[Prashant Narang is Deputy Director – Research and Programmes and Renuka Sane is Managing Director at TrustBridge Rule of Law Foundation]
Arbitration is meant to be a private, expeditious alternative to court-based dispute resolution. Yet two provisions of India’s Arbitration and Conciliation Act, 1996 (“Arbitration Act”) — section 11 (appointment of arbitrators) and section 29A (extension of time to make an award) create predictable points of court exposure before award-stage review. The overwhelming majority of these interactions, as the data show, produce no contested determination at all.
This piece argues that when routine procedural tasks such as constituting a tribunal and resetting a deadline are routed through adversarial court proceedings, litigation becomes a built-in stage of arbitration rather than an exception to it.
The Appointment Bottleneck: Section 11
Section 11 is triggered when parties’ agreed appointment procedure mechanism fails and they cannot agree on who should arbitrate their dispute. In principle, this is a limited procedural problem. Someone must constitute the tribunal so that the arbitration can begin. In practice, section 11 became an entry point for court contestation before the arbitral process had even started.
In SBP & Co. v. Patel Engineering Ltd. (2005), the Supreme Court held that the section 11 power was judicial in character, allowing courts to examine, at the appointment stage, whether a valid arbitration agreement existed and whether the dispute was arbitrable. The appointment gateway became a site of threshold adjudication. Parties could now contest the reference before an arbitrator was even named.
Parliament tried to correct this in 2015 by inserting section 11(6A), which limits the court’s inquiry to whether an arbitration agreement exists. Parliament went further in 2019 by amending Section 11 to vest the appointment power in designated arbitral institutions altogether, effectively legislating the administrative model proposed here. That amendment, however, has not yet been brought into force, leaving the 2015 position operative and the structural problem unresolved.
The Supreme Court later clarified in Vidya Drolia v. Durga Trading Corporation (2021) that referral could be refused only where the arbitration agreement is obviously absent or the dispute manifestly non-arbitrable. The Section 11 court is not meant to conduct a mini-trial. That said, Vidya Drolia also permitted courts, for legitimate reasons, to conduct an ‘intense yet summary prima facie review’. This formulation invited inconsistent application.
In In Re: Interplay between Arbitration Agreements under the Arbitration and Conciliation Act, 1996 and the Indian Stamp Act, 1899 (2023), the seven-judge bench (convened primarily to resolve the enforceability of unstamped arbitration agreements) confirmed as a consequential finding that the section 11 court is concerned with the prima facieexistence of an arbitration agreement, leaving more granular questions of validity, enforceability, admissibility, and jurisdiction to the arbitral tribunal under section 16. A three-judge bench in SBI General Insurance Co. Ltd. v. Krish Spinning (2024) subsequently reasserted the existence-only test and eclipsed the broader Vidya Drolia observations, restoring a narrower conception of the referral court’s role.
Section 11 performs two functions: an administrative one (appointing an arbitrator after a deadlock) and a provisional screening one (filtering out plainly unmeritorious references). The screening is deliberately narrow and preliminary. The tribunal retains authority under section 16 to decide its own jurisdiction, and courts can review the award at the end.
Yet both functions (the routine administrative one and the limited supervisory one) are processed through the same formal adversarial hearing before a constitutional court. The procedure does not distinguish between a case where parties simply forgot to agree on a name and one where someone is genuinely contesting the existence of an arbitration agreement. Design choices made in 2005 continue to shape the system.
The Deadline Treadmill: section 29A
Section 29A is meant to impose time discipline on arbitral proceedings. The Arbitration Act sets statutory deadlines for making an award. Once the deadline passes, parties must approach the court for additional time.
The provision gives courts real power: to examine who caused the delay, to reduce arbitrator fees, to replace the tribunal, and to award costs. On paper, this is a meaningful accountability mechanism.
There is also a practical defence of section 29A that should be taken seriously. The statutory deadline may discipline proceedings even when courts rarely impose sanctions. Tribunals can rely on section 29A to resist repeated requests for filing extensions. Parties know that delay may eventually require a court application. In that sense, the provision’s effect may be ex ante: it may change behaviour before the matter reaches court.
