A Bond Designed to Vanish at the Moment of Failure

Additional Tier 1 bonds, usually shortened to AT1 bonds, are perpetual securities issued by banks that count as regulatory capital because the contract behind each bond gives the issuing bank, and behind the bank its supervisor, the power to cancel interest payments, to convert the bond into common shares, or to write its principal down, in whole or in part, when the bank’s capital weakens past a stated point. In ordinary times the instrument behaves like debt. The bank pays a coupon, the holder collects it, and the principal sits on the balance sheet as a long term obligation that carries no fixed repayment date and no maturity at which the bank must hand the money back.

The design rests on a single bargain, stated in the prospectus and tested only in a crisis. AT1 bonds work only when the loss order promised in the contract is the loss order applied in the rescue, because the instrument is priced, sold and regulated on the assumption that equity absorbs losses before hybrid capital and that a write-down follows a stated trigger and a lawful process. Every other feature of the instrument, from its coupon to its place in the queue of creditors, is priced against that assumption, and a rescue that rearranges the order after losses arrive changes the value of every layer of bank capital at once.

In the queue of claims against a bank, an AT1 bond stands near the back. Depositors rank ahead of it, holders of senior unsecured debt rank ahead of it, and holders of Tier 2 subordinated debt rank ahead of it. Only common shareholders stand behind it, and common equity is the layer that the whole structure assumes will absorb losses first, because shareholders own the residual value of the bank and take the gains when lending goes well. The AT1 holder accepts the rear position in exchange for a coupon higher than the coupon on any safer layer of the same bank’s funding, and the size of that extra coupon is the market price of standing close to the front of the loss line while standing near the back of the payment line.

Loss absorption reaches the bond through two doors. The first door is mechanical: the contract names a level of the bank’s common equity Tier 1 ratio, and if the reported ratio falls to that level or below it, the principal is written down or converted by the terms the holder accepted at purchase. The second door is discretionary: the supervisor or resolution authority decides that the bank has reached the point of non-viability, the point at which failure follows without public support or a private rescue, and orders the loss absorption that the contract permits. The Basel III framework, published by the Basel Committee on Banking Supervision in 2010 and revised in 2017, requires both doors to exist before an instrument qualifies for the Additional Tier 1 category.

The record of rescues shows both doors in use and shows why the order of losses became a legal question rather than a bookkeeping entry. In June 2017 the Single Resolution Board resolved Banco Popular in Spain, writing down its AT1 and Tier 2 instruments after its equity had been extinguished, and the bank was sold for one euro. In March 2020 the Reserve Bank of India reconstruction of Yes Bank wrote down 84 billion rupees of AT1 bonds, and bondholders challenged the procedure in court for years afterwards. In March 2023 the Swiss supervisor FINMA ordered 16 billion Swiss francs of Credit Suisse AT1 bonds written down to zero inside a takeover in which shareholders received consideration, and that sequence, bondholders extinguished while shareholders received value, produced litigation in several courts. Each case turned on the same underlying question of whether the applied order matched the promised order.

The Problem AT1 Was Built to Solve

The financial crisis that began in 2007 exposed a gap between the capital banks reported and the losses they could actually absorb. Banks had funded long term and illiquid assets with short term borrowing, held thin layers of common equity against those assets, and relied on hybrid instruments whose principal remained a fixed obligation even as losses mounted. When property prices fell and interbank lending seized, the losses landed on equity layers too small to carry them, and governments faced a choice between letting large banks collapse into the payment system or rescuing them with public money on a scale that rewrote national budgets.

The failure of Lehman Brothers in September 2008 showed what an uncontrolled collapse did to money market funds, trade credit and banks far removed from the original losses. The United States acted through the financial crisis legislation that authorized the 2008 rescues, which authorized 700 billion dollars under the Troubled Asset Relief Program in October 2008, the United Kingdom took majority stakes in Royal Bank of Scotland and Lloyds, and governments across Europe guaranteed bank debt and injected capital into institutions whose private creditors emerged largely whole. The rescues stabilized the system, and they also taught creditors a lesson that regulators considered dangerous: lending to a large bank carried an implicit public guarantee, because the state had shown it would pay rather than let the loss order run.

Legislation after the crisis attacked parts of the problem without closing the gap in the capital stack itself. In the United States, the legislative answer included the Dodd-Frank Act of 2010, whose effects on bank structure, supervision and resolution planning are weighed in the Dodd-Frank banking impact assessment. Resolution powers mattered because an authority that can wind a bank down in an orderly way can impose losses, and the Financial Stability Board set out the expected powers in its Key Attributes of Effective Resolution Regimes in 2011. Powers alone, however, left a timing problem. A supervisor imposing losses in a crisis acts under pressure, in litigation risk and in markets that reprice every similar instrument within hours, and a loss imposed by improvised decision invites the court challenges that later rescues in fact produced.

The contractual answer was to move the decision to the date of issue. If the bond itself states, before any crisis, the trigger at which principal disappears, the authority that can order a write-down, and the place of the holder in the loss order, then a rescue applies a bargain rather than inventing one. Private creditors absorb losses in advance by contract, the taxpayer is moved behind them in the queue, and the supervisor’s room for discretion shrinks to declaring that a stated condition has been met. The Basel Committee on Banking Supervision built that answer into the Basel III framework, defining a category of capital, Additional Tier 1, whose entire purpose is to be available for loss absorption while the bank continues operating or at the edge of failure. AT1 is therefore best understood as a pre-signed loss allocation. The instrument exists so that the question of who pays for a bank failure is answered in a prospectus, in calm conditions, rather than in a finance ministry over a weekend.

What Makes an Instrument Count as Capital

Capital, in banking regulation, is not a pile of cash kept in a vault. It is the layer of funding that can absorb losses while the bank continues to meet its obligations to depositors and ordinary creditors, and its defining property is that nobody outside the bank can demand it back at the moment the bank can least afford to pay. A deposit fails that test completely, because the depositor may withdraw on demand. A senior bond fails it as well, because the holder can enforce repayment of principal on a fixed date against a bank that needs every resource to survive its losses.

The Basel III framework of the Basel Committee on Banking Supervision, published in 2010 and revised in 2017, converts that idea into a list of criteria an instrument must satisfy before a supervisor allows a bank to count it as Additional Tier 1 capital. Four properties do most of the work. The instrument must be permanent, with no maturity date and no term that obliges or pressures the issuer to redeem it. It must be subordinated, ranking behind depositors, general creditors and subordinated debt holders in a winding up. It must absorb losses, through conversion into common shares or write-down of principal, at a pre-specified trigger and at the point of non-viability. And its distributions must be fully discretionary, so that the bank can cancel a coupon at any time, for any period, without the cancellation counting as a default or giving the holder a claim to the missed payment later.

What must an instrument do to count as Additional Tier 1 capital?

An instrument counts as Additional Tier 1 capital when it is perpetual, ranks behind depositors and ordinary creditors, absorbs losses through cancellation of payments or reduction of principal, leaves every distribution to the issuer’s discretion, and carries no term that pressures the issuer to redeem it.

Each criterion closes a loophole that pre-crisis hybrid instruments had exploited. Permanence closes the refinancing loophole: an instrument with a maturity date is capital only until the date approaches, and banks before 2008 issued hybrids with step-up coupons that made redemption economically compulsory, so the supposed capital was scheduled to leave precisely as the bank aged into its risks. Subordination closes the queue loophole, because loss absorption is meaningless if the holder can sue for a senior claim in insolvency. Discretion over payments closes the drain loophole, since a coupon that must be paid in cash every year behaves like an expense of a going concern rather than a buffer held against a bad one.

The boundary between equity and debt runs through these criteria, and AT1 sits deliberately on the boundary. Common equity is the purest form of capital because it has no coupon, no maturity, no repayment right and no claim beyond the residue after every creditor is paid, which is why Basel III treats common equity Tier 1 as the highest quality layer. Ordinary debt is the opposite pole: fixed payments, fixed maturity, a right to trigger insolvency on a missed payment, and no obligation to absorb losses beyond the general risk of default. AT1 borrows the legal form of debt, a bond with a stated coupon and a denomination that pension funds and insurers can hold inside fixed income mandates, and bolts onto that form the economic substance of equity, permanence, discretion and contractual loss absorption. Regulators accepted the hybrid because banks could sell it to a far larger pool of buyers than common shares alone could reach.

The Capital Stack and the Order in Which Losses Fall

A bank’s balance sheet is a queue, and the capital stack is the part of the queue where the order of payment is also the order of sacrifice. On the funding side of the balance sheet sit, from safest to riskiest, insured deposits, senior unsecured debt, Tier 2 capital, Additional Tier 1 capital and common equity Tier 1. Payment priority runs in that order: in a winding up, each layer is paid in full before the layer behind it receives anything. Loss absorption runs in exactly the reverse order. Common equity absorbs losses first, continuously and automatically, because every loss reduces the bank’s retained earnings and therefore the value of the shareholders’ residual claim. Only after equity is exhausted, or after a trigger or resolution decision reaches further up the queue, do AT1, Tier 2 and senior creditors take their turn, and insured depositors stand protected behind the deposit insurance fund rather than behind the bank’s own resources.

The table below settles the comparison that the rest of the stack discussion depends on: for each layer, the form the funding takes, the mechanism by which it absorbs losses, and the stage of a bank’s decline at which that mechanism operates.

Capital layer	Form it takes	How it absorbs losses	When it absorbs them
Insured deposits	Customer deposit accounts covered by a deposit insurance scheme	The insurance fund pays depositors and takes their claim against the bank	After the bank has failed and the insurer is called
Senior unsecured debt	Bonds and wholesale borrowings ranking equally with general creditors	Holders bear losses in insolvency or under a resolution bail-in power	After capital instruments are exhausted or a resolution authority imposes bail-in
Tier 2 capital	Dated subordinated bonds with an original maturity of at least five years	Principal is written down or converted at the point of non-viability or in insolvency	At the point of non-viability, after Additional Tier 1 has absorbed losses
Additional Tier 1 capital	Perpetual bonds with fully discretionary non-cumulative coupons	Coupons are canceled, and principal is written down or converted into common shares	When the contractual capital trigger is breached or the point of non-viability is declared
Common equity Tier 1	Ordinary shares, retained earnings and disclosed reserves	Every loss reduces retained earnings and the value of the residual claim	Continuously, from the first loss the bank records

Two features of the stack explain most disputes about it. The first is that the order is a property of law and contract together, not of economics alone. Equity absorbs first because company law makes the share a residual claim, and AT1 absorbs before Tier 2 because the Basel III criteria and the issuing terms place it there, which means the order can be examined, priced and litigated as a set of documents rather than as a set of expectations. The second feature is that the layers differ in the manner of their absorption as much as in its timing. Equity absorbs losses silently, in accounting entries, with no decision and no announcement beyond the published accounts. AT1 absorbs losses by an act: a trigger calculation, a supervisory determination, a notice to holders. An act has an author, a date and a legal basis, and each of those can be challenged, which is why a mechanism designed to remove discretion from rescues still produces arguments about whether the act was properly performed.

A Perpetual Bond and the Call Date Convention

An AT1 bond has no maturity date. The prospectus promises repayment of principal on no date at all, and the instrument is designed to sit on the bank’s balance sheet permanently, absorbing losses for as long as the bank exists. Permanence is not an accident of drafting but a condition of the capital category itself: under the Basel III criteria of the Basel Committee on Banking Supervision, an instrument with a maturity date cannot qualify as Additional Tier 1, because funding that the bank must repay on a schedule is funding that leaves, and capital that leaves on a schedule cannot be relied upon at the point of failure. The perpetual form makes AT1 the only widely issued bond whose holder has no contractual right ever to be repaid by the passage of time.

