NARCL Bad Bank in India: IDRCL and the ₹30,600 Crore Guarantee

Every banking system that survives a bad loan cycle eventually asks whether to gather the wreckage in one place. Sweden did it in 1992, the United States after the savings and loan collapse, Malaysia with Danaharta after 1998. India asked the question for two decades, through the Narasimham Committee's asset reconstruction fund and the Economic Survey's proposals, and answered it in the Budget of 2021: a bad bank, built not as a state corporation by statute but as an asset reconstruction company under the existing Securitisation Act, owned mostly by the banks themselves and credible only because the Central Government guaranteed its paper.

1. Why aggregation, and why in 2021

• A large defaulted account is typically held by a dozen or more lenders under consortium arrangements; resolution stalls because every decision needs agreement among creditors with different provisioning, different exposures and different incentives. Aggregation puts the whole debt in one hand, so one owner negotiates, sues or sells.
• By 2021 the asset quality cycle had been through recognition (the asset quality review), provisioning and the insolvency code. What remained were old, fully or nearly fully provisioned accounts, written down in the books but still consuming management time and blocking resale. Selling them, even at a deep discount, frees capital and attention, and full provisioning means the sale price is mostly write-back, not fresh loss.
• Private asset reconstruction companies existed since 2003 but were too small for accounts of this size and unwilling to pay in cash. The design answer was a public aggregator with a sovereign guarantee standing behind deferred consideration.

2. The twin structure

Figure 1: Four levels of the bad bank

• The National Asset Reconstruction Company Limited (NARCL), incorporated in July 2021, is an asset reconstruction company registered with the Reserve Bank under the Securitisation Act, with public sector banks holding the majority of its equity (Canara Bank the sponsor) and private banks the rest.
• The India Debt Resolution Company Limited (IDRCL) is its service twin: majority private sector owned, engaged under an exclusive arrangement to manage and resolve what NARCL acquires, the split meant to marry a public balance sheet to private resolution skill.
• Acquisition follows the Securitisation Act's ordinary mechanics: NARCL offers for an identified account, the lead bank runs a Swiss challenge inviting better offers, and the debt passes to a trust whose receipts the selling banks hold.
• The consideration is the structure's signature: 15 per cent in cash, 85 per cent in security receipts, so the banks part with the asset but keep an upside claim on whatever the resolution realises.

3. The guarantee, and why the receipts needed it

• A security receipt is only as good as the eventual recovery, and banks had learned under the 2000s asset reconstruction model to distrust paper consideration. The Cabinet therefore approved a guarantee of up to ₹30,600 crore (September 2021) backing NARCL's receipts for five years.
• The guarantee pays the shortfall between the receipt's face value and the actual realisation when the account is resolved or the receipt redeemed; a guarantee fee, rising with time, pushes NARCL and IDRCL to resolve early rather than warehouse.
• For the selling bank the guarantee transforms the receipt: sovereign backing means the receipt does not sit as a doubtful asset, provisions write back on sale, and the capital released is immediate while the recovery is pursued at leisure.
• The five-year horizon from 2021 also sets the experiment's clock: acquisitions and resolutions have accelerated as the term's end approaches, and the guarantee's fate beyond it is a government decision outside the original design.

Figure 2: The numbers that define the experiment

4. Progress and the critique

Figure 3: The bad bank so far

• The identified universe was roughly ₹2 lakh crore of legacy exposures of ₹500 crore and above. Acquisition began slowly, then compounded: according to the Finance Ministry, by March 2026 NARCL had acquired 33 borrower accounts with aggregate debt of about ₹1.66 lakh crore, against a target of ₹2 lakh crore.
• Recoveries have begun: about ₹6,345 crore from 23 accounts by March 2026, more than 48 per cent of the acquisition cost of those accounts, with ₹4,364 crore of it in 2025-26 and three accounts fully resolved. Measured against face value, recoveries remain a small fraction, which is the arithmetic of accounts this old rather than a verdict on the vehicle.
• The critique, made since the Economic Survey debates: a bank-owned bad bank moves the loss around the system rather than out of it, since the sellers hold the receipts; the guarantee is a contingent fiscal cost; pricing between related parties is hard to keep honest; and aggregation can become warehousing, delay dressed as resolution.
• The defence: aggregation has in fact unlocked accounts the committee-of-creditors stalemate had frozen for years, fully-provisioned sales release capital and management time at once, and the guarantee fee plus the five-year term builds the incentive to resolve into the structure itself.

5. Quick Revision

1. The bad bank is an aggregator of creditor consent as much as of debt.
2. No new statute: licence, trust and guarantee, all existing law.
3. The guarantee converts paper consideration into saleable certainty.
4. Full provisioning is what makes selling at a discount painless.
5. Judge it by resolutions closed, not by assets acquired.

6. Frequently Asked Questions

How is NARCL different from the private asset reconstruction companies?

In law, barely at all, and that is the point of the design: it holds the same registration, uses the same trusts and receipts, and enforces with the same Securitisation Act powers. The differences are commercial. Its owners are the selling banks themselves, so the usual hard bargaining between seller and buyer happens inside one family, disciplined by the Swiss challenge. Its receipts carry a sovereign guarantee no private company's paper has, which is why banks will accept 85 per cent deferred from it. And its mandate is the large legacy account that private companies found too big, too litigated and too slow, leaving the newer, smaller flow of stressed loans to the private market. It is less a competitor than a clean-up crew for a specific historical stock.

Why do the selling banks hold the security receipts themselves?

Because the alternative was no sale at all. A cash sale at the true market price of a decade-old defaulted account would have crystallised prices banks found unacceptable, and no outside investor would pay more. Letting the banks keep 85 per cent as receipts defers the final reckoning: if resolution beats the price, the receipt pays the upside; if it falls short, the guarantee covers the gap on redemption. The honest criticism is that the system's exposure to the asset survives the sale, transformed from loan to receipt; the honest answer is that with full provisioning already done, what the sale releases, capital, provisions, management attention and a single resolution hand, is real even though the economics stay in the family until the receipt is extinguished.

Did the bad bank work?

On the narrow question, movement, yes: accounts that had sat in creditor deadlock for years have been aggregated past the consent problem, and about ₹1.66 lakh crore of face value have found a single owner with recoveries flowing back, figures that shift every quarter and are best checked against NARCL's latest disclosures. On the deeper question, whether India needed a standing bad bank or a one-time cleanup, the structure itself answers modestly: a five-year guarantee, a defined legacy universe and a rising fee are the anatomy of a scheme built to end, not an institution built to last. The durable lessons sit elsewhere, in recognition and provisioning done early, and in a resolution regime that would make the next cycle's wreckage smaller.

7. Related Topics and Provisions

• The Securitisation Act and asset reconstruction companies
• Non-performing assets: recognition and provisioning
• The Insolvency and Bankruptcy Code and the banker
• Committee landmarks: Narasimham to Nayak
• Resolution of financial firms and the FRDI debate
• Public sector banks: nationalisation and after