AT-1 conundrum, revisited

In India, the administrator of Yes Bank had written off `8,415 crore of AT-1 bonds in March 2020. Both write-downs have since been found wanting in court. In October 2025, a Swiss court ruled that the Credit Suisse write-down lacked a valid legal basis, and that ruling is under appeal. On October 15, the Supreme Court upheld the Bombay High Court’s 2023 decision quashing the Yes Bank write-down. The draft reconstruction scheme for Yes Bank included a clause writing down the bonds.

After investors objected, the central government dropped that clause from the final scheme notified on March 13, 2020. The next day, the administrator wrote the bonds off anyway. The court held that the omission was deliberate and could not be undone by the administrator. Any write-down had to happen before the bank was reconstituted, not after. And under the bond terms, the power to act at the point of non-viability lay with the RBI, not with an administrator standing in for the board.

That answers who and when. The harder questions still remain. The missing law Yes Bank was rescued through a reconstruction scheme under Section 45 of the Banking Regulation Act. From moratorium to final notification took roughly eight days. That pace was set by a run on deposits at a bank whose core capital ratio had already fallen to 0.6%. The scheme’s terms were tailored to Yes Bank’s situation, not drawn from a general resolution law.

India still has no dedicated legal framework for resolving financial firms including failing banks. A proper resolution law would do what a scheme drafted in a crisis cannot. It would name a resolution authority and spell out its powers. It would set out, in law, the order in which shareholders and each class of creditors absorb losses. It would provide tools such as conversion of debt into equity.

And it would carry the safeguard found in international standards: no creditor should end up worse off than in liquidation. Investors would then know in advance what can happen to them. The Financial Resolution and Deposit Insurance Bill, which sought to create a resolution authority, was introduced in Parliament by the government in 2017 but withdrawn in 2018. The Yes bank case shows a crying need to have such an authority.

They state that a write-down of common equity “shall not be required before” a write-down of AT-1, both at the pre-specified trigger and at the point of non-viability. The actual order is left to each bank’s offer document. Yes Bank’s bond terms put AT-1 first, and provided that holders could not insist on equity being written off first. In 2023, we argued that writing down AT-1 bonds before equity is conceptually flawed.

The judgment shows that this is written into the rules themselves. Requiring AT-1 to be written down before common equity violates the basic principle of capital hierarchy and creates a perverse incentive for shareholders. It need not be this way. Yes Bank’s own 2013 bonds, issued before the Basel III rules took effect, expressly required common equity to be written down before the bonds could absorb losses.

After Credit Suisse, regulators in the EU, the UK, Hong Kong, and Singapore publicly stated that shareholders should absorb losses before AT-1 bondholders. The RBI should make this explicit too. Institutions misread the risk too It is bewildering that even the institutions treated these bonds as safer than they were. Until 2021, mutual funds valued AT-1 bonds as if they would be repaid on the first call date, usually five years after issue.

That made a perpetual, loss-absorbing instrument look like short-term paper, fit even for low-risk schemes. Valuing to call had been quietly lowering banks’ cost of capital, with investors carrying risk that the price did not show. In March 2021, Sebi stepped in and told funds to value perpetual bonds as 100-year instruments. It also capped how much a scheme could hold and confirmed that closed-ended debt schemes could not buy them.

Later, Sebi softened its stance with a glide path but held on to the principle. The retail gap The retail story remains troubling. Some 1,346 individuals invested about `679 crore in these bonds. The Supreme Court criticised the sale to individuals, but the mis-selling was not before it for decision. The human cost lies less in the amount than in the 277 investors who broke fixed deposits to buy them.

The RBI allowed banks to sell these bonds to individuals only with warnings that they were not fixed deposits and carried no deposit insurance. The banking regulator wrote the safeguards, the bank ignored them. Who was watching the sales counter? Suitability rules for complex instruments should apply, and be enforced, whoever sells them. In April 2021, Sebi found that Yes Bank had sold the bonds as “super FDs”.

It fined the bank `25 crore and three executives. In a separate order in September 2022, it fined former chief executive Rana Kapoor `2 crore. Those penalties were stayed on appeal. What now AT-1 bonds remain a sizeable market, with over `1 lakh crore outstanding. They serve a purpose; namely they let banks raise loss-absorbing capital without diluting shareholders, and they spare taxpayers the cost of rescues.

But they work only if investors can price the risk with confidence. That needs four things: A bank resolution law with a legislated order of losses; an RBI rule that shareholders absorb losses before AT-1 holders; suitability standards that follow the product, whoever sells it; and strict and timely enforcement against those found mis-selling. The Supreme Court has done its part. The rest is for lawmakers and regulators.