What the holder receives instead of a maturity date is a call option held by the issuer. After a minimum period, five years from issue under the Basel criteria, the bank may redeem the bond at par on a stated call date, and on further dates thereafter, provided the supervisor consents and the bank either replaces the instrument with capital of the same or better quality or demonstrates that its capital position remains strong after redemption. The option belongs entirely to the bank. The holder cannot demand redemption, cannot put the bond back to the issuer, and has no remedy if the call date passes unused beyond holding an instrument whose next possible exit sits years further away. Basel III also forbids terms that create an incentive to redeem, such as a coupon that steps up sharply after the first call date, because an incentive to redeem would convert a permanent instrument into a dated one in economic substance.

Extension risk is the name for what happens when the convention fails. A bank approaching a call date in weakened conditions faces a supervisor reluctant to approve redemption, replacement capital priced at coupons far above the coupon on the outstanding bond, or a capital position in which redemption itself would breach a ratio. In those conditions the rational issuer extends, the bond is repriced in the market as a perpetual rather than a dated instrument, and its price falls toward the value of an endless stream of discretionary coupons from a weakened bank. The holder discovers that the five year horizon in the portfolio model was a courtesy, not a term. Episodes in which European banks allowed call dates on subordinated and AT1 instruments to pass without redemption, Deutsche Bank’s decision in 2008 to leave a subordinated bond outstanding past its call date being an early and much discussed example in the wider hybrid market, repriced the convention across the sector each time.

The gap between convention and contract is a permanent feature of the instrument rather than a defect that better drafting could remove. Remove the issuer’s option and the instrument gains a maturity and loses its capital status; guarantee the call and the same result follows. The Basel framework chose permanence and placed the option with the issuer, and the market chose to price a hope. A careful reader of an AT1 prospectus therefore treats the first call date as a statement about the bank’s funding intentions under favorable conditions, and treats the perpetual term as the statement that governs under unfavorable ones, which are the conditions under which the rest of the contract, the trigger, the coupon discretion and the write-down, also begins to operate.

Coupons That Can Be Canceled Without a Default

The coupon on an AT1 bond is a payment the bank may choose not to make, at any time, for any reason or for no stated reason, without breaking the contract. Full discretion over distributions is one of the Basel III criteria for Additional Tier 1 capital, and it is written into the terms in the strongest available form: cancellation is not an event of default, it gives the holder no right to accelerate repayment, no right to petition for insolvency, and no claim against the bank in a later winding up beyond the subordinated claim on principal that the holder always held. A missed AT1 coupon is legally comparable to a passed dividend on a preference share rather than to a missed interest payment on a bond, even though the instrument is sold, taxed in many jurisdictions and reported as a bond.

The canceled coupon is also gone for good. AT1 distributions are non-cumulative, which separates them from the cumulative preference shares of an earlier era, under which skipped dividends accumulated as arrears that had to be cleared before shareholders received anything. Under a non-cumulative term, a coupon canceled in a bad year is not deferred, not accrued and not compensated by a larger payment in a good year. The bank may resume payments when its board chooses, at the original rate, as if the canceled payment had never been scheduled. For the holder, each coupon is therefore an independent act of issuer discretion repeated across the life of a perpetual instrument, and the yield on the bond is compensation for a stream of payments every one of which the payer may lawfully withhold.

Discretion is wide but not unlimited, and the limits come from the bank’s own accounts rather than from the bond contract. Coupons may be paid only from distributable items, the accumulated profits and reserves that company law in the relevant jurisdiction allows the bank to distribute, so a bank whose losses have consumed its distributable reserves cannot pay an AT1 coupon even if its board wishes to pay one. In the European Union, the Capital Requirements Regulation ties distributions to the maximum distributable amount, a formula that restricts coupons on AT1 instruments, dividends and variable remuneration together when the bank’s capital falls into its combined buffer range, with the restriction tightening as capital declines toward the minimum. The mechanism converts coupon cancellation from a purely voluntary act into a semi-automatic consequence of a falling ratio, arriving before any trigger on principal and serving as an early, visible stage of loss absorption that costs the bank reputation rather than cash.

The reason regulators insist on the discretion is conservation. A bank sliding toward trouble destroys its own recovery prospects if it continues to pay out cash to hybrid holders while its equity erodes, and a contractual obligation to pay would force supervisors to choose between enforcing the coupon and preserving the bank. Cancelling coupons keeps cash inside the institution, slows the decline of the common equity Tier 1 ratio, and signals to every other creditor that the loss absorption machinery described in the prospectus is functioning in its mildest gear. The holder, in turn, prices the discretion at issue: AT1 coupons stand well above the coupons on the same bank’s Tier 2 and senior debt, and part of that spread is the market value of payments that can vanish without legal consequence for the payer.

The Common Equity Tier 1 Ratio, Worked Through

Every trigger, buffer and distribution restriction in the AT1 contract is expressed through one number, the common equity Tier 1 ratio, and the instrument cannot be understood until the ratio can be calculated by hand. The numerator is common equity Tier 1 capital: ordinary share capital, retained earnings and disclosed reserves, reduced by regulatory deductions such as goodwill, other intangible assets and certain deferred tax assets that supervisors judge unable to absorb losses in a crisis. The denominator is risk-weighted assets, the bank’s loans, securities and other exposures each multiplied by a percentage weight reflecting the risk of loss, with off-balance-sheet commitments converted into exposure amounts before weighting. The ratio divides the highest quality capital by the risk-adjusted size of the balance sheet, and Basel III requires the result to stand at or above 4.5 percent before buffers.

Why do risk-weighted assets matter more than total assets for bank capital?

Risk-weighted assets matter more because capital rules measure strength against the assets most likely to produce losses. A loan secured on sound property carries a low weight, a speculative loan carries a high one, and dividing capital by the weighted total shows how much loss the bank can absorb relative to its risk.

A hypothetical example fixes the arithmetic. Consider a hypothetical bank holding 40 billion in common equity Tier 1 capital against 320 billion in risk-weighted assets, figures chosen for clean division and stated in a single currency unit. The ratio is 40 divided by 320, which is 0.125, expressed as 12.5 percent. That figure stands comfortably above the Basel III minimum of 4.5 percent and above the level at which most contractual AT1 triggers are set, and it means the bank could absorb losses equal to 8 percent of its risk-weighted assets, 25.6 billion, before its ratio touched the 4.5 percent floor. Notice what the denominator does to comparisons between banks: a second hypothetical bank with the same 40 billion of capital and a portfolio of safer assets weighted at 200 billion reports a ratio of 20 percent, while a riskier bank whose 320 billion of total loans weight to 400 billion reports 10 percent, although all three hold identical capital and the second and third may hold identical total assets.

Losses attack the ratio through the numerator. A loss is charged against profits and therefore against retained earnings, which sit inside common equity Tier 1, so each billion of loss removes a billion of numerator while the denominator changes only as the damaged assets are written down, sold or reweighted. If the hypothetical bank records losses of 16 billion on soured loans, its capital falls from 40 billion to 24 billion, and against unchanged risk-weighted assets of 320 billion the ratio falls from 12.5 percent to 7.5 percent. A further 8 billion of losses brings capital to 16 billion and the ratio to 5 percent, below the 5.125 percent trigger floor discussed in the next section and inside the zone where an AT1 contract set at that trigger would fire. The slide is arithmetic rather than opinion, which is precisely why regulators built the trigger on the ratio: the parties can dispute valuations, growth prospects and management quality, and the ratio narrows the dispute to the two audited quantities that compose it.

The Trigger: Falling Capital and the Point of Non-Viability

An AT1 bond carries two distinct firing mechanisms, and confusing them is the source of much misunderstanding about the instrument. The first is the contractual trigger, a stated level of the common equity Tier 1 ratio written into the terms at issue. The second is the point of non-viability, a determination made by the supervisor that the bank will fail unless the instruments absorb losses or public support arrives. The contractual trigger operates on a number and can in principle fire without any official decision beyond the publication of the bank’s accounts. The point of non-viability operates on a judgment and can fire while reported ratios still stand above every contractual line, because a supervisor watching deposit flight and funding markets close may conclude that the audited ratio no longer describes a viable institution.

The Basel III framework, issued by the Basel Committee on Banking Supervision in 2010 and revised in 2017, fixes the floor for the contractual trigger: an instrument qualifies as Additional Tier 1 only if its terms provide for conversion or write-down when the common equity Tier 1 ratio falls to 5.125 percent or below, or to a higher level the issuer and supervisor choose. The figure is not arbitrary. It sits above the 4.5 percent minimum ratio, so the trigger fires while the bank still satisfies its bare minimum requirement, injecting loss absorption at the last point where the injection can restore the ratio rather than merely certify the failure. Jurisdictions transpose the floor into their own law: in the European Union the trigger requirement appears in the Capital Requirements Regulation governing instrument eligibility, and in India the Reserve Bank of India master directions on Basel III capital prescribe the trigger and the write-down mechanics for instruments issued by Indian banks.

Issuers choose where above the floor to set their own trigger, and the choice divides the market into two families. High-trigger instruments, commonly set at a ratio of 7 percent, fire early, while the bank is wounded but operating, and the loss absorption they deliver can genuinely recapitalize the institution as a going concern. Low-trigger instruments, set at the 5.125 percent floor, fire late, when the bank is close to the minimum and the absorption functions mainly to protect depositors and senior creditors in a gone-concern outcome. The pricing follows the design: a high trigger stands closer to ordinary fluctuations in the ratio, so the holder demands a higher coupon for the shorter distance to destruction. Supervisors have their own preferences in the argument, since an early trigger delivers private recapitalization before public authorities must act, while banks prefer the low trigger because it makes the bond cheaper to issue and less likely to fire during a passing downturn.

The point of non-viability clause exists because the contractual trigger can fail to fire in time. Accounting ratios are published quarterly, losses can be recognized slowly, and a bank can pass from reported health to a liquidity collapse between reporting dates. Basel III therefore requires every AT1 instrument to include terms under which the supervisor may order write-down or conversion at the point of non-viability, defined through the supervisor’s determination that the bank is unlikely to survive without such action or without extraordinary public support. The Financial Stability Board, in its Key Attributes of Effective Resolution Regimes of 2011, built the same determination into the international design for resolution powers, so the clause connects the individual contract to the authority that would conduct a rescue. In the European Union the corresponding judgment is made under the Bank Recovery and Resolution Directive by the resolution authority, and in India the determination rests with the Reserve Bank of India as the banking supervisor.

Write-Down and Conversion, the Two Loss Mechanisms

When a trigger fires or a supervisor declares the point of non-viability, the contract must do something to the bond, and the Basel III framework permits two mechanisms, write-down and conversion, which deliver the same accounting result by opposite routes. Both mechanisms increase the bank’s common equity Tier 1 capital or extinguish a liability that ranked ahead of it, and both therefore raise the ratio whose decline caused the firing. They differ in what the holder is left holding afterwards, in what happens to the existing shareholders, and in the legal questions each route leaves open after the rescue.

Write-down reduces the principal of the bond by the amount the terms specify, in the most severe form to zero, and the holder’s claim shrinks or disappears without any payment. The reduction is recorded as a gain in the bank’s accounts and added to the reserves inside common equity Tier 1, which is the mechanical sense in which destroyed debt becomes capital. A permanent write-down is final: the principal never returns, coupons on the written-down amount cease, and the holder’s only remaining route to recovery is litigation against the decision itself. A temporary write-down softens that finality by allowing a write-up, under strict conditions in the issuing terms and in European Union law, if the bank returns to profit; the principal may be restored in stages out of subsequent earnings, subject to limits tied to the bank’s distributable capacity and to the maximum distributable amount, so the holder holds a claim whose value depends on a recovery the write-down itself was meant to secure.

Conversion takes the other route: the bond is extinguished and the holder receives common shares in the bank at a conversion price or rate fixed in the terms, becoming a shareholder in the institution whose losses destroyed the bond. The bank’s debt falls and its share capital rises by the corresponding amount, and the ratio improves as it does under write-down. The costs land differently. Existing shareholders are diluted, because the new shares issued to converted holders reduce each old share’s fraction of the company, and the share price of a bank at its trigger point is usually depressed, so a conversion rate fixed at issue can deliver the holders a large block of shares for their canceled principal. Holders, for their part, receive an asset they may not want: an equity stake in a distressed bank, illiquid in size, volatile in value, and ranked behind every creditor they previously stood beside or ahead of.

