The Math Behind the Thin Cushion.

The number supervisors care most about is the CET1 ratio. CET1 stands for Common Equity Tier 1, the strictest definition of capital there is: ordinary shares and retained earnings, the money that belongs to shareholders and absorbs losses first and without argument.

That capital is measured against risk-weighted assets, usually shortened to RWA. Not every euro a bank lends carries the same risk. A loan to a highly rated government is not the same bet as an unsecured loan to a mid-sized company. A rule that treated them identically would push banks toward the riskier one. This is why each asset gets assigned a risk weight and the balance sheet is re-measured through that lens before the ratio is calculated.

Put numbers on it. Take a bank with $1,000M of assets and $80M of CET1 capital, and say the weights work out at 20% for highly rated government bonds and 100% for unsecured corporate loans. Held as government bonds, its risk-weighted assets are $200M and its CET1 ratio is 40%. Held as corporate loans, that same $80M produces a ratio of 8%.

Same bank, same capital, both correctly calculated. Which is why a ratio quoted with no reference to what the bank actually holds tells you very little. It’s a statement about the assets as much as the capital. It also means the useful follow-up, whenever a bank’s CET1 ratio improves, is whether capital went up or risk-weighted assets went down. Those are different events wearing the same number.

Basel III sets a CET1 minimum, and once you add the capital conservation buffer on top, the familiar headline figure is a little higher.

In the eurozone, the ECB reviews each bank it supervises through a process called the SREP, the Supervisory Review and Evaluation Process. It looks at that bank’s business model, governance, and risks, then sets a bank-specific add-on on top of the Basel minimum, known as the Pillar 2 Requirement. Other buffers can be layered on depending on the bank’s size and the state of the economy.

The requirement is personalised. Two eurozone banks reading the same rulebook can be working to different numbers and a ratio comfortable at one could be uncomfortably tight at the other.

What matters is the distance between the reported ratio and the number that bank has actually been told to hold. A bank sitting 40 basis points above its requirement and one sitting 400 basis points above it are in completely different situations, and both technically pass.

That distance has a direct consequence for anyone holding the shares. When a bank drops into its combined buffer, restrictions kick in on how much cash it’s allowed to lend. Dividends, share buybacks, and coupon payments on certain capital instruments all get capped. It’s called the Maximum Distributable Amount, the mechanism that links a bank’s capital position to whether shareholders get paid at all.

These get discussed as though they’re the same but they aren’t and separating them is crucial. Each has its own set of rules, because each fails differently.

A solvency problem is when the assets turn out to be worth less than everyone believed. Loans don’t get repaid, or securities lose value, and the losses absorb that thin equity cushion. The bank isn’t short of time. It’s short of value.

Solvency is policed by the capital ratios: CET1 against risk-weighted assets. Alongside it sits the leverage ratio, which measures capital against total exposure with no risk weighting at all. It exists precisely because risk weights depend partly on models, and models involve judgement, and judgement can be optimistic. The leverage ratio is the blunt instrument that doesn’t care how clever the modelling was.

A liquidity problem is when the assets are fine and the money still isn’t there. Depositors want their money faster than the loans mature, and the bank can’t turn a thirty-year mortgage into cash by Thursday afternoon. Nothing about the underlying value is wrong. The timing is.

Liquidity is policed by two others. The Liquidity Coverage Ratio, or LCR, asks whether a bank holds enough high-quality liquid assets to survive thirty days of a modelled stress scenario, with a minimum of 100%. The Net Stable Funding Ratio, or NSFR, asks a slower version of the same question over a full year: whether the long-dated assets on this balance sheet are backed by funding that’s actually stable enough to support them. Also a 100% minimum.

A capital ratio is a point-in-time photograph of value but says nothing about speed. It cannot tell you how quickly money can leave, how concentrated the depositors are, or whether the people funding this bank all know each other and read the same group chat.

Bank capital is deliberately built as a stack, and the order is the entire point. Everyone in it has agreed, at least in theory, to a position in the queue.

The instrument most people have never heard of is AT1. These are bonds that stop behaving like bonds the moment a bank’s capital falls below a pre-agreed level, at which point they can be written down or turned into shares. They pay a noticeably higher yield than ordinary bank debt, and the yield is not generosity. It’s the price of sitting that close to the front of the queue.

One thing worth noting is that this queue isn’t fixed by nature. It’s a legal arrangement, defined by each instrument’s documentation and by the rescuing country’s rules. And in some real cases, a troubled bank hasn’t actually paid out in the order investors expected.