Guarantee or Co-Debtor? NCLAT on Arbitral Awards and Personal Guarantor Liability under IBC

[Shreya and Utkarsh are students at NALSAR University of Law.]

On 17 September 2026, the Principal Bench of the National Company Law Appellate Tribunal (NCLAT) dismissed four connected appeals in Ravindra Gopalkrishan Agarwal v. Janakalyan Sahakari Bank Limited, upholding the admission of four applications under Section 95 of the Insolvency and Bankruptcy Code 2016 (IBC) against personal guarantors of Oasis Alcohol Limited, the corporate debtor (CD). The credit facilities in question had been extended by an eleven-member consortium of co-operative banks. The CD had subsequently entered a consent arbitral award with the consortium in 2017, six of the eleven lenders had assigned their shares to an asset reconstruction company, and the Respondent bank (which had not assigned its share) sought to enforce a continuing personal guarantee nearly six years later. On its reasoning, the NCLAT offers a compact and useful restatement of two propositions of general application: first, that a consent award altering the time of repayment does not, without more, alter the character of the underlying obligation from guarantee to co-debtorship; second, that within a fragmented consortium, the burden of proving assignment lies squarely on the party asserting it. Each proposition merits closer examination.

Arguments and Ruling

The appellants’ central and most ambitious submission was that Clause 7 of the 2017 consent award rendering the borrower and guarantors “jointly and severally” liable to the lead bank had novated their status from guarantor to co-borrower. The distinction is consequential as a guarantor’s liability under Section 128 of the Indian Contract Act 1872 (ICA) is accessory and contingent, arising only upon the principal debtor’s default, whereas a co-debtor’s liability is primary and immediate from inception, unmediated by anyone else’s default. 

The Appellants had invoked Vyravan Chettiar v. Official Assignee of Madras and Lima Leitao and Co Limited v. Union of India to argue that parties who are equally and primarily liable for the same debt fall outside the definition of a contract of guarantee under Section 126 of the ICA, and further relied on a 2021 letter from the lead bank stating that the award had “finally and conclusively” determined all rights and liabilities, such that earlier documents had lost their significance.

The tribunal rejected the submission, holding that the award’s own terms, which expressly preserved liability “in the event of default” and provided for cancellation of any concession granted, were inconsistent with the claim that the original guarantee had been extinguished. 

Joint and several liability is a description of the relationship among obligors inter se; it says nothing, without more, about whether any particular obligor’s duty to the creditor is primary or accessory in character. A guarantor may perfectly well be rendered jointly and severally liable alongside the principal debtor and yet remain, in substance, a guarantor, provided the obligation continues to be conditioned upon the principal’s default and provided no fresh, independent promise has been substituted for the old one. The appellants’ arguments thus establish only that a co-obligor is not, by definition, a guarantor; they do not establish that language of joint and several liability, standing alone, converts a guarantor into a co-obligor. The tribunal might have made this distinction explicit rather than resting principally on the award’s savings clause for default.

In English law, per Moschi v. Lep Air Services Limited, the House of Lords held that a creditor’s acceptance of the principal debtor’s repudiatory breach — an event that would ordinarily extinguish a primary contractual obligation entirely — left the guarantor’s secondary “see to it” obligation intact, because Lord Diplock treated that obligation as analytically independent of the continued existence of the principal contract. 

The judgment in Ravindra Gopalkrishan Agarwal stands for the mirror proposition: restructuring the form or timetable of the principal debt, however emphatically worded, does not discharge or recharacterise a guarantee absent an express act of release, novation, or substitution appearing on the face of the instrument. Clause 16 of the underlying guarantee agreements, providing that the guarantee was irrevocable and would survive “variations, alterations, modifications, waiver, dispensation with or release of security” without requiring fresh consent, reinforced this outcome under Section 129 of the ICA, and the Tribunal was right to treat it as decisive in the absence of any document effecting an actual release.