The Basel Committee framework permits either mechanism and judges the instrument by whether loss absorption at the trigger is certain, timely and sufficient, leaving the choice of machinery to issuers and national law. The jurisdictions diverge in practice. Under the European Union Bank Recovery and Resolution Directive, resolution authorities hold both write-down and conversion powers over capital instruments, and the eligibility rules in the Capital Requirements Regulation accommodate bonds designed for either route; European issuance has used both, with conversion features prominent in several markets and write-down structures, including temporary write-down, used widely by issuers whose company law made share issuance at a trigger difficult. In India, the Reserve Bank of India framework prescribes write-down as the operative mechanism for Additional Tier 1 instruments, and the Yes Bank reconstruction of March 2020 applied it, extinguishing 84 billion rupees of bonds rather than converting the holders into shareholders of the reconstructed bank. In Switzerland, the Credit Suisse instruments resolved in March 2023 were write-down bonds, and FINMA ordered their full write-down of 16 billion Swiss francs under the emergency ordinance and the contractual terms.

How a Rescue Uses AT1, Step by Step Through One Balance Sheet

A single worked example, built on a hypothetical bank with stated figures, shows the machinery operating in sequence and shows why the order of each stage matters. Consider a hypothetical regional bank, invented for this illustration, holding common equity Tier 1 capital of 3.2 billion against risk-weighted assets of 40 billion, a ratio of 8 percent. Behind the equity stand 1.2 billion of Additional Tier 1 bonds with a contractual trigger set at the Basel III floor of 5.125 percent, 1 billion of Tier 2 bonds, and a funding base of senior debt and deposits that the example leaves untouched, because a rescue that reaches the depositors has failed at its purpose. The trigger level, translated into capital, is 5.125 percent of 40 billion in risk-weighted assets, which is 2.05 billion: the AT1 bonds fire when common equity Tier 1 falls to 2.05 billion or below.

The losses arrive first, as losses in a real rescue always arrive before the legal acts. Suppose the bank’s commercial property portfolio deteriorates and the bank recognizes losses of 1.6 billion across two reporting periods. Every unit of loss reduces retained earnings and therefore the numerator of the ratio, so common equity Tier 1 falls from 3.2 billion to 1.6 billion while risk-weighted assets, after write-downs of the damaged loans, are assumed for clarity of the arithmetic to remain at 40 billion. The ratio becomes 1.6 divided by 40, which is 4 percent. The bank has breached its 4.5 percent regulatory minimum, passed through the zone in which the European Union style maximum distributable amount would already have restricted its AT1 coupons and dividends, and crossed the contractual trigger of 5.125 percent on the way down. The equity holders have absorbed the whole 1.6 billion of losses in the accounting sense: the value of their residual claim has halved, exactly as the capital stack promises, and no creditor has yet lost anything.

The trigger then converts the accounting position into a legal act. Because the reported ratio of 4 percent stands below the 5.125 percent line written into the bonds, the write-down clause operates, and in this example the supervisor also confirms the point of non-viability, since a bank at 4 percent with deteriorating assets cannot raise private capital on terms that would restore it. The full 1.2 billion of AT1 principal is written down to zero. The accounting effect is immediate: extinguishing the liability records a gain of 1.2 billion that flows into the reserves within common equity Tier 1, raising the numerator from 1.6 billion to 2.8 billion. Against unchanged risk-weighted assets of 40 billion, the restored ratio is 2.8 divided by 40, which is 7 percent. The bank has moved from below its minimum to a ratio above the trigger level without a unit of public money entering the balance sheet, and the cost has been borne entirely by the AT1 holders, whose bonds have ceased to exist, standing behind shareholders who lost value on paper and ahead of Tier 2 holders, senior creditors and depositors who lost nothing.

Banco Popular in 2017: The Hierarchy Applied as Written

The resolution of Banco Popular in June 2017 remains the clearest illustration of additional Tier 1 capital performing the task described in its contract. Banco Popular was the sixth largest bank in Spain, with a large stock of nonperforming real estate loans left from the Spanish property collapse and a deposit base that came under severe pressure in the first week of June 2017. On 6 June 2017 the European Central Bank, acting in its supervisory capacity for the European Union, determined that the bank was failing or likely to fail because liquidity outflows had left it unable to pay its debts as they fell due. That determination opened the door to resolution under the European Union Bank Recovery and Resolution Directive, the framework adopted in 2014 to ensure that shareholders and subordinated creditors, rather than taxpayers, bore the first losses of a bank failure.

The Single Resolution Board, the European Union resolution authority seated in Brussels, adopted a resolution scheme in the early hours of 7 June 2017. The scheme followed the order written into European Union law and into the instruments themselves. Existing common shares were written down to zero first. Additional Tier 1 instruments were written down next. Tier 2 instruments were written down after them. Only after those layers had absorbed losses in full were the remaining assets and liabilities transferred. Banco Santander acquired the resolved bank for one euro, a nominal price that reflected the value left after the write-downs had eliminated the deficit between assets and liabilities. Santander later raised fresh common equity to support the acquisition, but that capital came from private investors in a market transaction, not from a government grant.

Depositors were untouched. Senior creditors were untouched. Covered deposits and eligible deposits above the coverage limit continued without interruption, branches opened on the following morning under Santander ownership, and payment services continued. No public money was used in the resolution itself. Banco Popular tested whether a different legal promise would hold under time pressure, overnight, with a bank that served more than four million customers.

Several features explain why practitioners in the European Union treat the case as a textbook application. The trigger was a formal failing or likely to fail determination by the supervisor, not an improvised political decision. The authority exercised was the authority named in the directive and in Spanish implementing law. The loss order matched the contractual hierarchy that investors had bought: equity bore the first loss, additional Tier 1 capital bore the second, and Tier 2 capital bore the third. The transfer tool used was a sale of business, a resolution tool listed in the directive, rather than an open-ended public guarantee. Valuation work commissioned by the Single Resolution Board supported the write-down amounts, and the European Commission endorsed the scheme for compatibility with European Union state aid rules on the same day.

Litigation followed, as litigation follows almost every large write-down, but the litigation concerned valuation, procedure, and compensation, not a reversal of the operating result. Former shareholders and bondholders challenged the resolution in European Union courts and in Spanish courts, arguing that the bank had retained value and that the valuation had been too severe. Those challenges did not restore the written-down instruments, did not unwind the sale to Santander, and did not impose the losses on depositors. The distinction is instructive for reading any rescue. A court can examine whether an authority followed the correct procedure and whether compensation was due, while the loss-absorbing work of the instruments remains complete. In Banco Popular, procedure, hierarchy, and outcome pointed in the same direction, and that alignment is the reason the episode is cited in training materials and in later resolution planning across the European Union.

Yes Bank in 2020: A Write-Down Inside a Reconstruction

The rescue of Yes Bank in India in March 2020 used a different legal instrument, a reconstruction scheme, and placed additional Tier 1 capital in a more contested position. Yes Bank had expanded lending rapidly, accumulated a large share of stressed corporate loans, and lost depositor confidence in late 2019 and early 2020. On 5 March 2020 the Reserve Bank of India, the central bank and banking supervisor in India, imposed a moratorium under the Banking Regulation Act, 1949, capped withdrawals, superseded the board of directors, and appointed an administrator. On 13 March 2020 the Government of India notified the Yes Bank Limited Reconstruction Scheme, 2020, under which a consortium led by State Bank of India subscribed for fresh common equity and other lenders joined as investors.

Inside that reconstruction, the administrator wrote down additional Tier 1 bonds with a face value of 84 billion rupees in full. The write-down was permanent. The bonds ceased to exist as claims. Equity holders were treated differently. Existing shareholders were diluted heavily by the issue of new shares to the incoming investors, but their shares were not erased. The scheme imposed a lock-in on 75 percent of the holdings of shareholders who held 100 shares or more, preventing sale of that portion for three years. Depositors, by contrast, were protected in full, the moratorium was lifted on 18 March 2020, and withdrawal limits ended. The practical protection of depositors in a rescue of this kind works alongside the formal guarantee provided by the Deposit Insurance and Credit Guarantee Corporation, which covers eligible deposits up to the statutory limit in India when a bank is liquidated.

The contrast between the treatment of additional Tier 1 holders and the treatment of shareholders generated the central dispute. Under the Basel III framework adopted in India through Reserve Bank of India master circulars, additional Tier 1 instruments contain a point of non-viability clause. The clause allows a write-down when the Reserve Bank of India determines that a bank would become non-viable without a write-down or a public sector capital injection. Holders of the Yes Bank bonds argued, in petitions filed in Indian courts, that a reconstruction that preserved part of the existing equity while eliminating the bonds inverted that ranking. The Reserve Bank of India and the administrator maintained that the scheme and the contractual terms permitted the write-down at the point of non-viability and that equity dilution, combined with the lock-in, imposed a real loss on shareholders even though their shares survived.

Understanding the episode requires separating three questions that public discussion often merged. The first question was economic: whether Yes Bank needed loss absorption of that scale to restore its common equity Tier 1 ratio and to reassure depositors. The second question was contractual: whether the bond terms, read with the Reserve Bank of India regulations in force at issuance, authorized a permanent write-down in a reconstruction rather than only in liquidation. The third question was procedural: whether the decision to write down the bonds had been taken by the correct decision maker, in the correct sequence, under the scheme as notified. The later court judgments addressed mainly the third question, and to a degree the second, rather than re-deciding the economics of the rescue.

For the design of additional Tier 1 capital, Yes Bank supplied a lesson distinct from the lesson of Banco Popular. A write-down can restore a capital ratio on paper and protect depositors in practice, yet still leave the legitimacy of the instrument damaged if investors conclude that the order of loss and the process for imposing loss were unclear at the moment of sale. That combination, a reconstruction scheme, a permanent write-down, surviving equity, and a partly retail holder base, explains why the Yes Bank episode continued in litigation long after the bank had resumed normal operations and raised additional capital in a public offering in July 2020.

Credit Suisse in 2023: When Shareholders Were Paid and AT1 Was Erased

The rescue of Credit Suisse in Switzerland in March 2023 produced the sharpest dispute over loss order in the history of additional Tier 1 capital. Credit Suisse had suffered years of losses, risk management failures, and deposit outflows, including large withdrawals in October 2022 and again in March 2023 after the failures of Silicon Valley Bank and Signature Bank in the United States unsettled confidence in banks with large uninsured deposits. On 19 March 2023 the Swiss Federal Council, the Swiss government, announced a takeover of Credit Suisse by UBS, supported by liquidity assistance from the Swiss National Bank and by a federal loss guarantee for a defined portfolio of Credit Suisse assets. The takeover terms gave Credit Suisse shareholders one UBS share for every 22.48 Credit Suisse shares, consideration with a value of about three billion Swiss francs at announcement.

On the same day, the Swiss Financial Market Supervisory Authority, known as FINMA, instructed Credit Suisse to write down its additional Tier 1 instruments in full. The amount erased was 16 billion Swiss francs, a figure larger than the consideration paid to shareholders. Holders of those instruments received nothing. The write-down rested, according to statements issued by FINMA on 19 March 2023 and 23 March 2023, on two foundations. The first foundation was contractual. The prospectuses for the Credit Suisse instruments contained a viability event clause linked to extraordinary government support: if the Swiss Confederation granted exceptional assistance to keep the bank viable, the instruments could be written down. FINMA stated that the liquidity assistance loans, backed by a federal default guarantee, constituted that form of support. The second foundation was legislative. The Swiss Federal Council had adopted an emergency ordinance on 19 March 2023 under constitutional emergency powers, amending the Swiss Banking Act to authorize FINMA to order a write-down in connection with such support.

