Foreign companies are thus unlikely to want to operate and compete on this basis in one of the world's largest markets. The recent standoff between the EU and globally operating tobacco corporations over the operation of a vast tobacco-smuggling scheme involving tax fraud and money laundering illustrates this: in order to retain access to the EU market, all corporations seated outside the EU (with one exception) accepted the tough conditions imposed by the Commission in settlement of their involvement in the scheme.Footnote 106 The sanction of seriously complicating, if not actually denying, access to the EU market should provide enough leverage in most cases to obtain compliance with EU judgments giving effect to the overriding mandatory provision in question. Consequently, refusing to give effect to the arbitration or jurisdictional agreement is not an unsuitable sanction.
B. Necessity
If the regulatory measure pursues a legitimate policy objective and is not clearly unsuitable, it is nonetheless disproportionate if its impact on the individual's rights exceeds what is necessary to achieve the intended objective. This is a general principle of EU law that is explicitly enshrined in Article 5(3) TEU as well as in the case law relating especially to fundamental rights.Footnote 107 The question for present purposes is whether there is a less restrictive alternative measure that is equally effective in obtaining the policy objective.
1. Focusing on the problem: The choice-of-law agreement
The path to such alternative solutions can be found in Mitsubishi,Footnote 108 the very decision that pioneered the change of direction from general exclusions of certain matters from arbitration to ensuring the application of mandatory laws in matters of public policy. The intriguing question remains: what made the US Supreme Court reverse an order made by the Court of Appeals for the First Circuit which had found that antitrust claims brought by a car dealer in the context of the termination of a distribution agreement were not capable of settlement by arbitration? The US Sherman Act was clearly intended to govern these claims, but the Supreme Court accepted that they should be referred to an arbitral tribunal composed of three Japanese arbitrators sitting in Japan and called upon to apply Swiss law.
Arguably, this bold step was motivated not only by the manifold policy considerations which the Supreme Court set out in its judgment but also by an eminently practical consideration. The Court noted in footnote 19 ‘that in the event the choice-of-forum and choice-of-law clauses operated in tandem as a prospective waiver of a party's right to pursue statutory remedies for antitrust violations, we would have little hesitation in condemning the agreement as against public policy’.Footnote 109 This suggests that at the outset, the Court took more or less the same position as the Belgian and German courts.Footnote 110 However, a factual detail provided in the same footnote shows why the Supreme Court did not feel the need to condemn the parties’ agreement providing for arbitration in Japan according to Swiss law: ‘At oral argument, however, counsel for Mitsubishi conceded that American law applied to the antitrust claims and represented that the claims had been submitted to the arbitration panel in Japan on that basis.’Footnote 111
Irrespective of the controversy over what the US Supreme Court actually meant in footnote 19, the approach of seeking an undertaking that those—and only those—issues intended to be governed by the forum's overriding mandatory rules shall be governed by them irrespective of the lex contractus, is pragmatic and efficient.Footnote 112 Indeed, this approach is the most convincing way of solving the courts’ dilemma of seeking to respect the parties’ original intention to arbitrate whilst also respecting the legislatures’ intention to protect fundamental public interests. As shown above,Footnote 113 the courts simply cannot tolerate the deactivation of their overriding mandatory provisions by party autonomy since the former, necessarily and legitimately, place limits to the latter. However, before striking down the parties’ contractual construction, the courts must first verify whether the circumstances of the case may nevertheless allow them to uphold the arbitration agreement without compromising their duty to ensure the effectiveness of the fundamental policies in question. It is then the responsibility of the party requesting enforcement of the arbitration agreement to present the necessary clarifications.
