India's arbitration system has structural problems that no amount of legislation has been able to fix. In the second episode of Arbitration After the Bench, Justice AK Sikri offers no comfort. He offers candour.

Justice Sikri speaks with Bar & Bench's Debayan Roy in a conversation that moves from the philosophical foundations of arbitration to the government's retreat from arbitration clauses, from the Gayatri Balasamy judgment to geopolitical shocks now flooding international tribunals with force majeure claims. And he says, without hesitation, that making India an international arbitration hub will remain a distant dream.

Justice Sikri retired from the Supreme Court of India in March 2019. Within months, he was sworn in as an international judge of the Singapore International Commercial Court (SICC), where he has sat for six years alongside judges from England, France, the United States, Germany, China, Japan and Hong Kong. He simultaneously maintains an active arbitration practice, presiding over domestic and international disputes.

[Watch Video]

Edited excerpts follow.

Debayan Roy [DR]: Your career has taken you across three worlds that rarely speak to each other: the bench of the Supreme Court of India, the SICC and the arbitration room. What does each teach you that the other two cannot?

Justice AK Sikri: The Supreme Court vests you with powers to decide constitutional issues, social justice matters and questions of governance where there is enormous judicial discretion, including under Article 142. Those tools are simply not available in arbitration. The arbitral tribunal is a creature of the arbitration agreement. Its jurisdiction is derived from that agreement alone. Section 28(2) of the Arbitration Act specifically excludes the principles of equity. You go strictly by the contract.

SICC is a unique system. It is an international court where judges are drawn from both common law and civil law jurisdictions. That fusion of jurisprudence allows us to create a different kind of international commercial law, adopting best practices from both systems. It is also a flexible court: parties have considerably more say in procedural matters than in an ordinary domestic commercial court.

DR: Most Supreme Court judges who retire and enter arbitration have spent their careers hearing appeals, deciding on principles of law. Very few have had trial court experience. Arbitration at its heart is a trial. Are retired appellate judges actually equipped for it?

Justice Sikri: Many judges from High Courts with original side jurisdiction do have trial experience. I practised simultaneously in the trial court and High Court when I joined the profession. But you are right that judges from states without original side jurisdiction may not. And in most arbitration, the cases are fact-heavy. How to record evidence, marshal facts, analyse expert testimony on construction delays or quantum damages - these require real expertise.

Even if you lack that expertise, you must try to acquire it, through training if needed. One should not feel shy. There are many who think that 20 or 25 years on the bench makes them competent for arbitration. That should not be the approach. Every day, you learn. That applies to arbitration also.

The primary reason [why former judges are chosen as arbitrators] is faith. Litigants and business houses feel that retired judges will be more impartial, more rigorous. A survey by FICCI or ICA about seven or eight years ago found that almost 65 to 70 percent of respondents said they want retired judges. But I agree that it should not be only judges or senior lawyers. Anyone can be an arbitrator. Particularly in technically complex cases, a retired engineer may be far better placed than a judge to analyse construction delays or quantify damages. I have been the presiding arbitrator in cases where a technical member was on the tribunal and the assistance on technical matters was a great relief.

In England and Singapore, there is also a talk of gender diversity, wanting more inclusion of women arbitrators...That is good for various reasons. The combination of legal and technical expertise is important...

DR: When a retired Supreme Court judge passes an award, it is reversed or set aside under Section 34 or Section 37 by a judicial magistrate or a district judge. Does this not expose a fundamental tension in how we have structured arbitration?

Justice Sikri: I do not think so. When you accept the role of arbitrator, you know your award is subject to challenge under Section 34 and depending on the jurisdiction, the matter can go to a district judge. You accept the role knowing this ecosystem. I have seen for these reasons that there are many judges who do not want to become chairpersons of these tribunals or do not want to accept arbitration work, saying why should I do the work and have my reasoning overturned.

To answer you candidly: when I sit writing an award, does the thought come that a district judge may set it aside? No. I write the award on what I believe is the correct position in law. I have to come to a correct conclusion on the facts in my wisdom, without bias and treat both parties fairly.

DR: You had authored the Ayyasamy judgment in 2016 where you drew a difference between fraud simpliciter and serious fraud. Vidya Drolia in 2021 took the law further ahead. Now, when you see parties raising the ground of fraud as a ground to challenge the jurisdiction of an arbitrator, do you still apply Ayyasamy?