The apparent inversion drew immediate criticism from investors, from other regulators, and from commentators in several jurisdictions. In a conventional liquidation hierarchy, common equity absorbs losses before additional Tier 1 capital. Shareholders receiving value while additional Tier 1 holders received zero appeared to reverse that order. FINMA and the Swiss Federal Department of Finance responded that Switzerland had not conducted a resolution under the Swiss too big to fail regime in the form used for Banco Popular, but had instead executed an emergency commercial takeover supported by the state, and that the contractual trigger in the Swiss instruments had been drafted precisely for government support outside a formal insolvency.

Litigation followed in multiple forums. Bondholders filed claims in Swiss courts challenging the emergency ordinance, the FINMA instruction, and the constitutional basis for altering creditor rights without a vote of the Swiss Parliament. Other claimants pursued proceedings in courts outside Switzerland, including claims in the United States concerning disclosure in offering documents. Those proceedings were stated by the claimants as challenges to legality, proportionality, and property rights. As a matter of record, the claims sought restoration of value or compensation; they did not, by filing alone, restore the instruments. No responsible summary can treat a filed claim as a decided outcome, and the Swiss cases raised questions, concerning the scope of emergency powers and the interaction between contract and ordinance, that Swiss courts had not answered before March 2023 in an additional Tier 1 context.

The Credit Suisse episode therefore tested a different part of the mechanism from the part tested in Spain. The contractual language existed, the government support existed, and the write-down followed within hours. The dispute was whether a trigger drafted for viability support should operate when shareholders retained value in a negotiated takeover, and whether emergency legislation adopted on the day of the transaction could supply authority that investors believed the ordinary statute had withheld.

What the Courts Decided About the Yes Bank Write-Down

Court review of the Yes Bank write-down in India turned on sequence and authority, not on a general prohibition against writing down additional Tier 1 capital. Bondholders, trustees, and individual investors challenged the administrator decision in the Bombay High Court, the high court with jurisdiction over Mumbai, where the bank and the regulator maintained principal offices. In a judgment delivered in January 2023, the Bombay High Court quashed the administrator decision to write down the bonds. The court examined the Yes Bank Limited Reconstruction Scheme, 2020, as notified by the Government of India on 13 March 2020, and the steps taken under it. The scheme in its final notified form did not itself contain the write-down in the manner that an earlier draft had proposed. The administrator had announced the write-down before the scheme took effect in its final form and had treated a draft provision as a foundation for a completed decision. The court held that the decision sequence was defective: a final decision on the bonds had to follow the notified scheme and rest on a proper consideration under that scheme, not precede it on the basis of a draft.

That holding was procedural in character, even though its practical effect for holders was substantial. The Bombay High Court did not hold that additional Tier 1 bonds in India can never be written down. The court did not hold that the point of non-viability clause in the bond terms was void. A reader who follows questions of financial recovery and institutional design may also consult a companion reading reference alongside the judgments themselves, because the reasoning rewards close attention to chronology. What the court decided was that the particular decision, taken in that order and on that documentary basis, could not stand.

The dispute reached the Supreme Court of India, the apex court in India. In a ruling delivered in October 2026, the Supreme Court of India held the write-off of the Yes Bank additional Tier 1 bonds illegal. Reporting of the ruling in October 2026 described the holding as a conclusion that the write-down had not been carried out in accordance with the governing scheme and the applicable regulatory framework in India. The Supreme Court ruling closed the appellate path for the write-down as executed in March 2020. It did not establish, and no summary should claim that it established, a rule that a future write-down in India, ordered after a lawful point of non-viability determination, under a scheme that expressly provides for it, and through a decision taken in the correct sequence, would also fail. Distinguishing a holding about a specific exercise of power from a holding about the existence of the power is the basic discipline of reading resolution litigation in any jurisdiction.

Neither court decided several questions that investors sometimes assume were decided. Neither court priced the bonds. Neither court determined what recovery holders would have obtained in a liquidation of Yes Bank in March 2020, a counterfactual that would have required a valuation of a distressed loan book under forced sale conditions.

For writers of future instruments, the Yes Bank judgments carry a drafting and governance message as much as a legal one. A trigger must be matched by a visible decision path: the authority named in the contract must be the authority that acts, the scheme or resolution document relied upon must be the final document, and the date of the decision must fall after the date of the power under which it is taken. When any link in that chain depends on a draft, on an implied power, or on a press release rather than a notified instrument, a court in India, applying ordinary principles of administrative law, can set the decision aside without questioning the need for loss absorption itself. That narrowness is not a weakness in the judgments.

Who Buys AT1 Bonds and How the Market Is Built

The buyer base for additional Tier 1 bonds was institutional from the start, and the structure of that base explains much about pricing, liquidity, and dispute. Asset managers buying for bond funds form the largest single group in the European market. Private banks buying for discretionary wealth clients in parts of Europe and Asia form a second group, historically important in Switzerland, Singapore, and Hong Kong. Insurers and pension funds form a third group, usually with internal limits on subordinated bank capital, on exposure to a single issuer, and on instruments without a fixed maturity.

Stefan Avdjiev, Mary Anne McGuire and Patrick McGuire documented the expansion of the market in research published by the Bank for International Settlements, the institution in Basel, Switzerland, that hosts the Basel Committee on Banking Supervision. Their work traced how issuance of contingent convertible and additional Tier 1 instruments grew after Basel III created a standardized category for them, how issuance concentrated in large banks in Europe and in Asia, and how the investor base broadened from a narrow set of specialists toward mainstream fixed income portfolios during years of low policy interest rates. Low yields on senior bank debt and on government bonds pushed portfolio managers toward higher coupons, and additional Tier 1 capital, with seniority above equity and a coupon several percentage points above senior debt, filled a demand for yield inside portfolios that could not, or would not, buy shares.

Why do AT1 bonds pay higher interest than other bank debt?

Additional Tier 1 bonds pay higher interest because holders accept deeper subordination, a coupon that the bank may cancel, no fixed repayment date, and a possible permanent write-down, so the coupon compensates for risks that senior bondholders do not bear in the same form or degree.

The market is built as an over the counter market in large denominations rather than as a retail exchange market. New issues in the euro market have commonly used denominations of 200,000 euros, a size designed to restrict direct purchase to professional and wealthy investors. Issues in Asia have used denominations of 250,000 United States dollars or the equivalent in Singapore dollars or Hong Kong dollars. Regulators in several jurisdictions therefore looked beyond denomination to distribution practice, suitability assessment, and the way a coupon was presented to a client whose main goal was capital preservation.

Geography shaped the market in another way. European banks issued additional Tier 1 capital mainly in euros, United States dollars, and pounds sterling. Banks in Japan, Australia, China, and India issued in domestic currency under domestic implementations of Basel III. Terms that look identical in a summary, such as a 5.125 percent common equity Tier 1 trigger, can therefore sit inside contracts governed by different laws, with different point of non-viability language, different tax treatment, and different resolution authorities. Index providers grouped the instruments for performance measurement, and trading desks quoted them as an asset class, but the Credit Suisse write-down in March 2023 and the Yes Bank write-down in March 2020 each turned on jurisdiction specific language, a reminder that aggregation for trading does not create uniformity in law.

The Yield Premium and What It Actually Compensates

The coupon on an additional Tier 1 bond compensates a holder for four distinct risks, and separating them clarifies what a yield premium does and does not prove. The first risk is subordination. In liquidation and in resolution, the holder ranks below depositors, senior creditors, and Tier 2 holders, and above only common shareholders in the ordinary hierarchy. The second risk is extension. The bond has no maturity date, and the issuer may choose not to call it at the first call date, leaving the holder with a perpetual instrument whose coupon may be unattractive relative to market rates at that time. A decision not to call is not a default under the terms, yet it alters the economic life of the investment from a five year or ten year expectation to an indefinite holding.

The third risk is coupon cancellation. The bank may cancel a coupon in full or in part, the cancellation is not cumulative, and the missed payment never returns. Cancellation can occur by choice of the issuer, within the limits of distributable items, or by compulsion under the maximum distributable amount framework in the European Union and under analogous supervisory expectations in other jurisdictions. A senior bondholder faces a different bargain: a missed senior coupon is ordinarily an event of default with legal consequences. An additional Tier 1 holder accepts coupon fragility as a defining feature, and the yield premium includes payment for income that can disappear in a year in which the bank remains open and continues to lend. The fourth risk is principal loss through write-down or conversion. Conversion replaces a bond with shares whose market value at conversion may be a fraction of face value. A permanent write-down eliminates the claim entirely. Pricing that risk requires a judgment about trigger distance, about the credibility of the authority, and about the likely treatment of equity in the same rescue.

No forecasting follows from the premium either. A high coupon does not predict that a write-down will occur, and a low coupon does not predict that a bank will survive without support. During long periods of low interest rates, spreads on additional Tier 1 capital compressed in Europe and in Asia, even for banks with weak profitability, because demand for yield outweighed discrimination among issuers. Both patterns show the limits of price as information. Spread compression can understate risk, and spread widening can overstate the risk of an unrelated issuer. The premium is best read as the price of a bundle of contractual fragilities at a moment in time, negotiated between an issuer that needs the instrument to count as capital and a buyer who will accept fragility only for additional income. Nothing in that negotiation certifies the quality of bank management, the accuracy of loan loss recognition, or the wisdom of a resolution authority.

Mis-Selling Claims and the Retail Investor Problem

Mis-selling claims after the Yes Bank write-down in India exposed a structural problem that extends beyond one bank and one country: a loss-absorbing instrument had reached savers who, according to the claims filed, understood the purchase as a safe income product. Holders of Yes Bank additional Tier 1 bonds included individuals, family trusts, and small institutions in India. Claimants alleged that bank staff and distributors had presented the bonds as a high-yield substitute for a fixed deposit, emphasized the coupon and the reputation of the bank, and minimized or omitted the point of non-viability clause, the possibility of permanent write-down, and the absence of deposit insurance coverage for the bonds. Those statements are allegations made by named claimants in petitions and complaints in India. They are recorded here as claims. A court determination about the conduct of a particular seller toward a particular buyer depends on documents signed, disclosures given, and conversations proved in evidence, and no general account can convert an allegation into an established fact about every sale.

Regulatory responses in India addressed valuation, distribution, and disclosure after March 2020. The Securities and Exchange Board of India, the securities regulator in India, issued measures in October 2020 that treated perpetual bonds, including additional Tier 1 instruments, with a deemed maturity of 100 years for valuation by mutual funds, limited the exposure of a mutual fund scheme to such instruments, and restricted investment in bonds with special features. The effect of the valuation measure was to make price risk visible in fund accounting rather than to price a perpetual bond as if repayment in ten years were assured. The Reserve Bank of India, in guidance and in revised instructions for issuance in India, emphasized disclosure of risk, suitability for investors who could bear a total loss, and the responsibility of issuing banks for the conduct of distribution.

Other jurisdictions reached similar conclusions through different tools. In the European Union, product governance rules under the Markets in Financial Instruments Directive required manufacturers and distributors to define a target market and to exclude clients for whom a loss-absorbing, perpetual, coupon-cancellable instrument was unsuitable. The shared principle across jurisdictions was that complexity and loss absorption restrict the appropriate buyer base. A product designed to fail safely for the financial system can fail catastrophically for a household if the household placed retirement savings in it on the understanding that failure was impossible.

The retail problem also feeds back into resolution politics. An authority deciding whether to impose losses faces a different public reaction when holders are pension funds with diversified portfolios than when holders are elderly savers in the same city as the failing bank. Preventing unsuitable distribution in calm times is therefore part of resolution policy, not a separate consumer topic. Suitability checks, clear statements that a coupon can be canceled and principal can be written to zero, and a prohibition on deposit language in marketing material protect the credibility of a write-down long before a bank approaches the point of non-viability.

