# Stress testing and the risk appetite framework in action
Picture a mid-sized US bank on the last day of a good year. Common Equity Tier 1 (CET1) ratio, the core measure of loss-absorbing capital against risk-weighted assets, sits at a comfortable 12.5%. Now a supervisor hands the bank a scenario: unemployment jumps to 10%, house prices fall 30%, equities drop 55%, and corporate spreads blow out. Twelve months later, on paper, that same bank's CET1 has fallen to 8.1%.
Nothing real happened. The bank just ran a stress test. And that gap between 12.5% and 8.1% is the whole point of this lesson.
A stress test projects a bank's balance sheet, losses, and capital under a hypothetical adverse scenario. It answers one question: if the world turns ugly, does the bank still have enough capital to keep lending and stay solvent?
Two supervisory regimes dominate.
Both use a baseline scenario (expected path) and a severely adverse scenario (a deep recession). Banks project nine quarters forward.
A scenario is just a set of macro paths. A CCAR-style severely adverse scenario has historically included variables like:
The bank translates these macro paths into losses on every portfolio.
Let us run a simplified version. CET1 ratio is:
CET1 ratio = CET1 capital / Risk-Weighted Assets (RWA)
Risk-weighted assets weight each exposure by riskiness. A government bond might carry a 0% weight; an unsecured corporate loan, 100%.
Take our bank:
Now apply the severely adverse scenario over nine quarters:
1. Credit losses. Loan defaults spike. Say projected loan losses net of some reserves reduce capital by $7 billion.
2. Trading and counterparty losses. The market shock hits the trading book: $2 billion.
3. Pre-provision net revenue (PPNR). The bank still earns money (net interest income plus fees minus expenses), but stressed revenue is thin. Say PPNR adds back $4 billion over the period.
4. Dividends. Assume the bank pays $1 billion in dividends it cannot skip in the projection.
Stressed CET1 capital:
25, 7, 2 + 4, 1 = $19 billion
Meanwhile RWA often rises in a downturn as risk weights climb. Say RWA grows to $210 billion.
Stressed CET1 ratio = 19 / 210 = 9.05%
The bank fell from 12.5% to about 9%. The minimum CET1 requirement is 4.5%, but with buffers the effective bar is higher (often 7% or more once the capital conservation buffer and any SCB are stacked). The drawdown here is roughly 3.5 percentage points. That number becomes the raw material for the Stress Capital Buffer.
Under the US framework, the SCB is broadly the maximum decline in the CET1 ratio during the stress test, plus four quarters of planned dividends, with a floor of 2.5%. A bigger drawdown means a bigger required buffer, which means less capital free to lend or return to shareholders.
That is the lever. Stress testing is not an academic exercise. It sets how much capital a bank must hold.
A stress test tells the board how bad things could get. The Risk Appetite Framework (RAF) decides how much of that badness the bank is willing to tolerate, in advance, and enforces it through hard limits.
The centerpiece is the Risk Appetite Statement (RAS): a board-approved document stating the level and types of risk the bank will accept to hit its strategy. The Financial Stability Board's principles shaped how banks structure these.
The RAS is deliberately high-level. The magic is the cascade down to concrete limits.
Top of the house (board RAS): "We will maintain a CET1 ratio of at least 11% under our internal severely adverse scenario."
Translation step: If the stress test shows a 3.5 point drawdown, then to end above 11% the bank must start above roughly 14.5%. That gap drives everything below.
Cascade to business limits:
Here is the mechanism managers care about. The trader has a VaR limit. The lending officer has a concentration cap. Each limit traces back, through the cascade, to a board sentence about surviving a recession. When a desk breaches VaR, it is not just a policy violation; it is chipping into the buffer the board promised regulators.
Breaches trigger escalation: to the business head, then the Chief Risk Officer, then the board risk committee. Persistent breaches force position reductions.
Knowledge check
1. In the opening example, a bank's CET1 ratio drops from 12.5% to 8.1% during a stress test even though 'nothing real happened.' What does this gap primarily represent?
2. What core question is a stress test fundamentally designed to answer?
3. Why do supervisors use both a 'baseline' and a 'severely adverse' scenario rather than only one?
4. Select ALL correct answers about how US and European stress-testing regimes are structured.
Select all the correct answers.
5. Select ALL correct answers about what defines a stress-test scenario.
Select all the correct answers.
If you are assessing a bank (as an investor, counterparty, or acquirer), stress testing and the RAF are prime diligence targets.
Check the results, not just the ratio. A 12% CET1 today means little if the stressed CET1 is barely above minimum. Read the drawdown. US firms disclose SCB values; a rising SCB signals the Fed sees more risk.
Check for consistency. Does the RAS match actual behavior? If the statement says "low appetite for commercial real estate" but CRE is 300% of capital, the framework is decorative.
Check limit utilization. Ask how often desks run near their limits and how many breaches occurred last year. Chronic breaches signal either a broken RAF or a mispriced business.
Check the qualitative side. The Fed can fail a bank on qualitative grounds (weak modeling, poor governance) even if the numbers pass. In Europe, the ECB's Supervisory Review and Evaluation Process (SREP) can impose extra capital (a "Pillar 2" add-on) based on judgment about a bank's risk management.
Given disclosed figures, sanity-check the arithmetic:
Starting CET1 ratio 12.5%
Minus stressed drawdown -3.5 pts
= Stressed minimum = 9.0%
Regulatory minimum + buffer 7.0%
= Headroom = 2.0 ptsTwo points of headroom on $200 billion of RWA is about $4 billion of surplus capital. Thin headroom means the bank has little room to grow lending or absorb a surprise. Fat headroom may mean the bank is over-capitalized and under-earning. Either way, the number tells a story.