Justice Sikri: The substratum remains. In Ayyasamy, our main focus was that a mere allegation of fraud should not be sufficient to exclude a dispute from arbitration. Unless it is of an egregious nature meeting the penal definition of fraud, it should be arbitrable. Avitel fine-tuned this by adding a second test: whether the fraud has a public law element.

In Drolia, the emphasis shifted to procedure: should the court decide the fraud issue at the Section 8 or Section 11 stage? Drolia said no. Apply only the eye of a needle test at that stage and leave the substantive question to the tribunal. When such cases come before me as an arbitrator, I apply the same two tests from Ayyasamy and Avitel.

DR: In December 2018, the Government of India had nominated you to the Commonwealth Secretariat Arbitral Tribunal in London while you were still a Supreme Court judge. When the matter became public, you withdrew your nomination. Has India resolved this issue of post-retirement judicial appointments or have we just ignored it?

Justice Sikri: It is a very tricky situation. It is always talked about and it is said that judges should normally not accept any such positions after retirement. The perception is that in their last months or maybe last one year before retirement, in order to get some good assignment, they may start deciding cases in favour of the government...maybe in some cases it has happened also.

First, let me clarify: that was not an arbitration position. It involved deciding employment disputes of Commonwealth Secretariat staff. There was no remuneration. I was even reluctant when it was offered. The controversy that arose was totally unrelated to the nature of the appointment. When I saw what people were thinking, I immediately said I should not be given this and withdrew.

The broader issue has always remained and it is tricky. Many tribunals and quasi-judicial bodies, under statute, must be headed by retired judges. If these positions must be filled by former judges, how can you say judges should not accept post-retirement appointments? One good solution was suggested by Arun Jaitley when he was in opposition: increase the retirement age to 70 and send judges on deputation to these tribunals while they are still sitting judges. That can solve the problem. Otherwise, any appointment will always be scrutinised for which judgments the judge gave in favour of the government before retirement. Unfortunately, there have been some appointments in these tribunals also of those persons whom people call pro-government in their approach.

DR: The government has removed arbitration clauses from infrastructure and public procurement contracts worth ₹10 crore or more. However, under the MSME Act, arbitration is statutorily mandated and the government cannot opt out. Are we adopting arbitration only when the outcome suits us?

Justice Sikri: I would say it is a knee-jerk reaction. On one hand, we speak of ease of doing business and India as an arbitration hub. On the other, because some high-value awards came against the government, you remove the clause. ₹10 crores today is nothing. Any significant contract dispute now goes to court, which foreign investors will not accept.

The real problem is internal. Government officers know that a contractor is entitled to something. But they will not recommend settlement because tomorrow someone may ask why they recommended it. The fear of CAG, CVC and vigilance inquiry is so paralysing that the matter goes to arbitration instead. The contractor who might have accepted 60 or 70 rupees for a legitimate 100-rupee claim now files for 500 rupees. Interest runs for five, seven, ten years. And then the government complains about large awards.

I must add that not necessarily every time are awards given against the government. Many times, we have rejected the claims totally and allowed the counterclaims of the government...I may add here very frankly that there is an issue of arbitrators' bias. There are issues about the impartiality and integrity of some arbitrators also. Those should be taken up at a different level. Party autonomy gives you the power to appoint a person, so appoint a person of impeccable integrity. Many times, I have seen awards given against the government by those who were earlier in the government itself.

DR: Section 34 courts continue to revisit arbitral merits under the language of patent illegality and public policy, despite decades of Supreme Court corrections. Is the problem the law or the culture?

Justice Sikri: The law is not the problem. Section 34 is clear: awards cannot be reviewed on merits. The Section 34 court is not a court of appeal. Even interpretation of a contract clause given by the tribunal is binding. The grounds are very limited.

Lord Mustill, in his foreword to OP Malhotra's book on arbitration, captured this very well. He said that when a very bad award comes before a court, the judge's instinct is to do justice. But how? The judge then reaches for patent illegality as the legal hook. In doing so, to correct a genuine error in one case, a proposition of law is laid down which in future is misused or misapplied in routine cases where there was nothing wrong with the award. Bad cases always lead to bad law. By and large, the Supreme Court has sent very clear signals in judgments over the last 10-15 years that courts will interfere in the least possible manner. But some cases come where it happens and those send a wrong signal.