Coupon Restrictions and the Maximum Distributable Amount

Coupon payments on additional Tier 1 bonds in the European Union sit inside a system of automatic restrictions that links capital buffers to payments to investors and staff. The Capital Requirements Directive, the European Union directive governing prudential supervision, requires a bank to maintain a combined buffer above its minimum capital requirements, built from the capital conservation buffer, the countercyclical buffer, and, for some banks, systemic buffers. The buffers consist of common equity Tier 1 capital. A bank that operates inside its combined buffer has not breached its minimum requirement and has not triggered resolution, yet European Union law treats buffer use as a condition that demands restraint. The restraint operates through the maximum distributable amount, a calculated ceiling on the total amount a bank may pay out in dividends on common shares, in coupons on additional Tier 1 instruments, and in variable remuneration to staff.

The calculation is mechanical in structure. A bank that fails to meet the combined buffer requirement calculates the maximum distributable amount from its distributable profits, scaled by a factor that depends on how deeply the bank has entered the buffer range. A bank near the top of the buffer range may distribute a larger fraction. A bank in the lowest quartile of the buffer range has a maximum distributable amount of zero under the European Union formula. Inside the ceiling, the board allocates payments among dividends, additional Tier 1 coupons, and bonuses, subject to supervisory oversight and to a capital conservation plan that the bank must submit to its supervisor. A coupon canceled under this framework is canceled by operation of a prudential rule, not by a finding that the bank has defaulted, and the cancellation does not give holders a right to accelerate repayment, because the instrument has no maturity to accelerate toward.

For holders, the maximum distributable amount changes the meaning of a coupon promise. The relevant question is not only whether a bank earned a profit in a quarter, but whether the bank holds enough common equity Tier 1 capital above its combined requirement to permit a distribution of the planned size. Banks therefore disclose the distance to the maximum distributable amount trigger in investor reports, expressed as the amount of common equity Tier 1 capital above the threshold at which restrictions begin. A shrinking distance warns holders that coupons are at risk before any trigger for write-down is approached. The framework also creates an ordering debate inside the restricted stack. European Union law places dividends, additional Tier 1 coupons, and bonuses inside the same ceiling rather than ranking the coupon as a mandatory payment ahead of dividends, so a board that wishes to preserve cash can reduce each element, and a supervisor can press for restraint across all three.

The approach in the United Kingdom offers a close comparison with a different institutional accent. The Bank of England, through the Prudential Regulation Authority, implements a capital buffer framework in the United Kingdom that also restricts distributions when buffers are used, and the Prudential Regulation Authority Rulebook contains the calculation for the maximum distributable amount in United Kingdom law after the departure of the United Kingdom from the European Union. The Bank of England has emphasized, in Financial Stability Reports and in speeches by officials of the Bank of England, the usability of buffers in stress: a bank should be able to enter the buffer range to support lending without facing market stigma so severe that management defends the buffer as if it were a minimum. The maximum distributable amount sits exactly at that tension, converting a capital shortfall measured in percentage points into a concrete decision about cash leaving the bank.

The Complication: The Case That AT1 Is Debt Pretending to Be Capital

The strongest objection to additional Tier 1 capital holds that the instrument solves a political problem rather than an economic one: governments wanted banks to appear better capitalized after the financial crisis of 2008 and 2009 without requiring shareholders to supply, or accept dilution from, the far larger amounts of common equity that genuine loss absorption would demand. Anat Admati, a professor at Stanford University, and Martin Hellwig, a scholar at the Max Planck Institute for Research on Collective Goods in Germany, gave that objection its most developed form in The Bankers’ New Clothes, published in 2013. Admati and Hellwig argued that banks could and should operate with equity funding in the range of 20 percent to 30 percent of total assets, far above Basel III requirements, and that high equity would absorb losses automatically, in any order of events, without a trigger, without a viability determination, and without litigation over contract language.

The Squam Lake Group, a group of financial economists in the United States that published a report on contingent capital in 2010, offered a related but distinct criticism. The Squam Lake Group preferred a form of contingent capital whose conversion would be triggered automatically by a regulatory declaration tied to system wide stress, rather than by the capital ratio of an individual bank alone or by discretionary action against one institution. A regulatory trigger linked to a declaration that the financial system as a whole faced stress was intended to convert debt to equity across many banks at once, recapitalizing the system rather than singling out the weakest institution for a punishment that markets might treat as a death sentence.

The record of rescues gives the critics material that cannot be dismissed as theory. Discretionary and contract based triggers produced litigation in India after the Yes Bank write-down in March 2020 and in Switzerland after the Credit Suisse write-down in March 2023, litigation that continued for years and that left investors unsure whether the hierarchy printed in a prospectus would govern in an emergency. Discretion also invites political pressure in both directions. An authority may delay action to protect a national champion or to avoid losses for domestic savers who hold the bonds. An authority may act abruptly, with emergency legislation adopted on a Sunday, to protect depositors and a buyer. Admati and Hellwig would add that none of those dilemmas arises with common equity already issued and paid in: a share price can fall to near zero without a clause, a hearing, or an ordinance, and the loss is absorbed as it occurs.

The thesis of the preceding sections must therefore be narrowed in response. Additional Tier 1 capital performs as gone-concern capital only where the trigger, the authority, and the loss order are mechanical and pre-committed. Gone-concern capital means capital available to absorb losses when a bank can no longer continue as a going concern without support. Where the trigger depends on a contested valuation, where the authority acts on a draft scheme or on emergency legislation written for one transaction, or where shareholders retain value while hybrid holders are erased, the instrument behaves less like capital and more like a subordinated loan with an uncertain expropriation risk. Where, by contrast, equity is eliminated first, the trigger follows a published ratio or a formal failing determination, and the acting authority is the authority named in law and in contract, as in the Banco Popular resolution in June 2017 in the European Union, the instrument does the work claimed for it. The narrowing does not concede that additional Tier 1 capital is useless. The narrowing concedes that usefulness is conditional, and that the condition is institutional discipline rather than financial engineering.

What Would Settle the Argument: Evidence From Rescues Compared

A comparison of Banco Popular, Yes Bank, and Credit Suisse on the same criteria offers the best available evidence about additional Tier 1 capital, and the limits of that comparison define what remains unsettled. The first criterion is whether the loss order applied was the loss order promised. In the Banco Popular resolution in June 2017 in the European Union, the applied order matched the promised order: common shares were eliminated, additional Tier 1 instruments were written down, Tier 2 instruments were written down, and senior creditors and depositors were protected. In the Yes Bank reconstruction in March 2020 in India, the applied order departed from the expectation of many holders, because additional Tier 1 bonds worth 84 billion rupees were erased while existing equity survived in diluted and locked-in form. In the Credit Suisse takeover in March 2023 in Switzerland, the applied order departed more sharply in appearance, because additional Tier 1 instruments worth 16 billion Swiss francs were written to zero while shareholders received consideration in UBS shares.

The second criterion is whether the authority acted inside the scheme that authorized the action. The Single Resolution Board acted under the European Union Bank Recovery and Resolution Directive and a resolution scheme adopted for Banco Popular. The Yes Bank administrator acted under a reconstruction scheme notified by the Government of India, but the Bombay High Court in January 2023 and the Supreme Court of India in October 2026 each found defects in the exercise of that authority, with the Supreme Court of India holding the write-off illegal. FINMA acted under contract language and an emergency ordinance adopted by the Swiss Federal Council on the day of the Credit Suisse transaction, a foundation whose legality bondholders challenged in Swiss courts as a use of emergency power to alter creditor rights.

The third criterion is whether litigation followed, and the fourth criterion is whether depositors or public money bore losses. Litigation followed all three rescues, although the scale and the target differed: valuation and compensation claims after Banco Popular, hierarchy and procedure claims after Yes Bank, and legality and property claims after Credit Suisse. Depositors bore losses in none of the three cases, a consistency that reflects a shared policy priority in the European Union, India, and Switzerland. Public money bore losses in none of the three resolutions in direct form. Banco Popular was sold for one euro without a state capital injection. Credit Suisse received Swiss National Bank liquidity backed by a federal guarantee and a federal loss guarantee for UBS on a defined asset portfolio, support that created a contingent public exposure even though the guarantee was not a payment to additional Tier 1 holders.

What the comparison shows is that the instrument itself was constant in broad design across the three cases, while the institutional setting determined whether the design was trusted afterward. Trigger language, hierarchy, and authority, stated in advance and applied without improvisation, produced acceptance of a severe outcome in Spain. Ambiguity about sequence in India and emergency improvisation in Switzerland produced durable challenges to legitimacy, even though depositors were protected in each country. What the comparison cannot show is equally important. Three cases in three different legal systems, involving banks of different size and business models, cannot establish how additional Tier 1 capital would perform in a system wide crisis affecting many issuers at once. No case supplies a counterfactual: whether Banco Popular would have caused greater contagion as a liquidation, whether Yes Bank would have survived without a write-down, or whether UBS would have acquired Credit Suisse on any terms that preserved additional Tier 1 value. Evidence from rescues compared can discipline the argument and expose the conditions for success, but settlement of the argument would require a larger record, including rescues that have not happened and crises whose shape cannot be specified in advance. The capital stack table compresses that promised order into one view, and each rescue can be scored against it layer by layer.

AT1 and Tier 2: Two Kinds of Debt-Like Capital Compared

Additional Tier 1 capital and Tier 2 capital sit in neighbouring layers of a bank capital stack, and both take the legal form of debt, yet the resemblance ends at that shared form. Tier 2 instruments under the Basel III framework have a fixed maturity of at least five years, a known repayment date, and a claim that ranks ahead of Additional Tier 1 instruments in liquidation. Additional Tier 1 instruments have no maturity date at all. An issuer may redeem an Additional Tier 1 instrument after a minimum period specified in the contract and only with supervisory approval, but a holder has no contractual right to demand repayment on a set date. The distinction matters for permanence. Capital that can leave a bank on a fixed schedule cannot be counted on during a prolonged period of losses in the same way as capital that stays unless a supervisor permits redemption, and Basel III treats permanence as a defining quality of the higher layer.

Loss absorption separates the two layers by timing as much as by amount. Tier 2 capital is designed principally as gone-concern capital, meaning that losses fall on Tier 2 holders when a bank has failed or entered resolution and the remaining assets are distributed or transferred. Additional Tier 1 capital is intended to absorb losses on a going-concern basis as well, meaning that a write-down or conversion can occur while a bank continues to operate, take deposits and make payments. A bank can therefore restore a capital ratio by reducing an Additional Tier 1 liability before any decision to close the institution has been taken. Tier 2 instruments also carry a point of non-viability clause in jurisdictions that apply the Basel III requirement, so Tier 2 holders are not protected at the moment of failure, but the ordinary path for Tier 2 loss absorption runs through resolution rather than through an early contractual trigger linked to a capital ratio.

Trigger design completes the comparison. An Additional Tier 1 instrument carries a contractual Common Equity Tier 1 trigger, and Basel III sets the floor for that trigger at 5.125 percent of risk-weighted assets for instruments that rely on a ratio trigger. When the ratio falls to or below the stated level, the instrument is written down or converted in the amount specified in the contract. Tier 2 instruments do not carry an equivalent going-concern ratio trigger as a standard Basel III feature. Both layers can be exposed to a statutory or contractual point of non-viability determination by the relevant authority, but only Additional Tier 1 combines that official determination with an earlier mechanical trigger and with routine coupon cancellation. The practical result is an ordering that a holder can read in advance: Common Equity Tier 1 absorbs losses first, Additional Tier 1 absorbs losses next, Tier 2 follows, and senior debt stands above Tier 2 unless a resolution authority applies bail-in powers under the law of the governing jurisdiction.

Why Banks Prefer AT1 to Issuing New Shares

A bank that needs more loss-absorbing capital has a choice between issuing common shares and issuing Additional Tier 1 instruments, and the arithmetic of that choice explains much of the instrument appeal to issuers. Common equity is the most reliable form of capital because a share has no coupon, no maturity and no claim that ranks ahead of depositors, yet equity is also the most expensive form of funding for an issuer to raise and maintain. Investors in shares require a return that compensates for standing last in the loss order, and a bank that issues a large volume of new shares at a depressed price transfers value to new holders and dilutes the ownership share and earnings per share of existing holders. Additional Tier 1 instruments offer a middle position. The coupon is higher than the rate on senior debt, reflecting subordination and cancellation risk, but in many issuance programmes the all-in cost has been lower than the cost of equity implied by share prices and required returns, and the instrument adds regulatory capital without changing voting control.