This should be unproblematic in the absence of a choice-of-law clause so long as the applicable conflict-of-laws rules can be expected to lead to the selection of the law of the country whose fundamental policy is at stake.Footnote 114 On the other hand, when the parties have submitted their contract to foreign law, the party requesting arbitration has the opportunity to show the court that the protection afforded by the lex contractus is equivalent to that of the lex fori, thereby ensuring an outcome that will not clash with the forum's public policy. If such equivalence cannot be established or remains doubtful, the requesting party's last resort would be—as in Mitsubishi—to undertake that it expects and accepts the application of the overriding mandatory laws of the forum by the arbitrators. Once recorded by the court, this would constitute a variation of the original choice-of-law agreement by confirming the applicability of the forum's overriding mandatory laws to the regulatory issues at stake. The issue created by the original choice-of-law agreement would thus be eliminated, and the court could refer the parties to arbitration without being concerned that its overriding mandatory laws would not be applied because the varied choice-of-law agreement would be binding on the arbitral tribunal. The effectiveness of the forum's fundamental policies and market regulations would then no longer be at stake.
2. Objections and consequences
One objection to this contractual solution could be that a unilateral undertaking cannot cause the variation of the original agreement. This objection could arise if the party that initiated the court proceedings continues to resist arbitration. However, at this stage, an expression of consent to the variation by the party resisting arbitration is irrelevant: that party's consent can be presumed, since the non-application of the forum's overriding mandatory rules is its only legitimate reason for resisting the arbitration agreement. Accordingly, the resisting party would be prevented from rejecting the undertaking given by the party moving to stay the court proceedings. Moreover, inverse concerns that the latter party had given the undertaking under duress because of the court's threat to ignore the arbitration agreement would also miss the point. Such concerns would ignore the fundamental nature of both overriding mandatory laws and party autonomy: the foreign party was not free to contract out of the application of the overriding mandatory laws in the first place. The undertaking is merely the reconciliation of the parties’ autonomy with the legal framework in which that autonomy is exercised. The undertaking is no different from, for example, a behavioural undertaking given by a merged entity in order to obtain clearance of the merger from the competition authority. Finally, another false theoretical objection would be that lois d'application immediate themselves define their own scope of application so that their application cannot be dependent on the parties’ intention. The parties do no more than acknowledge, and undertake to respect, the (in any case overriding mandatory) scope of application of the rules upon which they had previously cast doubt by their original choice of law. From that perspective, their agreement merely serves the purpose of avoiding the legal sanction (non-enforceability of the arbitration agreement) that the court would have to impose if they insisted on their original choice.
Once the court has referred the parties to arbitration on this basis, the variation of the choice-of-law agreement thus obtained is binding even on the most party-autonomy-minded arbitrators. The arbitral tribunal is no longer torn by the dilemma between loyalty to the parties and regulatory threat. The object of the variation agreement is technically a dépeçage: although the originally chosen law generally remains the lex contractus, the parties accept its derogation by the overriding mandatory laws of the affected market (only) insofar as the scope of the latter claims mandatory application as lex specialis. This then defines the arbitral tribunal's mandate to decide the dispute without risking later incompatibilities with the public policy of the affected market.
When seen in the context of the ‘public law’ requirement of necessity, courts are arguably obliged to request that such a clarification of the law be applied by the arbitral tribunal when confronted with arbitration agreements whose enforcement could lead to the violation of certain fundamental policies of the forum. So long as a defendant is willing to give an undertaking that ensures the applicability of the forum's overriding mandatory provisions, a categorical refusal to enforce the arbitration agreement would constitute an excessive, because unnecessary, interference by the State with the parties’ party autonomy—at least in its procedural dimension. This solution therefore catalyses the synthesis of the antithetical relationship between party autonomy and overriding mandatory provisions, allowing both for the efficiency of the former and the effectiveness of the latter: the defendant must sacrifice the right to plead the exclusive application of the law originally chosen for the sake of keeping the arbitration agreement alive, whereas the claimant can invoke the protection of the overriding mandatory laws but can do so only by accepting arbitral jurisdiction.