DR: in April 2025, a Constitution Bench held that the courts have a limited power to modify arbitral awards. However, Justice Viswanathan in his dissenting judgment said that it leads to judicial legislation. Does the pragmatism of the majority comfort you or does the dissent alarm you more?

Justice Sikri: I would lean in favour of Justice Viswanathan's dissent. The pragmatism of the majority is understandable: why send an award back and reconstitute a tribunal for years when a simple error can be corrected? But it is easy to state that proposition. When it comes to applying it in a given situation, where to draw the line will become very, very difficult. Justice Viswanathan rightly says we need law specifying the contours of interference. As of today, there is no law.

The earlier position - that courts have no power to modify - was the correct interpretation on the letter of the law. By laying down this principle, we are opening a Pandora's Box. If law is there, law will specify clearly the contours of interference and under what circumstances you can modify, which would then be applied with more certainty. Today, that certainty has gone.

DR: The world is now dealing with a US tariff war, an Israel-Iran military conflict, a blocked Strait of Hormuz and sanctions that are preventing parties from paying arbitral awards. Arbitration was designed for two parties and one contract. Is it built for a world where a contract fails for a cascade of simultaneous geopolitical shocks?

Justice Sikri: War-like situations have always been treated as force majeure and these issues will arise. COVID gave us force majeure claims. Whenever such natural or manmade calamities are there, force majeure clauses will be pressed into service. Generally speaking, a war-like situation would fall within force majeure under both international law and Indian law, which is broadly in tune with international law on this.

But another interesting question may arise: contractually, a particular event is not treated as force majeure, but the law treats it as such. So law can still be invoked because you cannot contract out of legal rights.

But tariffs are a different matter entirely. There are judgments that if a contract has merely become economically more difficult, that is no ground to avoid it. Suppose a tariff is increased on US imports. Because of that, the rate at which I agreed to sell now puts me at a loss. That may not be a ground to invoke force majeure. The force majeure doctrine in India depends heavily on what events are listed in the contract itself. If an event is not listed, you cannot invoke it contractually, though law can still be invoked where you cannot contract out of statutory rights.

DR: From inside the SICC, what does India look like as an arbitration seat to international parties and counsel?

Justice Sikri: India's lawyers are doing extremely well now in international arbitration. I have sat with arbitrators from other jurisdictions who had genuine praise for Indian counsel. As far as the Bar is concerned, we have lawyers who can compete with anyone in the world.

But the ecosystem is the problem. The main issue is enforcement of awards. The investor knows that if an award is challenged in Singapore, it will be decided in one to two years. In India, nobody knows how long it will take. That uncertainty, coupled with cases where courts have interfered even when the Supreme Court preaches restraint, creates an environment of unpredictability. The DMRC curative was exactly one such case. Any businessman, whether Indian or foreign, cannot accept delays and uncertainty. If we are not able to tackle these two aspects, I am very frank in saying that making India an international hub for arbitration would remain a distant dream.

DR: The SIAC Rules 2025 introduced ex parte emergency relief, mandatory third-party funding disclosure and a 90-day award drafting deadline. Each fills a gap that India's domestic framework has not addressed. India's 2024 Amendment Bill is silent on third-party funding. Is India watching Singapore or ignoring it?

Justice Sikri: I do not think India is ignoring Singapore. The 90-day award timeline has been in the ICC Rules for many years; Singapore has now adopted it. Emergency arbitration provisions exist in most international institutions including our own Mumbai Centre For International Arbitration (MCIA). The distinction is important: these are institutional rules, not statutory provisions. In other countries also, it is institutions, not legislatures, which have made these rules. In India's context, because 60 to 70 per cent of arbitration is ad hoc and institutional rules do not apply, the amendment to the Act becomes more relevant.

On third-party funding (TPF), there was a Privy Council judgment from the 1930s in the Indian context holding that funding for legitimate purposes is acceptable unless extortionate. TPF has now become an acceptable norm internationally. But the conflict of interest dimension is real: many times it has transpired that a funder had a relationship with one of the arbitrators, creating conflict issues.

Another good point of third-party funding is that when a third party would fund, they look into the merits of the case deeply. Only those cases they will take up because they have to have some share when the award is given in favour. The pros and cons are both genuine.

