Bail-in

The most significant institutional legacy of the 2008 global financial crisis is the collective reckoning with bail-outs. Following the Lehman Brothers failure — whose collapse the US government permitted before observing the resulting systemic shock — the US Treasury injected $85 billion into AIG, $45 billion into Citigroup, and $45 billion into Bank of America. The UK poured tens of billions of pounds into RBS and Lloyds. All of this was funded by taxpayers.

The problems with bail-outs are structural. First, requiring taxpayers to absorb losses from private financial institutions raises fundamental fairness objections. Second, when large banks expect an implicit government guarantee — the "Too Big to Fail" assurance — they have an incentive to take on excessive risk, knowing that the downside will be socialized. This moral hazard distorts risk-taking behavior throughout the financial system.

Bail-in is the regulatory response to these problems. When a bank faces insolvency, the first loss-bearers should be the bank's own shareholders and creditors — private capital that invested in the institution knowingly and was compensated for the risk it accepted. This principle was codified in the EU's BRRD (Bank Recovery and Resolution Directive, enacted 2014), the TLAC (Total Loss-Absorbing Capacity) requirements applicable to global systemically important banks (G-SIBs), and the MREL (Minimum Requirement for Own Funds and Eligible Liabilities) framework for European banks. All of these regulatory regimes were architecturally designed around the bail-in principle as their foundational logic.

When a bail-in is activated, losses are absorbed sequentially from the most subordinated layer of the capital structure upward. Understanding this sequence is essential for any investor in bank debt instruments.

The first loss-absorption layer is CET1 (common equity). Shareholders absorb losses first; equity value is extinguished before anything else. The second layer is AT1: once CET1 is fully wiped out — or if the regulator declares PONV — AT1 absorbs losses next, through equity conversion or principal write-down. The third layer is Tier 2 (subordinated bonds): if AT1 is insufficient, Tier 2 bondholders bear losses. The fourth layer includes Senior Non-Preferred bonds — a category of senior debt specifically designed to be eligible for bail-in for resolution purposes, sitting below ordinary senior unsecured in the hierarchy. Standard Senior (secured and unsecured) bonds and deposits are protected in theory and typically the last resort.

TLAC and MREL were introduced to ensure that banks maintain adequate minimum bail-in capacity at all times. TLAC applies to G-SIBs (Globally Systemically Important Banks — JPMorgan, HSBC, BNP Paribas, etc.) and requires them to hold loss-absorbing capacity of at least 18% of risk-weighted assets (from 2022 onward). MREL imposes comparable requirements on EU banks. Both frameworks achieve the same goal: by mandating a pre-positioned stack of bail-inable capital, they ensure that when a bank fails, an orderly restructuring is possible without taxpayer involvement. The bail-in mechanism is the machinery through which this theoretical pre-positioning is actually executed.

The Credit Suisse episode of March 2023 stands as one of the most consequential bail-in events in modern banking regulatory history — though notably, it was not a textbook bail-in in the purest sense, and that distinction is precisely what generated lasting controversy.

That a partial bail-in occurred is unambiguous. CHF 16 billion of CS AT1 bonds were written to zero. The core bail-in mechanism — private creditors (AT1 holders) absorbing losses rather than taxpayers — functioned as designed. The Swiss government-backed liquidity support (a central bank lending guarantee of over CHF 100 billion) served as a backstop deployed after the bail-in, not as a bail-out deployed instead of it.

The "abnormal" element was the sequencing. Under standard bail-in principles, equity should be extinguished before CoCo/AT1 absorbs losses. In the CS-UBS merger, shareholders received some residual value — UBS shares — while AT1 holders received zero. This inversion of the standard loss-absorption hierarchy triggered an immediate and significant reaction from international regulators. The ECB, EBA, and PRA collectively issued statements within hours, explicitly affirming that in their jurisdictions, equity would absorb losses before AT1 in any future resolution.

The institutional aftermath involved two main dimensions. First, prospectus language and jurisdiction-specific loss absorption sequencing came under fresh scrutiny. Switzerland maintained its legal framework permitting FINMA discretion, while debate continued about ensuring that the associated risks are communicated transparently to investors. Second, AT1 coupon spreads widened structurally across the market — investors repriced jurisdictional and PONV risk upward. The enduring lesson: the bail-in principle only functions as intended when investors genuinely understand, in advance, what the contractual terms and applicable law actually prescribe for every scenario. The label "bail-in compliant" is not a uniform standard — the details in each prospectus and each jurisdiction's statute book are what ultimately determine outcomes.