VI. Conclusion
Overriding mandatory laws, at least those of the EU, are not a kind of ghost haunting the efficiency of international arbitration. When framed and applied correctly, they are, indeed, the lois de police that enforce respect for core national and EU public policies. Although they may appear antithetical to international arbitration, it is their public law nature that allows a synthesis to be developed between the interests at stake: ‘private’ and ‘public’ legal certainty.
It is worth noting that the Scherk rhetoric of legal certainty in international trade that requires the enforcement of jurisdictional and choice-of-law agreements applies both ways.Footnote 115 Leaving all public interest issues to ex post facto control at the stages of enforcement would indeed enhance ‘orderliness and predictability’ in international contracts. However, it would be at the expense of the ‘orderliness and predictability’ that the legislature intended to guarantee to market participants, contractual stipulations notwithstanding, by defining a stable basis of their operations in an overriding mandatory rule of law.Footnote 116 Expressed differently, enforcing arbitration and choice-of-law agreements in order to avoid ‘imperil[ing] businessmen's willingness to enter into international commercial agreements’Footnote 117 could result in ignoring the fact that the EU overriding mandatory laws are designed precisely to ensure the establishment and good functioning of the Internal Market—and thereby also foreign businessmen's opportunity to access, and benefit from, the level playing field of these regulated markets.Footnote 118
The problem boils down to the question of why ‘private’ legal certainty should be given primacy over ‘public’ legal certainty. This becomes more tangible when considering the reproach inherent in the criticism of the European court decisions, notably that they authorize a breach of good faith by one of the parties.Footnote 119 The fallacy is that statuta institutionalia do not, and cannot, care about individual arrangements but are concerned about ensuring the functioning of the market through their rule over the multitude of individual contracts that they regulate.Footnote 120 That is precisely why they constitute overriding mandatory laws. This becomes even more palpable when considering Article 101(2) TFEU, the archetypical overriding mandatory rule of EU law: any agreements prohibited by EU competition law, because they distort the level playing field of the Internal Market, ‘shall be automatically void’. There is simply no good faith to start with—and thus no legitimate expectations—when parties enter into agreements that undermine competition in the Internal Market. A smart use of party autonomy cannot possibly change that outcome.
Counsel who draft international agreements must do their homework. In virtually all the cases discussed here, starting with Ingmar, there is no indication of an intent to evade EU laws. Counsel for the North American companies seem to have simply ignored the problem of overriding mandatory laws in other jurisdictions and (understandably) have desired uniformity in their contracts with agents or distributors around the globe—rather than being aware of, and accepting, the different regulations of the foreign markets in which they wish to operate. Here, as well, the rhetoric of the US Supreme Court cuts both ways. The parties that simplistically insist on compliance with the jurisdictional and choice-of-law clauses must accept that:
We cannot have trade and commerce in world markets and international waters exclusively on our [preferred] terms, governed by our [preferred] laws, and resolved in our [preferred] courts.Footnote 121
The heart of the problem is perspective. Arbitration practitioners are rightly worried about the abusive invocation of certain public interests reflected in overriding mandatory laws outside the chosen lex contractus Footnote 122—after all, it is arbitration practitioners who raise such issues if that is in the interest of their clients. From this (private law) perspective of pacta sunt servanda, the understandable preference is to allow the arbitration to move forward and limit control to ex post review by courts at the challenge or enforcement stage in order to dissuade obstructive practices. What if there is nothing to the party's argument based on public interests?
From a (public law) perspective of economic regulation, the focus is on the other side of the coin: what if a public interest is at stake? If so, a private transaction should not be allowed to compromise the regulatory efficiency and the public interest in the market which is affected. The public law perspective is compelling for courts—and ultimately unproblematic for arbitration if properly tempered by the limitations that public law itself provides for the protection of private interests, notably proportionality. Arbitration is not hindered provided that the application of overriding mandatory laws is ensured. What potentially remains is unease at having to trust national courts to address and dispose of these issues properly. However, this may be the price which has to be paid in order that national courts will trust arbitrators to uphold public interests in international arbitration.