Return on equity arithmetic sharpens the incentive. Return on equity divides profit by common equity, so an expansion of the common equity base lowers the ratio unless profit rises in proportion. A bank managed against a return on equity target, whether set by a board or watched by shareholders, faces pressure to meet a capital requirement with the smallest increase in common equity that the rules permit. Additional Tier 1 instruments count toward total Tier 1 capital without enlarging the common equity denominator in the same way, although the coupon reduces profit available to common shareholders. Tier 2 instruments count only toward total capital and not toward Tier 1 capital, so Additional Tier 1 fills a specific regulatory bucket that Tier 2 cannot fill. That bucket logic, rather than any single price comparison, accounts for repeated issuance by large banks with access to institutional bond markets.

Tax treatment adds a jurisdiction-specific element that an issuer weighs alongside regulatory value, and treatment varies rather than following a single international rule. In some jurisdictions a coupon on an instrument classified as debt for tax purposes has been treated as a deductible expense, while a dividend on a share is paid from after-tax profit, and that difference lowers the effective cost of the bond form. The position has been debated in India and in the European Union, where finance ministry documents and regulator documents have addressed classification, deductibility and the boundary between debt and equity for hybrid instruments in general terms, without a uniform outcome across member states or across instrument designs. A careful reading therefore separates two questions that marketing material often merges. Regulatory classification decides whether an instrument counts as Additional Tier 1 capital, and tax classification decides how a coupon is treated, and the two classifications do not have to match.

Rating agency treatment supplies a further incentive, also stated with limits. Rating agencies including Moody’s Investors Service, S and P Global Ratings and Fitch Ratings have at different times assigned partial equity credit or equity content to qualifying hybrid instruments in their capital adequacy assessments, based on permanence, subordination and coupon deferral features described in agency methodology documents. Equity credit from a rating agency is not regulatory capital and does not replace a Basel III test, but favorable agency treatment can support a rating outcome while a bank raises a layer that a supervisor does count, subject to the limits in each agency method. The combined effect is a set of reinforcing incentives: regulatory recognition without voting dilution, a coupon cost below the implied cost of equity in favorable markets, possible tax deductibility depending on the governing jurisdiction, and agency recognition that varies by method. Each incentive depends on conditions outside the instrument label alone, and each incentive weakens when coupons rise, triggers are questioned in litigation, or a supervisor restricts distributions.

The Design Debate Before Basel: Flannery, Calomiris and Herring, and the Squam Lake Group

Contingent capital began as an academic proposal for solving a timing problem in bank regulation. Supervisors often identified a failing bank after losses had consumed common equity, and a rescue at that late stage placed costs on deposit insurers, other creditors or public funds. Mark Flannery, an economist at the University of Florida, proposed reverse convertible debentures in 2005 as a market-based corrective. Under the Flannery design, a bank would issue debt that converted automatically into common shares when a market measure of bank value fell to a stated threshold. Conversion would recapitalise a bank without a supervisor having to declare failure, replenish equity from an existing creditor layer, dilute existing shareholders at the moment of weakness and reduce the debt burden that a distressed bank had to service. The elegance of the proposal lay in pre-commitment. Conversion terms would be fixed in the contract during a period of calm, so a crisis would execute a bargain already made rather than open a negotiation over who should bear a loss.

Charles Calomiris of Columbia University and Richard Herring of the University of Pennsylvania developed a related design in 2012, arguing for contingent capital with a market trigger linked to the market value of bank equity relative to a measure of liabilities or assets. In the Calomiris and Herring design, a sustained fall in a market-based ratio would trigger conversion of a specified portion of contingent debt into equity, and conversion in stages would create an incentive for managers and shareholders to raise equity or reduce risk before the trigger was reached. A market trigger had an informational claim behind it. Share prices aggregate judgments from many traders about asset quality and future losses, and a price signal can move faster than audited accounts or a quarterly regulatory ratio based on risk-weighted assets. The same authors recognized a design problem that any market trigger must answer. A trigger based on a single trading day could fire on noise, manipulation or a general market fall, so a workable design required averaging, thresholds and conversion amounts specified with care.

The Squam Lake Group, a group of academic economists convened after the financial crisis of 2007 to 2009, took a more supervisory path in a 2010 report on regulatory contingent capital. The Squam Lake proposal contemplated a class of hybrid capital that would convert into equity when a regulator declared a systemic crisis or when institution-specific conditions in the design were met, linking conversion to the state of the wider financial system as well as to the condition of a single bank. The emphasis differed from a pure market trigger. A systemic event changes the social cost of a bank failure because many institutions face funding pressure at the same time, and a regulator holds information about interconnections, liquidity support and resolution capacity that a market price does not reveal. Regulatory contingent capital in the Squam Lake form traded automaticity for context, accepting official discretion as a feature where a mechanical rule might convert capital at banks that could recover or fail to convert capital at banks whose market price had been supported by an expectation of public assistance.

Basel III adopted neither a pure Flannery design nor a pure Calomiris and Herring design. The Basel Committee on Banking Supervision framework published in 2010 and revised in 2017 relied on an accounting and regulatory capital ratio, the Common Equity Tier 1 ratio, with a contractual trigger floor of 5.125 percent, combined with a point of non-viability clause exercised by a supervisor or resolution authority. Market prices inform supervision, but a fall in a share price alone does not convert an Additional Tier 1 instrument under the standard Basel III structure. One reason lies in panic dynamics. A visible market trigger can become a target. As a price approaches a conversion threshold, short selling, hedging by holders and withdrawal by short-term creditors can push a price down, trigger conversion in disorderly conditions and accelerate the very run dynamic that contingent capital was meant to calm. Supervisor discretion avoids a single public cliff edge, at the cost of uncertainty about whether an authority will act early, act at all, or act in an order that bondholders regard as consistent with the hierarchy stated in the contract.

Stress Tests and the Supervisory Layer Around AT1

A stress test places a bank balance sheet inside a severe hypothetical scenario and estimates whether capital remains above required minima after losses on loans, trading positions and other exposures. The exercise surrounds Additional Tier 1 capital with a supervisory layer that a contract alone does not provide. In the European Union, the European Banking Authority has coordinated Union-wide stress tests in cooperation with national supervisors and the European Central Bank, publishing bank-level results for capital ratios under baseline and adverse scenarios. In the United Kingdom, the Bank of England has run annual stress tests of major banks through the Prudential Regulation Authority and the Financial Policy Committee, with results expressed in Common Equity Tier 1 ratios and leverage ratios after modelled losses and management actions allowed under the test rules. In the United States, the Board of Governors of the Federal Reserve System has run supervisory stress tests under the Dodd-Frank Act framework, a comparative case in which large bank holding companies rely more heavily on common equity and preferred shares within a capital plan and capital review process.

A stress test can reveal information that bears directly on Additional Tier 1 coupons and triggers, within the limits of the scenario chosen. A published adverse-scenario ratio shows how far a Common Equity Tier 1 ratio falls toward a contractual trigger set at 5.125 percent or higher, after the losses assumed in the scenario have been applied to risk-weighted assets. A test can also show whether a bank breaches a combined buffer requirement that feeds into a maximum distributable amount calculation in the European Union, a breach that restricts dividends, bonuses and Additional Tier 1 coupons by a formula linked to available distributable items and buffer use. Supervisors can require a bank that performs poorly in a test to submit a capital plan, restrict distributions or raise common equity before a contractual trigger is reached. In that supervisory role, a test operates as an early intervention ladder. A deteriorating test result gives an authority a documented basis for action while a bank still reports a ratio above a trigger, and disclosure gives creditors a common set of figures for comparing depletion across banks.

A stress test cannot show several things that matter most in a real failure. A scenario is a model, and a model depends on assumptions about house prices, unemployment, interest rates, market liquidity and correlations that a crisis can violate in combination. A test conducted on a reporting date cannot capture a deposit run that develops in days, an intraday liquidity shortage, a legal dispute over a contract clause, or a decision by a finance ministry about the terms of a merger arranged over a weekend. The Credit Suisse rescue in March 2023 illustrates the limit without proving that a test failed on the test terms. Long-running concerns about profitability, risk controls and deposit outflows preceded a rescue in which the Swiss Financial Market Supervisory Authority ordered a write-down of 16 billion Swiss francs of Additional Tier 1 instruments, and the speed of the final transaction owed as much to liquidity and confidence as to a modelled capital ratio at a fixed date.

Jurisdictions Compared: India, the European Union, Switzerland, the United Kingdom and the United States

India writes the Additional Tier 1 trigger and the point of non-viability power through the Reserve Bank of India, acting under the Banking Regulation Act, 1949 and under master directions that adapt the Basel III framework to Indian banks. An Indian Additional Tier 1 instrument carries a contractual loss absorption clause linked to the Common Equity Tier 1 ratio and a separate power for the Reserve Bank of India to require a write-down when the bank reaches a point of non-viability, a judgment tied to resolution, reconstruction or the survival of the bank as a going concern. The Yes Bank reconstruction in March 2020 placed that power inside a government-notified scheme. An administrator appointed under the scheme wrote down 84 billion rupees of Additional Tier 1 instruments, and litigation in the Bombay High Court and later in the Supreme Court of India tested whether the administrator, rather than the Reserve Bank of India itself, held lawful authority for the write-down and whether required procedure had been followed. The Indian case therefore separates two questions that a single trigger clause can appear to merge: whether losses should fall on a layer, and which named official holds legal power to impose the loss in the required form.

The European Union writes the trigger through the Capital Requirements Regulation, Regulation (EU) No 575/2013, and writes resolution action through the Bank Recovery and Resolution Directive, Directive 2014/59/EU. An Additional Tier 1 instrument issued by a bank in the European Union states a Common Equity Tier 1 trigger at or above the Basel III floor of 5.125 percent, gives an issuer full discretion to cancel coupons, and sits inside a buffer framework that restricts distributions through the maximum distributable amount. A point of non-viability or failing-or-likely-to-fail determination involves a competent supervisor, including the European Central Bank for significant institutions in the Banking Union, and resolution action for a failing bank falls to the Single Resolution Board for institutions within its remit, working with national resolution authorities. The Banco Popular resolution in June 2017 applied that architecture in a single night. The Single Resolution Board wrote down equity, Additional Tier 1 instruments and Tier 2 instruments and sold Banco Popular to Banco Santander for one euro, with depositors and senior creditors protected in the transfer and with the loss order matching the hierarchy that the instruments described.

Switzerland writes the instrument through the Swiss Financial Market Supervisory Authority, known as FINMA, under Swiss banking law and capital ordinances that adapt Basel principles to Swiss institutions, including systemically important banks subject to Swiss too-big-to-fail requirements. Swiss Additional Tier 1 instruments have included high-trigger and low-trigger forms, with contractual language on viability events and on action by FINMA when a bank faces a point of non-viability or receives extraordinary public support. The Credit Suisse transaction in March 2023 tested the boundary between contract language and the ordinary expectation that equity absorbs losses before hybrid capital. FINMA ordered a complete write-down of 16 billion Swiss francs of Additional Tier 1 instruments under the terms that FINMA identified in the instruments and in emergency law adopted for the transaction, while shareholders in Credit Suisse received shares in UBS as merger consideration rather than a zero outcome. Bondholder litigation followed in Swiss and foreign courts, with claims centerd on the legal basis for the write-down and on the treatment of the hierarchy, and the episode remains a reference point for drafting viability clauses with explicit language on shareholder treatment.

