Arbitrability of Fraud in India – Is Ayyasamy only about “Seriousness”?

[Shubham Jain and Prakshal Jain are V Year B.A., LL.B. (Hons.) students at National Law School of India University, Bangalore]


The question of arbitrability of fraud in case of domestic seated arbitration in India was addressed by the Supreme Court in A. Ayyasamy v. A. Paramasivam (“Ayyasamy”). Ayyasamy has been subjected to much discussion and criticism here, here, here and here. There has been an outcry around the requirement of “seriousness of allegations of fraud” laid down by the Court and its application as the all-important test to determine whether a dispute is arbitrable or not. However, commentators have missed an essential qualifier that the Court imposed in addition to the requirement of “seriousness”. In this post, we argue that, as per Ayyasamy, only those disputes which involve very serious allegations of fraud and have an implication in the public domain are non-arbitrable. Such an interpretation is also consistent with the law on subject matter arbitrability in India.

Law on Subject Matter Arbitrability in India: The Booz-Allen test

The legal position with respect to the arbitrability of a particular subject matter was laid down by the two judge bench of the Supreme Court in Booz-Allen & Hamilton Inc v. SBI Home Finance (“Booz-Allen”). It was held: “Generally and traditionally all disputes relating to rights in personam are considered to be amenable to arbitration; and all disputes relating to rights in rem are required to be adjudicated by courts and public tribunals, being unsuited for private arbitration.” The Court reasoned that certain categories of disputes are reserved exclusively for public fora as a matter of public policy. While there might be shortcomings of the test laid down in Booz-Allen, it cannot be denied that since then the test has been widely accepted in India and the bench in Ayyasamy (being a two judge bench) was bound by the decision in Booz-Allen.

Radhakrishnan: The Source of “Seriousness”

Prior to Ayyasamy, the test of “seriousness” to determine arbitrability of fraud was accepted by the Supreme Court in N. Radhakrishnan v. Maestro Engineers (“N. Radhakrishnan”). In doing so, the Court had relied on a few High Court decisions and a three judge bench decision of the Supreme Court (Abdul Kadir Shamsuddin Bubere v. Madhav Prabhakar Oak), which was rendered in context of the erstwhile Arbitration Act, 1940. While the Court’s reliance on Abdul Kadir was misplaced (246th Report of the Law Commission of India, ¶50), N. Radhakrishnan was accepted as the authority on the subject and was followed in subsequent cases.  

Ayyasamy: The Art of Balancing N. Radhakrishnan and Booz-Allen

In addition to Booz-Allen, the Court in Ayyasamy was also bound by the decision of the two judge bench in N. Radhakrishnan. Inevitably, the Court did lay considerable stress on the requirement of “seriousness of allegations of fraud” for a dispute to be arbitrable. However, while doing so, the Court performed a balancing act by incorporating the principles laid down in Booz-Allen into the test of determining arbitrability of fraud. Reading the judgement in its entirety shows that, for both Justice Sikri and Justice Chandrachud, the requirement of the fraud having an implication in public domain was an important consideration. The following passage from Justice Sikri’s opinion (¶20) is instructive:

It is only in those cases where the Court, while dealing with Section 8 of the Act, finds that there are very serious allegations of fraud which make a virtual case of criminal offence or where allegations of fraud are so complicated that it becomes absolutely essential that such complex issues can be decided only by civil court on the appreciation of the voluminous evidence that needs to be produced, the Court can sidetrack the agreement by dismissing application under Section 8 and proceed with the suit on merits…..Reverse position thereof would be that where there are simple allegations of fraud touching upon the internal affairs of the party inter se and it has no implication in the public domain, the arbitration clause need not be avoided and the parties can be relegated to arbitration. While dealing with such an issue in an application under Section 8 of the Act, the focus of the Court has to be on the question as to whether jurisdiction of the Court has been ousted instead of focusing on the issue as to whether the Court has jurisdiction or not. It has to be kept in mind that insofar as the statutory scheme of the Act is concerned, it does not specifically exclude any category of cases as non-arbitrable. Such categories of non- arbitrable subjects are carved out by the Courts, keeping in mind the principle of common law that certain disputes which are of public nature, etc. are not capable of adjudication and settlement by arbitration and for resolution of such disputes, Courts, i.e. public for a, are better suited than a private forum of arbitration.” (emphasis supplied)

A close reading of the above passage suggests that the Court has indeed laid down the requirement of “very serious allegations of fraud” while determining arbitrability. However, the same has been qualified by the requirement that the alleged fraud must have an implication in the public domain i.e., the dispute must be dealing with actions in rem. Those matters which deal with internal affairs of the party i.e., actions in personam, can be submitted to arbitration without any difficulty. Similar tenor is found in Justice Chandrachud’s judgement as well. In fact, he seems to indicate that the underlying rational of the Court in N. Radhakrishnan was grounded in public policy as well (¶14). Therefore, it can be said that the Court has incorporated the test of Booz-Allen where it was held that disputes dealing with actions  in rem are generally non-arbitrable.

