Introduction

An arbitral award is final and binding on the parties, and Article III of the New York Convention requires Contracting States to recognise and enforce it, subject only to the exhaustive grounds for refusal in Article V. Yet finality cannot be absolute. Due process and the right of access to justice require that an award remain open to some judicial scrutiny of its legality and fairness. For the losing party, this scrutiny takes the form of an application to annul, or set aside, the award before the courts of the seat of arbitration.

For the prevailing party, by contrast, an award has practical value only once it is given effect. Where the losing party does not comply voluntarily, the prevailing party must seek recognition and enforcement before the national courts of the State where enforcement is sought, typically where the debtor’s assets are located. These courts may, however, refuse recognition and enforcement on limited grounds.

For “annulment” and “refusal of recognition and enforcement” (hereinafter refusal of R&E) of an arbitral award, there are prerequisites evoked in several documents in international scale. New York Convention on the Recognition and Enforcement of Foreign Arbitral Awards (Hereinafter New York Convention) and UNCITRAL Model Law on International Commercial Arbitration (Hereinafter UNCITRAL Model Law) are primary examples to these.

While the New York Convention is legally binding on its Contracting States, the UNCITRAL Model Law serves as a legislative model for national legal systems. For the purposes of this piece, the New York Convention is particularly relevant through Article V, which sets out the grounds on which recognition and enforcement of an arbitral award may be refused. It does not, however, provide grounds for the setting aside of an award. The UNCITRAL Model Law, by contrast, addresses both situations. Article 34 sets out the grounds on which an arbitral award may be set aside, while Article 36 governs the refusal of recognition and enforcement. The grounds contained in Article 36 are, in substance, closely aligned with those set out in Article V of the New York Convention and substantially overlap with the grounds for setting aside under Article 34.

Among these grounds, “conflict with the public policy of the State” requires closer examination, particularly with regard to its definition and scope. The principal difficulty lies in determining the extent to which public policy review may permit review of the merits of the arbitration, thereby calling into question the arbitral character and finality of the decision.

Prohibition of Review of the Merits

Concerning the issue of refusal of R&E, The New York Convention Art. V does not offer local jurisdictions a scope of a review of the merits/révision au fond according to the general acceptance (see: Autonomous Versus Domestic Concepts under the New York Convention, p. 244). In fact, the New York Convention aimed to promote the finality of arbitral awards, thus applying respect towards the decisions of a constituted arbitral tribunal, which results in a general prohibition of a review of the merits during an examination of an arbitral award (see p. 257).

Given that the UNCITRAL Model Law is highly consistent with the New York Convention as regards the refusal of recognition and enforcement, an examination of Articles 34 and 36 of the former instrument reveals that the grounds for both setting aside and refusal of recognition and enforcement are, in general, predominantly procedural or jurisdictional in nature. However, a particular issue arises under Articles 34(2)(b)(ii) and 36(1)(b)(ii). The question to be examined in the following subsection is whether the concept of “public policy” may constitute an exception to the prohibition of review on the merits, and whether national courts may, under the guise of public policy review, interfere with the substantive content of an arbitral award.

While certain doctrinal sources take the view that this concept should not give rise to a révision au fond (see Wolff, para. 561), it has also been observed in the literature that some national courts, including those of England, which will be examined below, have in certain circumstances engaged with matters touching upon the merits of the dispute in case of an annulment (see Born, p. 5405). However, this does not necessarily imply that “public policy” may serve as a basis for review on the merits, especially since some sources characterise the prohibition of such review as a general principle aimed at preserving the autonomy and binding force of arbitral awards.

It is frequently mentioned in judicial decisions that the prohibition of inspecting the merits exists as a fundamental principle.

Among the relevant judicial decisions, the decision of the Casablanca Court of Appeal has been included in this piece for the particularly clear and unequivocal manner in which it addresses the issue. The following passage from the judgment offers a concise and informative summary: “Consequently, the Court cannot examine the merits of the dispute, nor review the relevance of the assessments made by the arbitral tribunal. Therefore, the arguments raised by the applicant on this basis fall exclusively within the sovereign power of the arbitral body, which has definitively ruled on these points, and must therefore be dismissed as inadmissible.”

Public Policy

It should be borne in mind that the notion of the “public policy of the State” leaves a considerable degree of discretion to national courts in the context of international arbitration, particularly when determining whether an arbitral award may be incorporated into the domestic legal order, or may be, set aside.