DR: On one side, retired Supreme Court judges sit at the apex of the arbitration market. On the other, very young lawyers with just a few years of practice are being appointed as arbitrators. No minimum standards, no mandatory training, no accountability. Is this not a serious crisis of standards?

Justice Sikri: You are right that appointing a very young person without experience may not be good. But in other countries, young people are appointed if they have undergone proper training. In India, there is an Institute of Arbitrators. Why should not youngsters receive that training and learn how to conduct proceedings and write awards? We should encourage them. They are the future. There are also many Section 11 disputes involving only a few lakhs of rupees where it would be disproportionate to appoint a retired High Court judge. A junior advocate taking that matter gets real experience, which is good for the ecosystem. Having said so, proper training should be made mandatory before appointment.

DR: The Singapore Court of Appeal set aside an award in an Indian railway dispute because substantial portions of the reasoning were reproduced from earlier awards. Global guidelines now say that arbitrators must not delegate their personal mandate to any AI tool. Where do you draw the line?

Justice Sikri: Ultimately, arbitrators must apply their own mind. AI can only be a facilitator. I would go further and say: do not depend upon AI even for your conclusions in a particular case. You may use it to summarise a 500-page pleading into 15 pages. But subject that summary to your own scrutiny. We have seen AI committing blunders, citing cases that do not exist. Everything must be verified by you. For analysis of evidence, analysis of law and your conclusions, you must apply your own mind.

On the railway award: it involved a standard format dedicated freight corridor contract. The same contract came before the arbitrator a second time. He had already interpreted it. His reasoning was that if I have already interpreted these provisions, why should I write different words? But the mistake was that while copying, he copied even the arguments of parties that were never actually made in the new case. And the new case had a different nuance that he never addressed. Had he addressed that nuance, his result would not have been the same. It was clearly without application of mind. I am surprised the other two arbitrators on the tribunal did not point this out. It damaged the image of Indian arbitrators. It became ammunition for those who were already looking for reasons to condemn.

DR: India has amended its Arbitration Act in 2015, 2019 and 2021. There is now a Draft Amendment Bill for 2024 proposing an Appellate Arbitral Tribunal. You have said India needs to breathe. Is the 2024 proposal breathing or hyperventilating?

Justice Sikri: Some amendments became necessary because of judgments, because of interpretation issues, because of practical experience. That is understandable. But over-activism in legislative amendment is also there. Some aspects could be addressed on the judicial side without requiring parliament.

On the appellate arbitral tribunal specifically, conceptually, it is a good idea. There are many high-stakes private disputes where parties genuinely want one more chance at getting it right. In investment treaty arbitration, many global south countries feel ICSID tribunals have treated them as if sovereignty does not matter and want a second look. SICC now offers itself as that forum.

But in the Indian context, the appellate tribunal comes with a problem: orders of an appellate tribunal have no precedential value, just like awards. At least today, Sections 34 and 37 and the Supreme Court create jurisprudence. An appellate tribunal will not. You can cite a judgment but you cannot cite an award. The same thing will apply even to the appellate tribunal. So that is a concern.

DR: Arbitration was introduced in India as an escape from court delays - faster, cheaper, more final. But today, you appoint an arbitrator and that is contested. The award goes to Section 34, then Section 37, then the Supreme Court. Parties fight in arbitration and in courts, simultaneously and sequentially. Has arbitration in India become just another layer of litigation?

Justice Sikri: To a great extent, yes, no doubt about it. We are a litigating community. As I always say, we are all Amartya Sen's Argumentative Indian: we are never satisfied and want to argue further. Even after an award, parties go to Section 34, then Section 37, then Supreme Court by SLP.

But the statement would be incomplete without empirical data. We only know about awards that are challenged. Has anyone studied out of 100 awards given, how many are challenged? It may be 20 or 30 per cent. 60-70 per cent may never be challenged. I have seen many cases settle after the award, many Section 34 challenges settle mid-way. Just 2 weeks ago, I had a very high-stakes matter where after arguments were heard and parties were to file closing submissions, we were told that they had settled. We passed an award in terms of the settlement. In the last 2-3 years, in 4-5 of my own international arbitrations, parties settled after full proceedings...

...The perception that every award is challenged exists because most high-stakes awards are challenged. But a settlement culture is also growing. To answer your question categorically, there has to be an empirical study. We do not have one.