In fact, this requirement of “implication in the public domain” was acknowledged by the Delhi High Court in Rishabh Enterprises v. Ameet Lalchand Shah. However, in that case, while Justice Endlaw considered it to be one of the requirements, he fell short of characterising it as an essential requirement. In our opinion, the relevant portion of Ayyasamy as highlighted above makes it very clear that absence of implication in public domain would mean that the dispute is arbitrable, thereby making it an essential requirement for declaring fraud to be non-arbitrable.

The Confusion Surrounding Complexity Simplicitor

It is important to clarify another myth prevailing with respect to the judgement. It has been argued that Ayyasamy’s reasoning behind making serious allegations of fraud non-arbitrable was the fact that these disputes involve complicated issues of fact which require reviewing of voluminous evidence. We believe that this is not the case. First, as we have argued above, the rationale of the Court is grounded in public policy concerns and, even if a dispute involves complicated issues, courts will still see if it has implications in public domain. Second, the Arbitration and Conciliation Act, 1996 ensures that complexity due to the need for appreciation of voluminous evidence can never be a concern for an arbitral tribunal. Section 27 of the Act allows the arbitral tribunal to apply to the Court for assistance in taking evidence. This section has been used abundantly by the Supreme Court as well as High Courts. For instance, see Alka Chandewar v. Shamshul Ishrar Khan; .

Therefore, in light of section 27, there seems to be no need to declare fraud inarbitrable only because of complexity. This further strengthens our argument that the dispute will be non-arbitrable only when allegations of fraud are very serious and have an implication in public domain.


We acknowledge that the Supreme Court laid substantial emphasis on the seriousness of allegations for a dispute to be declared non-arbitrable. However, it is equally important to note the underlying importance that the Court accorded to the necessary requirement of implications in public domain. While Ayyasamy did not elaborate upon the scope and meaning of the phrase “implication in the public domain”, recognising it as an essential requirement while determining arbitrability can be a helpful tool in reducing the wide scope accorded by another ambiguous requirement, “very serious allegations of fraud”.

Unfortunately, this requirement has been overlooked in academic discourse and otherwise. The neglect of this important qualifier poses a major threat. This is because any and all cases with serious allegations of fraud, even those which merely deal with internal affairs of the party and have no in rem impact, may be relegated to litigation. This would be inconsistent with the basic principles of non-arbitrability laid down and approved by the apex court in previous cases. Only when the requirement of implication in public domain is read as an integral part of the test will we be able to avoid a situation where parties are devoid of the option of arbitration merely because one party was able to show serious allegations of fraud.

– Shubham Jain and Prakshal Jain

About the author


  • Hello Shubham and Prakshal,

    Thank you for your insightful article. I would concur with you on the fact that capability of a tribunal to review voluminous evidence need not be the factor to determine inarbitrability and I would also agree that the public policy based test is also implicit in the Ayyasamy judgment. However, in the scheme of the Arbitration and Conciliation Act, 1996 following a public policy test for determining inarbitrability is an erroneous method. We have argued this in our article [Semantic and doctrinal restructuring of ‘arbitrability’: examining Brekoulakis’ arguments in the Indian context, 33(4) Arbitration International 609 (2017)]

  • Hi Shubham and Prakshal,

    You make an interesting point about Ayyasamy laying down a twofold test for arbitrability of fraud- (i) public policy considerations, and (ii) serious allegations of fraud. I take your point to the extent that the test definitely incorporates the element of public policy; what I disagree with you on is the extent of presence and dominance of that element. Let me illustrate:

    If serious allegations of fraud have been made, but these allegations do not have any implications on the public domain (or involve any rights in rem), the dispute would be arbitrable, from my understanding of your analysis. However, switch things around– if the allegations of fraud have implications on the public domain (or involve any rights in rem), but are simple and straightforward rather than serious, would the dispute still be arbitrable? Based on my understanding of Booz-Allen, it would not, simply because matters having implications on the public domain are not arbitrable, irrespective of the nature of allegations of fraud. Would this not convert your twofold test of (i) public policy considerations, and (ii) serious allegations of fraud, into the unitary test of public policy considerations already established under Booz-Allen? Would this not render irrelevant the nature of allegations of fraud? I only seek to extend your argument to what I see as its conclusion– serious allegations or not, public policy implications will definitely render a dispute inarbitrable, and the lack of public policy implications will render a dispute arbitrable.

    In fact, I belive that this understanding of Ayyasamy (the understanding that the nature of the allegations made is not only of limited importance, but of no importance) would better solve the major threat of incorrect refernce to litigation that you point out.

    There is, however, one hitch with the understanding of Ayyasamy that I propose– it completely overlooks and disregards the fact that the court in Ayyasamy clearly distrusted the ability of arbitrators to evaluate serious allegations of fraud involving voluminous evidence.

    I guess there are no easy answers?

Subscribe to Blog via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Recent Posts


Recent Comments


web analytics