Majority of the jurisdictions seem to adopt resembling methods to define “public policy” concerning the annulment of arbitral proceedings. According to that, an annulment is possible only if the award directly contravenes a “clearly-defined and mandatory rule which is fundamental to the legal or moral order of the state in question” (Born, p.5386). Given that the term “public policy” is also used both in Articles 36 of the UNCITRAL Model and the New York Convention; this explanation might sound suitable for refusal of R&E according to   the embraced perspective of this paper.

The conceptual uncertainty surrounding “public policy” may give rise to concerns that the notion could be used as a gateway to a review of the merits. Nevertheless, judicial decisions frequently reaffirm that public policy does not, in itself, create an exception to the general prohibition of such review. By way of illustration, the Swiss Federal Supreme Court has emphasised the distinction by rejecting scrutiny of a “distinctively appellate nature”, thereby providing a particularly clear example of the limits imposed on judicial review of arbitral awards.

In this regard, national legal systems in the world have embraced different approaches about the choice of law governing these mechanisms. Most of them preferred to use their own “public policy” due to the extract of “…the public policy of that country/State” (see relative articles in UNCITRAL Model and New York Convention), whereas some of them have chosen an approach in a concept of private international law. “The following sections comparatively analyse the common law and civil law traditions, primarily through the lens of English and French legal systems.”

France and International Public Order

In French law, the notion of “public policy” is generally expressed through the concept of “ordre public international” in proceedings concerning the setting aside of arbitral awards or the refusal of their R&E. Although the two concepts are broadly comparable in function, “ordre public” reflects the particular doctrinal vocabulary and structure of civil law systems rather than merely constituting a linguistic equivalent of the common law notion of “public policy”. Thus, in order to underline the semantic difference of these terms; “public order” will be preferred instead of “public policy” during translation of “ordre public” under the French Law.

Under the French Code of Civil Procedure (CPC), the condition for recognition and enforcement and the annulment of an arbitral award are addressed in separate provisions. According to CPC Article 1514:

“Arbitral awards are recognized or enforced in France if their existence is established by the party relying on them and if such recognition or enforcement is not manifestly contrary to international public order (ordre public international).”

Also, according to CPC Article 1520:

“An appeal for annulment is only available if:

(…)

5° Recognition or enforcement of the award is contrary to international public order.”

These two provisions indicate that the national procedural code of France chose to consider both situations procedurally distinct but substantively interconnected. Also, the CPC Art. 1525, reinforces this statement as well:

“The decision ruling on a request for recognition or enforcement of an arbitral award rendered abroad is subject to appeal.

(…)

The Court of Appeal may refuse recognition or enforcement of the arbitral award only in the cases provided for in Article 1520.”

Then, it is necessary to evaluate the “ordre public international” criteria and answer the question whether this causes an interference to the merits of the case or not.

The terminological explanations and doctrinal discussions concerning the notion of “ordre public international” as incorporated into the French legal system do not appear to provide a single, precise and uniform definition. According to Simma, the notion of “ordre public international” under the French Code of Civil Procedure can only refer to the French conception of international public policy, namely “a set of values a breach of which could not be tolerated by the French legal order, even in international cases” (see footnote 34). Thara appears to support a similar understanding in his monograph (see Section 2, I, B, 1–2). The same monograph (see Section 2, I, B, 3), as well as the piece of Loquin (see para. 437), also appears to include lois de police—that is, overriding mandatory rules—within the broader scope of ordre public international, particularly where such rules are intended to govern situations involving an international or external element.

Despite attempts to define “ordre public international” in abstract terms and in a manner encompassing other branches of law, it may be more effective to turn to judicial decisions in order to understand how the concept operates in practice, particularly in private international law as a standard of review in proceedings concerning the annulment of arbitral awards or the refusal of their recognition and enforcement.

The Cour de Cassation seems to inspect the violation of public order through assessing the existent proofs, though, not being strictly limited to them. In Belokon v. Kyrgyzstan, a case of annulment concerning the question of money laundering and its conflicting nature against international public order; the Supreme Court emphasized: “The Court Of Appeal rightly held that such research, conducted for the protection of international public order, was neither limited to the evidence produced before the arbitrators nor bound by the findings, assessments and qualifications made by them, its sole function in this regard being to ensure that the production of the evidence before it respected the principle of adversarial proceedings and that of equality of arms.” (see p. 5).

In light of this passage, the Cour de Cassation appears to accept a relatively broad form of judicial analysis, thereby raising the question whether such scrutiny may amount to a review on the merits. It is therefore necessary to determine whether the Court acknowledges any possibility of a révision au fond.