But it rarely works that way. Our study finds, of 202 reported orders on arbitration time-extension applications from the Delhi High Court between 2015 and 2024, 198 (that is, 98%) ended in an extension being granted. Fees were never reduced. Arbitrators were replaced in just four cases. Costs were awarded in six. The dataset captures reported court orders, rather than the entire universe of arbitrations governed by Section 29A.
Most extension applications are uncontested. Both sides want more time, or no one is seriously alleging that the arbitrator acted improperly. In those cases, the Court has little basis to impose sanctions and no strong reason to refuse an extension. So it grants one quickly and moves on. But because the statute routes all extensions, contested and uncontested alike, through formal judicial proceedings, the court’s docket fills with procedural housekeeping.
The adjudicative power built into section 29A is real and occasionally useful. The question is whether it needs to be triggered every time a deadline lapses, including in the overwhelming majority of cases where no fault is alleged and no sanction is sought. If the main value of section 29A lies in the existence of a statutory deadline, then we might ask whether the same discipline can be preserved without routing every routine extension through court. A short consent extension in a complex arbitration, where no party alleges prejudice and no one seeks sanctions against the tribunal, is not the same as a contested application alleging tribunal inaction or tactical delay. Yet the statute sends both to the same forum.
Administration Versus Adjudication: The Missing Distinction
Reform debates on Indian arbitration tend to focus on the ad hoc versus institutional divide. Should parties default to institutional arbitration to reduce court intervention?
The recurring court exposure under sections 11 and 29A is a consequence of how certain procedural tasks are classified. Judicial proceedings are designed to resolve contested legal questions — to hear both sides, apply legal standards, and produce findings reviewable on appeal. Administrative mechanisms are designed to manage routine process tasks: schedule-setting, appointment, coordination.
When the latter are channelled through the former, the costs and formality of adjudication attach to tasks that do not require them. Appointments after a deadlock, deadline extensions by mutual consent, uncontested timeline resets — these are process-management tasks. Routing them through adversarial court hearings before constitutional courts is a category error, and an expensive one.
What Reform Could Look Like
The reform question is which procedural steps genuinely require judicial authority and which could be handled through structured administrative alternatives.
On appointments, the administrative approach is well-established globally. Designated neutral bodies or rotating appointment authorities handle deadlocks automatically, without adversarial proceedings. Courts remain available to review outcomes on grounds of legality, neutrality, or fairness, but are not the default first port of call.
On timelines, where parties mutually agree that more time is required, the tribunal’s mandate could continue upon an administrative filing recording limited procedural-status information: constitution of the tribunal, completion of pleadings, expiry of the statutory period, length of extension, and expected date of award. For institutional arbitrations, the filing would be made to the administering institution; for ad hoc arbitrations, it could be routed to the appointing authority already designated under the arbitration agreement or appointed by the court under section 11. This sunlight clause would not permit the authority to review the tribunal’s conduct or the merits of the dispute. It would only make timeline extensions visible and auditable. At the section 34 stage, when a court hears a challenge to the arbitral award, it should exercise its power to impose costs under section 31A purposively, treating significant, unexplained delay as relevant conduct, rather than treating costs as a formality.
If any party alleges prejudice, tribunal-caused delay, conflict, or misconduct, the matter would move to court. Substitution remains available under the existing provision.
Conclusion
India’s experience under sections 11 and 29A shows that safeguards can inadvertently produce the very congestion they are meant to prevent.
Courts remain essential when serious legal questions arise: when there are challenges to an arbitrator’s independence, when an award conflicts with public policy, when jurisdictional questions are genuinely contested. The point here is that procedure should match function. Administrative tasks should be handled administratively. Adjudication should be reserved for genuine disputes.
Getting that allocation right is the difference between arbitration that occasionally touches courts and arbitration that requires them.
– Prashant Narang & Renuka Sane