The United Kingdom writes the trigger through the Prudential Regulation Authority rulebook, which applies the capital framework inherited from European Union law and adapted after departure from the European Union, and resolution powers rest with the Bank of England under the Banking Act 2009. A point of non-viability assessment in the United Kingdom involves the Prudential Regulation Authority and the Bank of England in defined roles, and a bail-in in resolution follows the creditor hierarchy in insolvency subject to safeguards in the Banking Act 2009, including a principle that a creditor should not be worse off than in liquidation. The United States presents the contrasting structure. Large United States banks have relied more heavily on non-cumulative perpetual preferred shares, which are equity instruments with discretionary dividends and no maturity, than on Additional Tier 1 bonds of the European contractual write-down form, and total loss-absorbing capacity for global systemically important banks in the United States has been built principally from common equity, preferred shares and long-term debt that can be bailed in at a holding company under a single-point-of-entry strategy described by the Federal Deposit Insurance Corporation and the Board of Governors of the Federal Reserve System. Preferred shares absorb losses as equity without a ratio trigger clause of the Additional Tier 1 bond type, and the difference explains why a comparison across jurisdictions must compare functions in a capital stack rather than instrument names alone. A label travels across borders more easily than the legal power behind a label, and a holder who compares only coupon and trigger level across India, the European Union, Switzerland, the United Kingdom and the United States misses the authority, the court and the hierarchy that decide a contested case.

How to Read a Bank Rescue and Find the AT1 Layer

A bank rescue announcement compresses a capital stack, a legal power and a set of instrument contracts into a few paragraphs written under time pressure, and a methodical reading separates the three elements before any judgment about fairness or loss. The first step is the capital stack in the annual report and in the Pillar 3 disclosure of the bank named in the announcement. A reader can locate a table that lists Common Equity Tier 1 capital, Additional Tier 1 capital and Tier 2 capital as amounts and as ratios to risk-weighted assets, together with a list of outstanding instruments by issue date, coupon, trigger level and governing law. That table gives scale. A write-down figure reported in a headline acquires meaning only beside the size of the layer written down, the size of the common equity layer beneath the layer and the size of total assets and deposits that a rescue aims to keep functioning. A reader who starts with the stack can state which layer was touched, which layer was left untouched and which layer remains available if losses continue.

The second step is the prospectus or offering circular for the specific instrument, a document filed by an issuer and stored by a stock exchange, a securities regulator or an issuer investor relations archive, depending on the listing jurisdiction. In that document a reader can find the trigger clause stated as a Common Equity Tier 1 ratio, the definition of a trigger event, the amount and method of a write-down or conversion, the ranking clause, the coupon cancellation language and the point of non-viability clause that names the authority entitled to make the determination. A short checklist kept beside a prospectus, or recorded with a companion reference tool, keeps the search focused on clauses rather than on summary marketing pages. Particular attention belongs to three drafting details. A clause may permit a temporary write-down with a later write-up under stated conditions, or a permanent write-down with no restoration. A conversion clause may deliver a fixed number of shares, a variable number based on a floor price, or shares only in a jurisdiction where an issuer has authorized share capital available for delivery. A point of non-viability clause may define the event by reference to a regulator decision, to receipt of public support, or to both, and the wording decides which official act starts loss absorption.

How can a reader tell which investors lose money first in a bank rescue?

You can tell by reading the ranking clause in the prospectus together with the resolution decision: common shareholders stand first in the loss order, Additional Tier 1 holders stand second, Tier 2 holders stand third, and senior creditors and depositors rank above those layers unless a named authority applies a different statutory power in writing.

After that direct answer, the reading method continues with a comparison that you can perform on documents alone. Place the promised order from the prospectus beside the applied order in the official decision, line by line, for equity, Additional Tier 1, Tier 2, senior debt and deposits, and mark each layer as wiped out, partly written down, converted, transferred at value or left in place. A mismatch at the equity line deserves the closest attention, because consideration paid to shareholders while an Additional Tier 1 layer is written to zero raises the hierarchy question that litigation after the Credit Suisse transaction placed before courts in several jurisdictions, while a full write-down of equity before any hybrid loss, as in the Banco Popular resolution, shows the promised order executed in the form a prospectus described. Dates matter in the comparison as well. A trigger measured at a quarter end, a regulator determination made on a weekend and a merger consideration fixed by negotiation belong to different clocks, and a reader who records each date can separate a mechanical trigger outcome from a negotiated transaction outcome that a resolution authority later ratified under emergency powers.

What AT1 Cannot Do: The Limits of Hybrid Capital

Additional Tier 1 capital absorbs a defined amount of loss for a defined set of creditors, and the definition itself marks the boundary of what the instrument can achieve. A bank run is a liquidity event driven by a belief among depositors and short-term creditors that others will withdraw first, and a capital instrument with no maturity does nothing to supply cash on the day withdrawals accelerate. A bank can report a Common Equity Tier 1 ratio above a contractual trigger and still lose access to funding markets, lose correspondent relationships or face deposit outflows that force asset sales at distressed prices. The Federal Deposit Insurance Corporation resolution of First Republic Bank in May 2023, handled by sale and deposit transfer to another bank rather than by bail-in of a hybrid layer, belongs in the same analytical frame, because the decisive problem in the final phase was a concentration of uninsured deposits and unrealized losses on long-term assets, addressed by a transfer that kept branches open and deposits available on the following business day. Additional Tier 1 capital did not feature as the mechanism that preserved continuity, and continuity was the outcome depositors experienced.

Liquidity requirements exist precisely because capital alone does not answer a run. The Basel III liquidity coverage ratio requires a bank to hold high-quality liquid assets against modelled net cash outflows over a thirty-day stress period, and the net stable funding ratio requires stable funding matched to the liquidity profile of assets over a one-year horizon, as described in Basel Committee on Banking Supervision standards published in 2013 and 2014. A supervisor also holds tools that operate in hours rather than in accounting quarters, including central bank lending against collateral and restrictions administered by a resolution authority after appointment. Additional Tier 1 loss absorption operates on a slower and narrower track. A write-down improves a capital ratio by reducing a liability or creates common equity on conversion, but a write-down produces no cash inflow, sells no asset and reassures no depositor standing at a branch or moving funds through an electronic transfer. Treating a capital ratio as a substitute for liquidity confuses solvency, a comparison of assets and liabilities over time, with liquidity, the capacity to pay obligations as obligations fall due.

Scale imposes a fourth limit that a simple balance sheet comparison makes plain. A large international bank can carry total assets measured in trillions of dollars, Swiss francs, euros or rupees, while an Additional Tier 1 layer is sized as a percentage of risk-weighted assets, a denominator far smaller than total assets because sovereign exposures, collateralised claims and other assets carry low risk weights. A complete write-down of an entire Additional Tier 1 layer can restore a ratio and impose a substantial loss on the holders of that layer, as the write-down of 16 billion Swiss francs at Credit Suisse in March 2023 demonstrated for the holders affected, without covering losses that a severe recession, a property collapse or a fraud could generate across a full loan book. Additional Tier 1 capital is best understood as a buffer with a specific place in a sequence, sitting between common equity and Tier 2 capital and standing far below the volume of deposits and senior liabilities that a resolution must either protect, transfer or bail in under law. A buffer of that size can change incentives, allocate a first tranche of creditor loss and give a supervisor a tool short of liquidation, and a buffer of that size cannot by itself make a large bank safe against every loss that a balance sheet of that scale can produce. When private loss absorption of every layer is exhausted, the remaining options are public ones, and the state either recapitalises the institution from public funds or supports the households and firms around it, the machinery examined in how government cash payments work.

The Structural Significance of Loss-Absorbing Debt

Loss-absorbing debt changes the politics of a bank failure by moving a contested decision from a crisis weekend into a contract signed years earlier. A prospectus that states a trigger at 5.125 percent, names the authority that can declare a point of non-viability and ranks Additional Tier 1 below deposits and senior debt and above nothing except common equity gives every later actor a text against which conduct can be measured. A supervisor can point to the text when imposing a loss, a holder can point to the text when challenging a loss in court, and a court can decide a dispute by reading the text beside the statute that empowered the supervisor, as courts in India did after the write-down of 84 billion rupees at Yes Bank in 2020 and as courts seized of claims after the write-down of 16 billion Swiss francs at Credit Suisse in 2023 were asked to do. The discipline is imperfect because emergency law can alter a transaction and negotiation can deliver value to a layer that a ranking clause placed first in the loss order, yet even a contested application reveals the function of the instrument. Additional Tier 1 capital exists to make a loss allocation legible before a loss occurs.

The deeper significance lies in the distribution of cost that every banking system must settle in some form. A failure generates losses that exceed common equity in a severe case, and the excess must fall on hybrid holders, on other creditors, on a deposit insurance fund financed by banks, on acquiring shareholders in a rescue merger, or on public funds supplied by taxpayers under authority of a finance ministry and a legislature. Earlier systems often settled the distribution during a crisis, under time pressure, with incomplete information and with strong incentives to protect every creditor class that could spread panic by withdrawing funds. A contractual loss-absorbing layer settles part of the distribution in advance, in public documents, at a price reflected in a coupon that compensates a holder for a stated risk. The argument over Additional Tier 1 design, over trigger levels, over market triggers against supervisory discretion and over the treatment of shareholders in March 2023 and in March 2020, is therefore an argument over who pays for bank failure, decided in advance by contract or decided in a crisis by politics.

Frequently Asked Questions

Q: What are AT1 bonds in simple terms?

Additional Tier 1 bonds are subordinated debt instruments that a bank counts as regulatory capital under the Basel III framework because the terms allow losses to fall on holders while a bank continues operating. A holder lends money for an indefinite period, receives a coupon that an issuer may cancel, and accepts a place near the bottom of the creditor ranking, above common shares alone. If a stated capital trigger is breached or a supervisor declares a point of non-viability, principal can be reduced, canceled or exchanged for shares under the contract. The higher coupon compared with senior debt reflects subordination, coupon cancellation risk and the possibility of a total principal loss in a rescue or resolution governed by the law stated in the prospectus.

Q: Are AT1 bonds the same as CoCo bonds?

The two labels describe overlapping instruments rather than two separate capital layers. CoCo, short for contingent convertible, refers to a design feature: a bond that converts into equity or absorbs losses when a contingent event defined in a contract occurs. Additional Tier 1 refers to a regulatory classification under Basel III, describing where an instrument sits in a capital stack and which criteria on permanence, subordination and coupon discretion an instrument satisfies. Many Additional Tier 1 instruments are CoCos because conversion or write-down follows a capital trigger, and a contingent convertible instrument can also be structured to qualify as Tier 2 capital in a particular jurisdiction. Classification depends on the full terms and on recognition by the supervisor named in the governing framework.

Q: Why do AT1 bonds have no maturity date?

Perpetuity serves a regulatory purpose connected to availability during distress. A dated instrument must be repaid or refinanced, and a repayment obligation that falls due during a period of losses drains resources at the moment a bank needs resources most. Basel III therefore requires Additional Tier 1 capital to be perpetual, with redemption possible only at an issuer option after a minimum period, only from a stated call date in the contract, and only with prior supervisory approval in jurisdictions applying the Basel criteria. An issuer also faces a rule against creating an expectation of redemption that would undermine permanence. Permanence allows a supervisor and a resolution authority to treat the layer as available to absorb losses across an indefinite horizon rather than as funding scheduled to depart.

Q: What happens to AT1 bonds when a bank fails?

Treatment in a failure follows the creditor hierarchy and the resolution law of the jurisdiction where a bank is incorporated or resolved. Common equity absorbs losses first under the hierarchy described in Basel III and in the Financial Stability Board Key Attributes of Effective Resolution Regimes from 2011, followed by Additional Tier 1 instruments, followed by Tier 2 instruments, with senior debt and deposits ranking higher. A resolution authority can write principal down to zero, convert principal into shares in a recapitalised institution, transfer assets and protected liabilities to a purchaser while leaving capital instruments behind, or combine methods under statutory powers. Banco Popular in Spain in 2017 showed a full sequence, with equity and hybrid layers eliminated and a sale completed for one euro under a Single Resolution Board decision.

Q: What is a point of non-viability trigger?