In Botas v. Tepe, however, the French Cour de cassation expressly reaffirmed that the annulment judge may not engage in a review on the merits. The Court held that:

“First, given that, in noting that BOTAŞ merely criticised the arbitrators for having disregarded the principle of good-faith performance of agreements, the Court of Appeal, which carried out the allegedly omitted inquiry, correctly inferred that, in the absence of any demonstration as to how the solution adopted by the arbitral tribunal violated international public policy, the ground for annulment in fact sought a review on the merits of the partial award, which is prohibited for the annulment judge.”

(Cour de Cassation, First Civil Chamber, Judgment No. 1592 F-D, Appeal No. A 11-13.269, 19 December 2012, p. 148).

When the two decisions are read together, the position of the French Cour de Cassation appears to be that judicial scrutiny of an arbitral award may, in terms of its intensity, extend beyond the evidence presented before the arbitral tribunal and may therefore come close to a review of the merits. At the same time, the Court maintains that such scrutiny does not formally constitute a révision au fond, since the annulment judge is not empowered to reopen the merits of the underlying dispute.

Appearance of the Concept in English Law

English arbitration law, as one of the principal representatives of the common law tradition, is primarily governed by the Arbitration Act of 1996 (hereinafter “Arbitration Act 1996” ), which was partially modified by the later Arbitration Act of 2025.

Section 67 of the Arbitration Act 1996 provides a mechanism for challenging an award on the ground of substantive jurisdiction. Section 68(2), on the other hand, sets out a number of grounds under the heading of “serious irregularity”, most of which concern procedural or formal matters. Section 68(3) further specifies the measures that the court may take where a serious irregularity is established, including the power to “set aside” the award, which may, in this respect, amount to annulment.

It should also be noted that, among the grounds of “serious irregularity”, paragraph (g) is particularly noteworthy, as it refers to “the award being obtained by fraud or the award or the way in which it was procured being contrary to public policy. ”The notion of “public policy” employed here may be regarded as functionally comparable to the concept of ordre public.

As another option, S.69 offers a possibility of “appeal” in case of agreement of parties or with the leave of the court. In the latter situation; S.69(3) enounces that:

“Leave to appeal shall be given only if the court is satisfied –

(…)

(c)that, on the basis of the findings of fact in the award –

(i)the decision of the tribunal on the question is obviously wrong, or

(ii)the question is one of general public importance and the decision of the tribunal is at least open to serious doubt, and

(d)that, despite the agreement of the parties to resolve the matter by arbitration, it is just and proper in all the circumstances for the court to determine the question.

(…)”

Although these criteria are distinct from the public-policy ground under s.68, they raise a related question for the purposes of the present discussion: to what extent can a court assess whether an arbitral tribunal was “obviously wrong” or its decision “open to serious doubt” without engaging with the merits of its legal reasoning? This suggests that the conceptual ambiguity surrounding ordre public, as well as the debate concerning review of the merits, may not be confined to continental European legal systems.

Indeed, how can a court determine whether an error is “obvious”, whether a question is of “general public importance”, or whether a decision is “open to serious doubt”, without engaging, at least to some extent, with the merits of the underlying dispute? Also, what does “being just and proper in all circumstances for the court to determine the question” mean?

In the case of HMV UK Ltd v. Propinvest Friar Ltd Partnership, the Court indicates that there shall be no admission by the virtue of S.69 unless the arbitrator made a major intellectual aberration. The Court also elaborated that the term “obvious” must fulfill the quality of being a severely egregious error, ensuring a high standard.

A useful illustration of the limits of judicial scrutiny under S.69 can be found in Sharp Corp Ltd v. Viterra BV. In this case, the Supreme Court clarified that an appeal under S.69 is confined to questions of law and must proceed on the basis of the findings of fact contained in the arbitral award. The Court further emphasised that an appellate court has no jurisdiction to correct errors of fact or to make its own factual findings. At most, it may recognise a factual finding that, although not expressly stated in the award, can necessarily be inferred from the findings already made by the arbitral tribunal. Such an inference does not amount to the court making a new finding of fact, but merely to identifying a finding that must already be implicit in the award (see paras. 71-74). It is also worth noting that Jacobs J considered the appeal to raise a question of “general public importance”, although no precise definition or detailed elaboration of that notion is provided in this context (see para. 43).