Assignment, Locus Standi, and Section 133: The Burden of Proof

The second, and evidentially more interesting, contest concerned locus standi. Of the eleven consortium banks, six had assigned their shares to CFM Asset Reconstruction Private Limited (CFM ARC); the respondent had not. The appellants sought to impute the assignment to the Respondent regardless, relying on CFM ARC’s subsequent sale of the appellants' mortgaged property and on the respondent’s failure to produce any deed of assignment, and invited the tribunal to draw an adverse inference from that non-production. The tribunal declined, holding that an assignment agreement between another consortium member and CFM ARC cannot be attributed to the respondent without direct proof, and that the respondent could hardly be expected to produce a document to which, on its own uncontroverted case, it was never a party. 

The episode nonetheless exposes a genuine structural vulnerability in syndicated co-operative lending of this kind. Where a single mortgaged asset secures obligations owed collectively to multiple lenders operating under an inter-se agreement, and only some of those obligations are subsequently assigned to a resolution vehicle, a guarantor resisting enforcement is often poorly placed to disentangle which specific creditor retains standing against him, particularly where, as here, the realisation of the shared security by the assignee is capable of creating a superficial appearance of universal assignment. The tribunal disposed of the point cleanly by allocating the burden to the appellants, but a fuller engagement with the inter-se agreement of 21 June 2010, and with what collective enforcement of shared security by one assignee-in-part implies for the standing of the remaining, non-assigning members, would have made the finding more resilient to challenge on a fuller evidentiary record.

A closely related plea under Section 133 of the ICA that CFM ARC’s unilateral restructuring of the debt it held amounted to a material variation discharging the guarantee failed for the same reason: there was no proof that the Respondent, as distinct from CFM ARC or the assigning banks, had varied any term of its own contract with the Appellants. Because the variation was never shown to be attributable to the Respondent, the Tribunal did not need to confront the harder question, recently addressed by the Supreme Court in Bhagyalaxmi Co-operative Bank Limited v. Babaldas Amtharam Patel, of whether Section 133 discharges a surety wholesale or only “as to transactions subsequent to the variance” — the Supreme Court there rejecting an all-or-nothing reading in favour of a transactional, prospective discharge. Agarwal’s unstated premise, that a variation must first be shown to be attributable to the very creditor being sued before Section 133 can be engaged at all, is a necessary logical precondition to the inquiry, even though the Tribunal never reached the point of applying it.

Broader Significance and Conclusions

Agarwal is best read not in isolation but alongside Lalit Kumar Jain v. Union of India, where the Supreme Court held that approval of a resolution plan for a corporate debtor does not “ipso facto discharge” its personal guarantors, whose contract of guarantee survives the corporate insolvency process as an obligation independent of the principal debt. Together, the two judgments describe a remarkably consistent judicial posture across very different procedural contexts — one concerning the CD’s formal insolvency resolution, the other concerning an out-of-court consent restructuring: a personal guarantor’s exposure is treated as genuinely autonomous from whatever vicissitudes befall the principal obligation, whether restructuring, resolution, partial assignment, or partial recovery, and will yield only to a clear, affirmatively proved act of release, novation, or variation attributable to the specific creditor being sued.

For consortium lenders operating through co-operative banking statutes, the practical lesson of Agarwal is that locus standi survives the fragmentation of a syndicate so long as the suing creditor’s own share of the debt remains unassigned, irrespective of what other members of the consortium may separately have done with theirs. 

For guarantors and their counsel, the lesson is the converse as neither an imaginative recharacterisation of a consent award, nor an inference drawn from a co-lender’s unrelated dealings, nor a bare assertion of restructuring by a third-party assignee, will dislodge a continuing guarantee without direct documentary proof that the specific creditor in question executed the release, assignment, or variation relied upon. Of course, when consent awards are often used as a tool to wipe out old securities and create a clean slate for ARCs, by strictly enforcing the savings clause, the NCLAT may be making it harder for ARCs to restructure distressed assets out-of-court, as guarantors remain perpetually on the hook despite the commercial reality that the original bank relationship has been entirely replaced.

In a lending landscape increasingly populated by partial assignments to asset reconstruction companies and multi-lender consent arbitrations, that discipline is likely to prove the judgment’s most durable contribution.