A two-member bench split. A third member produced a third answer. The first two then held that no majority existed at all, and the matter went to a fresh bench of five who had never heard it. One plan, three judgments, five judges. It is an irresistible headline, and it invites an irresistible conclusion: the institution is wrong for the job, so build a new one.

Resist it. Almost nothing that went wrong in the Chandra matter was caused by the shape of the tribunal, and almost nothing that a new tribunal would fix is what actually went wrong.

Two failures, neither architectural

Strip the case down and there are two distinct breakdowns. The first is procedural. Section 419(5) of the Companies Act tells the President to refer a difference of opinion to "one or more of the other Members", and says the point is then decided by the majority of members who have heard the case, "including those who first heard it". The provision contemplates addition to the original bench. It says nothing about what happens when the referee agrees with neither side, and nothing about who decides whether a majority exists. That is a gap in a 2013 statute, and it would have opened in a company petition, a merger, or an oppression case just as readily. It has nothing to do with insolvency.

The second is substantive, and it is the one that matters. Five entities holding 61.78 per cent of the vote were alleged to be the debtor’s associates. Section 109(4)(b) of the Code disqualifies an associate from voting. Section 79(2)(g) defines the term to catch "a company" in which the debtor, with his associates, owns more than half the share capital or controls board appointments. Three of the five disputed entities are limited liability partnerships. On the reported break-up they carry roughly 55.6 of those 61.78 points.

That is not an adjudication problem. That is a definition that does not cover the vehicle everybody uses. A dedicated insolvency authority staffed by dedicated insolvency judges reading the same clause would reach the same place, because the clause says what it says. The remedy is a line of drafting: import the wider "related party" test the Code already uses in Section 5(24), extend it to LLPs, trusts and sub-fifty-per-cent vehicles, and add an "accustomed to act on directions" limb. That is a monsoon-session amendment, not a new institution.

Parliament already built the dedicated authority

The argument for a bespoke forum rests on the claim that the Code never created an adjudicating authority of its own and simply borrowed the NCLT from company law. This is not quite right, and the inaccuracy is load-bearing.

Section 79(1) of the Code designates the Debt Recovery Tribunal as the adjudicating authority for Part III. Parliament did choose a separate forum for individual insolvency. The reason Chandra and Anil Ambani are before the NCLT is Section 60(1), which routes guarantors of corporate debtors to the same tribunal hearing the corporate case. That was a deliberate co-ordination choice, defended at the time and vindicated in practice: the guarantee and the underlying corporate default are one economic event and benefit from one pair of eyes.

Create a third authority for guarantors and personal insolvency fragments into three: the DRT for individuals, a new authority for guarantors, the NCLT for companies, with jurisdictional skirmishing at every seam. The proposal solves an institutional untidiness by adding one.

Single-member benches already exist

The reform is said to require a body that can sit as a single member, because bankruptcy elsewhere is decided by one judge. But Section 419(3) of the Companies Act already permits single-member benches for prescribed classes of case. The Central Government and the President can constitute them by rule. Indeed the third-member reference in Chandra was itself heard by a single member sitting alone. If a single-member track for personal-guarantee matters is desirable — and there is a case for it — that is an exercise in delegated legislation, achievable this quarter, not an exercise in institution-building achievable in this decade.

And single members may be the wrong answer anyway

There is a deeper problem with the single-member prescription. It sits awkwardly beside the concern that produced it.

If the worry is that an insider bloc can capture a creditors’ meeting and extract a 99.97 per cent discharge, then concentrating that decision in one adjudicator raises the stakes on every individual listing. It removes the second reader. It increases variance across benches and raises the return on forum selection. The Chandra split was undignified, but it is also the reason the associate question is now in front of the President of the NCLT and the appellate tribunal rather than buried in a routine approval order. It was the Technical Member who wanted to reject the plan. On the authors’ own account of what went wrong, the mixed bench worked.

The deference frame does not travel to Part III

The article’s underlying model is that creditors decide commercial questions and the adjudicating authority polices statutory boundaries. In Part II that has a rationale. A committee of creditors is deciding the fate of a going concern with enterprise value to preserve, on information the tribunal does not have and cannot cheaply acquire. Deference buys speed and it buys value.

Part III has no going concern. There is an estate, a set of claims, and a discharge. The question is not which of several business futures to back; it is how much a person must pay to be released from what he owes. Every developed system treats that as a question on which the court has something to say. An English individual voluntary arrangement can be challenged for unfair prejudice or material irregularity under Section 262 of the Insolvency Act 1986. A United States Chapter 13 plan must be proposed in good faith and must beat what creditors would receive in liquidation before a judge will confirm it. Neither system parks the outcome in creditor commercial wisdom.

So the corrective may run in the opposite direction from the one proposed. Part III does not need an authority designed to defer more cleanly. It needs a statutory ground on which a creditor can say the arrangement is unfairly prejudicial or the process was materially irregular, and a floor beneath dissenters comparable to Section 30(2)(b). At present Section 115 binds a bank that voted no to whatever the meeting produced, with no minimum entitlement whatsoever. That is the hole. It is not a hole a new courtroom fills.

Capacity is the binding constraint

There is also the unglamorous point. The chronic complaint about insolvency adjudication in India is delay, and the proximate cause of delay is vacancy and listing load. A new authority would need its own members, its own registry, its own rules and its own case-management architecture, drawn from the same thin pool of people who understand this law. Personal-guarantee matters are a small share of the docket. Building a second tribunal to serve them, while the first remains under-staffed, is a poor use of the only scarce input that matters.

The five-member bench may be evidence of illegality, not of design

One more caution. The Special Bench that stayed the approval order on 1 September consists entirely of members who had not heard the case. On a plain reading of Section 419(5), the deciding majority must include those who first heard it. The Supreme Court read a cognate provision purposively in Askari Hussain last month, but that case concerned what a validly constituted referee may decide, not whether the referee bench was properly composed.

If the appellate tribunal holds the Special Bench was irregularly constituted, the lesson of the episode is that the statute was not followed. That argues for procedural discipline and a clarifying amendment to Section 419. It does not argue for demolition.

The part nobody wants to say

Finally, the outcome that produced the outrage may not be the scandal it looks like. If the personal estate is genuinely worth less than the six and a half crore on offer, and the resolution professional’s valuation said so,  then rejecting the plan pushes everyone into bankruptcy under Chapter IV for less, with a discharge at the end regardless. A 99.97 per cent haircut is a grotesque number and may still be the value-maximising one.

Which relocates the failure. Twenty-two thousand crore of guarantees were taken from a man whose personal estate was always a rounding error against them. Lenders took those guarantees as a behavioural lever, not as a recovery asset, and priced nothing for the difference. The Code did not cause that. It merely made it visible, in a public order, with a percentage attached. Rebuilding the tribunal will not make the underlying arithmetic any less embarrassing.

What survives

One objection in the original piece survives all of this, and it is the strongest sentence in it. If insiders can vote, the commercial wisdom being deferred to is not the creditors’ wisdom at all. A corrupted franchise cannot be repaired by good faith further down the line, and no adjudicating authority, however designed, can hand back a decision that was never really taken.

That is right. Note only that it does not depend on the institutional proposal it has been recruited to support. Fix the electorate. Widen the associate definition and make it cover LLPs and trusts. Require the resolution professional to certify the affiliations of every voting creditor, on documents. Measure the seventy-five per cent against total admitted value rather than against whoever showed up. Give dissenters a floor. Give them a statutory ground of challenge.

Do those five things and the Chandra matter does not recur, whichever tribunal hears it. Build a new authority instead and it recurs on day one, in a nicer building.