Furthermore, although the Arbitration Act 1996 had not yet entered into force at the time of the dispute, Lord Diplock’s observations in The Nema case also provide useful guidance as to whether arbitral review may amount to a révision au fond and as to the extent to which factual findings should remain confined to those made in the arbitral award (see pp. 4-5): “Where, as in the instant case, a question of law involved is the construction of a ‘one-off’ clause the application of which to the particular facts of the case is an issue in the arbitration, leave should not normally be given unless it is apparent to the judge upon a mere perusal of the reasoned award itself without the benefit of adversarial argument, that the meaning ascribed to the clause by the arbitrator is obviously wrong: But if on such perusal it appears to the judge that it is possible that argument might persuade him, despite first impression to the contrary, that the arbitrator might be right, he should not grant leave; the parties should be left to accept, for better or for worse, the decision of the tribunal that they had chosen to decide the matter in the first instance (…) In deciding how to exercise his discretion whether to give leave to appeal under section 1 (2) what the judge should normally ask himself in this type of arbitration, particularly where the events relied upon are ‘one-off’ events, is not whether he agrees with the decision reached by the arbitrator…”

As regards S.69(3)(d), the decision in The Coal Authority v. F W Davidson is also instructive. In substance, the Court appears to treat the requirement that it be “just and proper” for the court to determine the question as one concerned with the legal nature of the issue, thereby implicitly indicating that the provision does not create an independent basis for a broader review of the merits. In reaching that conclusion, the Court also addressed the “obviously wrong” standard under S.69(3)(c) by reference to the circumstances of the particular case in Section E, paragraph 40.

“Going back to section 69 for a moment, it is, therefore, clear that his erroneous decision fundamentally affected the rights of both parties and concerned the principal question that he had been asked to determine. In these circumstances, since the issue is a pure point of law, it is just and proper for the court to determine the issue afresh, notwithstanding the agreement of the parties to go to arbitration. Accordingly, it seems to me that the Authority has made out every necessary ground under section 69 of the Arbitration Act 1996 and I, therefore, grant them permission to appeal.”

The same decision also addresses the notion of “general public importance” not through an abstract or universally applicable definition, but by reference to the circumstances of the case. In Section B, paragraph 7, the Court also decided that the parties’ agreement that the issue was one of general public importance.

“I should record at this stage that the principal issue on this application (…) is agreed by the parties as being a point of general public importance. I respectfully agree with that. Thus, for the purposes of leave, the test is whether the arbitrator’s decision is at least open to serious doubt, a broader requirement than that propounded by Lord Diplock in The Nema.”

As noted in the section titled “Prohibition of Review of the Merits”), it has been observed in the literature that certain national courts may interpret public policy in a manner that allows for a degree of scrutiny approaching a review of the merits, with England being cited as one such example. However, when the domestic mechanisms that could potentially permit such scrutiny are examined more closely under this section, that proposition appears to be less firmly established than it might initially seem.

Indeed, the concepts of “obviously wrong,” “general public importance,” and “open to serious doubt” are generally understood in the doctrine as relating primarily to the application and interpretation of legal rules rather than to a reassessment of the underlying facts. Accordingly, although the legal reasoning of the tribunal may be analysed, the factual basis of the dispute is not reopened; in this sense, S.69 does not amount to a full révision au fond.

Conclusion

Ultimately, the question of whether national courts may engage with the merits when reviewing arbitral awards that benefit from a presumption of validity – whether in annulment proceedings or in proceedings concerning the refusal of recognition and enforcement – centres, at the international level, on the notion of public policy; in French law, on ordre public international; and, in English law, partly on public policy and partly on the concepts embodied in s.69 of the Arbitration Act.

Although the prohibition of review on the merits is generally recognised in the literature, the semantic uncertainty surrounding these concepts makes a more detailed and case-specific examination necessary. In French law, the Cour de cassation’s refusal to confine judicial scrutiny strictly to the factual findings made by the arbitral tribunal tends to blur the boundary separating public-order review from a review of the merits. At the same time, this appears to sit in some tension with the Court’s express position that the annulment judge may not engage in a révision au fond.

English law, by contrast, appears to approach the issue primarily through the review of legal questions, while maintaining that factual findings must remain confined to those made by the arbitral tribunal. Nevertheless, with regard to the standards embodied in s.69(3)(c) – namely the “obviously wrong” threshold and, where the question is one of “general public importance”, the alternative requirement that the tribunal’s decision be “open to serious doubt” – the case law appears to clarify their operation primarily through context-specific thresholds rather than through exhaustive definitions. Section 69(3)(d) introduces a further and distinct safeguard by requiring that it also be “just and proper in all the circumstances” for the court to determine the question despite the parties’ agreement to arbitrate.

The comparison therefore suggests that the prohibition of review on the merits is not defined solely by the intensity of judicial scrutiny. French law appears willing, in the context of international public-order review, to permit a degree of independent factual assessment without characterising it as a révision au fond, whereas English law under S.69 permits scrutiny of the tribunal’s legal reasoning while expressly preserving its factual findings. The decisive boundary may therefore lie not merely in how deeply a court examines an award, but also in the object and legal basis of that examination.