A point of non-viability trigger is a clause that allows loss absorption when a supervisor or resolution authority judges that a bank will become non-viable without a write-down, a conversion or public support. Basel III required the clause in capital instruments so that a layer could absorb losses at the moment of failure even if a ratio trigger had not been breached under reported figures. The determination rests with an authority named in law or in the instrument, such as the Reserve Bank of India in India, FINMA in Switzerland, or the Bank of England and the Prudential Regulation Authority in the United Kingdom under their respective statutory roles. A declaration activates the remedy written into the contract or the resolution statute, and affected holders can challenge legality and procedure in a competent court.

Q: Can AT1 bondholders lose all their money while shareholders keep value?

Contract ranking points in the opposite direction, because common shares rank below Additional Tier 1 instruments and a hierarchy-respecting rescue eliminates share value before hybrid principal is reduced. A transaction structured as a merger rather than as a liquidation or a standard bail-in can complicate the outcome, because shareholders in a target bank can receive shares in an acquiring bank as negotiated merger consideration while a supervisor orders a hybrid write-down under a viability clause. The Credit Suisse transaction in March 2023 produced that pattern, with 16 billion Swiss francs of Additional Tier 1 instruments written to zero by order of FINMA and with Credit Suisse shareholders receiving UBS shares under merger terms. Bondholder claims filed after the transaction argued that the legal basis and the hierarchy treatment were defective, and courts were asked to examine those arguments.

Q: What happened to Credit Suisse AT1 bonds in 2023?

In March 2023 the Swiss government, the Swiss National Bank and FINMA arranged a takeover of Credit Suisse by UBS after deposit outflows and a collapse in market confidence placed the bank under acute pressure. FINMA ordered a complete write-down of Additional Tier 1 instruments with a nominal value of 16 billion Swiss francs, identifying contractual viability provisions and emergency measures adopted for the transaction as the legal basis for the order. Credit Suisse shareholders received UBS shares as merger consideration, a result that placed the episode at the center of a debate about creditor hierarchy in Switzerland and in other jurisdictions. Holders and holder groups pursued claims in Swiss proceedings and in other forums, alleging defects in authority, procedure and property protection, with outcomes depending on the forum and the instrument governing law.

Q: What happened to Yes Bank AT1 bonds in 2020?

In March 2020 the Reserve Bank of India placed Yes Bank under a moratorium and advanced a reconstruction scheme notified by the Government of India after the bank failed to raise sufficient capital and faced severe governance and asset quality problems. An administrator appointed under the scheme wrote down Additional Tier 1 instruments valued at 84 billion rupees, while a consortium led by the State Bank of India injected equity and the bank resumed operations under new leadership. Holders challenged the write-down, arguing that the administrator lacked authority for the decision and that required steps had not been followed. The Bombay High Court quashed the write-down in January 2023 on procedural grounds identified in the judgment, and the dispute continued through appellate proceedings in India concerning authority, procedure and the standing of the reconstruction scheme.

Q: How are AT1 bonds different from Tier 2 bonds?

The central differences concern maturity, coupon discretion and the stage at which losses are imposed. Tier 2 instruments under Basel III have an original maturity of at least five years and a scheduled repayment date, while Additional Tier 1 instruments are perpetual and repayable only under an issuer call approved by a supervisor. Additional Tier 1 coupons are discretionary and non-cumulative, so a canceled payment disappears, while Tier 2 coupons follow ordinary debt terms outside resolution. Additional Tier 1 instruments also carry a going-concern capital trigger linked to the Common Equity Tier 1 ratio, with a Basel III floor of 5.125 percent, while Tier 2 loss absorption operates principally at failure or resolution. Both layers rank below senior debt, and Additional Tier 1 ranks below Tier 2 in liquidation under the standard hierarchy.

Q: Who regulates AT1 bonds in India?

Regulation in India rests principally with the Reserve Bank of India for prudential capital treatment and with the Securities and Exchange Board of India for listed securities disclosure and market conduct, each acting within a statutory mandate. The Reserve Bank of India sets eligibility criteria for Additional Tier 1 instruments under directions that implement Basel III in India, approves inclusion in Tier 1 capital, supervises issuing banks and holds powers connected to reconstruction and to a point of non-viability determination under the Banking Regulation Act, 1949. The Securities and Exchange Board of India governs offer documents, listing obligations and trading disclosures for instruments listed on a recognized exchange. Courts in India, including the Bombay High Court and the Supreme Court of India, review legality when holders challenge an exercise of regulatory or administrator power after a write-down decision.

Q: Why do banks issue AT1 bonds instead of shares?

Issuance reflects a balance among regulatory recognition, cost, control and market capacity. Additional Tier 1 instruments count toward Tier 1 capital under Basel III without issuing voting shares, so existing shareholders avoid dilution of ownership and earnings per share that a large common share issue would produce. A coupon, even at a rate well above senior debt, can cost an issuer less than the return demanded by new shareholders for capital that ranks last for repayment, although market conditions can narrow or reverse the comparison. Treatment for tax purposes differs by jurisdiction and by instrument classification, and rating agency methods can assign partial equity content to qualifying hybrids. A bank with deep access to institutional bond demand can also raise a hybrid layer faster and in larger blocks than a share sale during weak equity markets.

Q: What is a CET1 trigger in an AT1 bond?

A Common Equity Tier 1 trigger is a threshold in a contract that activates loss absorption when a bank Common Equity Tier 1 ratio falls to or below a stated level. The ratio divides Common Equity Tier 1 capital, principally common shares and retained earnings after regulatory deductions, by risk-weighted assets calculated under rules set by the Basel Committee on Banking Supervision and implemented by a national supervisor. Basel III sets a floor of 5.125 percent for a trigger in a qualifying instrument, and many issuers set a higher level, such as seven percent, in the terms of a particular issue. Breach of the level obliges or permits an issuer, depending on contract wording, to write principal down or convert principal into common shares in the amount and sequence the prospectus specifies.

Q: Do AT1 coupons have to be paid every year?

Payment is not guaranteed in the manner of a senior bond coupon. Basel III criteria require Additional Tier 1 distributions to be fully discretionary, non-cumulative and payable only from distributable items, and an issuer decision to cancel a coupon does not constitute a default under a qualifying instrument. A supervisor can also restrict payment through capital buffer rules. In the European Union the maximum distributable amount mechanism limits dividends, variable pay and Additional Tier 1 coupons when a bank uses combined buffer capital, applying a formula tied to profits and to the extent of buffer consumption under the Capital Requirements Directive. Cancellation preserves cash and capital inside a bank during strain, and a holder receives no arrears for a canceled period because non-cumulative terms extinguish the skipped distribution rather than deferring payment to a later date.

Q: What is bail-in and how is it different from a bailout?

Bail-in describes a resolution method in which losses fall on shareholders and designated creditors of a failing bank under statutory powers, through a write-down or conversion of claims in hierarchy order. Bailout describes support supplied from outside a creditor hierarchy, principally public funds or central bank assistance provided under government authority, where protected creditors receive payment or continuity at public expense or with public backing. The Financial Stability Board Key Attributes of Effective Resolution Regimes from 2011 promoted bail-in capacity so that a large bank could be recapitalised or wound down without an assumption that taxpayers would absorb losses beyond common equity. European Union law in the Bank Recovery and Resolution Directive and United Kingdom law in the Banking Act 2009 both contain bail-in powers, with safeguards, exclusions and compensation principles defined by the governing statute.

Q: Who buys AT1 bonds?

Buyers are principally institutional investors, including asset managers, pension funds, insurance companies, hedge funds and private banks acting for clients, with composition varying by jurisdiction, listing venue and denomination size. A large minimum denomination in many European Union and United Kingdom issues limits direct retail access, while distribution through wealth channels can create indirect retail exposure that regulators in several jurisdictions have examined after mis-selling complaints and after losses in particular rescues. Pricing in a primary issue reflects the buyer assessment of subordination, trigger distance, coupon cancellation risk, issuer profitability and the legal record of the governing jurisdiction. Secondary trading takes place in over-the-counter bond markets rather than on a central equity exchange in most cases, and liquidity in a stressed period can fall sharply as buyers reassess trigger and resolution risk for an entire issuer group.

Q: Why are AT1 bonds considered risky for retail investors?

Risk for a retail holder arises from complexity, subordination and information asymmetry combined in a single instrument. A prospectus can run to hundreds of pages, define loss absorption through cross-referenced capital rules, and grant a supervisor a point of non-viability power whose exercise depends on facts that a small holder cannot observe directly. Coupon cancellation can occur without a default and without arrears, principal can fall to zero while a bank continues to serve depositors, and secondary prices can decline steeply before any trigger is breached. Regulators in the United Kingdom, in the European Union and in India have at different times restricted marketing, raised denomination expectations or issued warnings about suitability for ordinary savers, with the Financial Conduct Authority in the United Kingdom maintaining restrictions on promotion of contingent convertible instruments to retail clients under rules described in regulator publications.

Q: What does AT1 mean in Basel III?

AT1 is an abbreviation for Additional Tier 1, the second component of Tier 1 capital in the Basel III capital framework published by the Basel Committee on Banking Supervision in 2010 and revised in later framework consolidations. Tier 1 capital consists of Common Equity Tier 1, the highest quality layer built from common shares and retained earnings after deductions, plus Additional Tier 1 instruments that satisfy criteria on subordination, permanence, coupon discretion and loss absorption. Tier 1 capital sits above Tier 2 capital in quality because Tier 1 layers absorb losses on a going-concern basis, while Tier 2 functions principally as gone-concern capital. Minimum requirements in Basel III are expressed as ratios to risk-weighted assets, including a Tier 1 minimum of six percent, and Additional Tier 1 instruments help a bank meet the Tier 1 requirement beyond the Common Equity Tier 1 minimum.

Q: Can AT1 bonds be converted into shares?

Conversion is one of two principal loss absorption mechanisms found in Additional Tier 1 contracts, with write-down as the alternative, and some instruments combine a partial write-down with conversion provisions in a sequence set by the terms. A conversion clause specifies a trigger, a conversion price or a floor price, the number or value of shares delivered and any limits linked to authorized share capital in the jurisdiction of incorporation. Conversion recapitalises a bank by replacing a liability with common equity, dilutes pre-existing shareholders and gives a former bondholder an equity claim whose value depends on the condition of a bank after distress has emerged. Swiss, European Union and Indian instruments have used different mechanisms across issues, so treatment in a rescue depends on the prospectus for the specific series rather than on the Additional Tier 1 label alone.

Q: What is the maximum distributable amount for AT1 coupons?

The maximum distributable amount is a restriction mechanism in European Union capital law that limits distributions when a bank fails to meet a combined buffer requirement built on top of minimum capital ratios. Article 141 of the Capital Requirements Directive, Directive 2013/36/EU, sets the framework, and a calculation based on distributable profits and on the quartile of buffer consumption determines how much a bank may pay in total across dividends on common shares, variable remuneration and coupons on Additional Tier 1 instruments. A bank inside the restriction zone must prepare a capital conservation plan for supervisor review. The mechanism operates before a contractual capital trigger is reached and can therefore stop a coupon at a Common Equity Tier 1 ratio well above 5.125 percent, depending on risk-weighted assets, buffers and available profits in the calculation period.

Q: What should an AT1 prospectus disclose about write-down risk?

A prospectus governed by securities law in the European Union, the United Kingdom, India or another listing jurisdiction is required to disclose material risks in a form that allows an investor to understand the nature of loss absorption for the specific issue. Disclosure for a write-down instrument covers the trigger level and definition, whether a write-down is permanent or temporary, any write-up conditions, the authority empowered to declare a point of non-viability, ranking in liquidation and in resolution, coupon cancellation discretion, buffer restrictions on distributions, tax treatment statements, selling restrictions and the absence of a maturity date. Risk sections in prospectuses filed for issues by large banks also describe litigation risk, market liquidity risk and the possibility that a resolution authority exercises statutory powers beyond contract terms under the law named in the offer document.