A bank's investment committee debates two trades. Trade A returns 18% on $100M of regulatory capital. Trade B returns 11% on $25M of regulatory capital. Headline return picks A. Risk-adjusted return on capital, the discipline forced by Basel and CCAR, picks B. The exam tests whether you can reframe headline numbers into capital-adjusted numbers and pick the trade that creates shareholder value rather than the one that looks bigger.
Til Schuermann's 2014 paper traces the modern stress-testing era from the Supervisory Capital Assessment Program (SCAP) of 2009 through the Comprehensive Capital Analysis and Review (CCAR) and Dodd-Frank Act Stress Tests (DFAST), with parallel work by the European Banking Authority. SCAP was an emergency response: a one-time test to determine which large U.S. banks needed to raise capital after the financial crisis. CCAR and DFAST are the institutionalized successors, run annually for large bank holding companies.
The methodologies differ in design choices but share a common shape. The supervisor publishes a baseline scenario, an adverse scenario, and a severely adverse scenario.
Common mistakes
- Treating stress testing as a worst-case scenario calculation. Stress testing asks what happens under a defined narrative; it does not produce a confidence-level number. VaR and stress testing answer different questions. Trap: a question gives stress test loss and asks about confidence: there is no confidence level associated with a stress scenario.
- Confusing risk capital, economic capital, and regulatory capital. Risk capital is the firm's belief about needs; economic capital is the model output for risk capital; regulatory capital is the legal minimum. The three numbers can differ materially.
- Forgetting the capital charge income in the RAROC numerator. The capital allocated to the activity earns the risk-free rate even if the activity does nothing. Forgetting to add it back understates RAROC and biases the firm against capital-intensive activities.
Bottom line
- Stress testing answers a different question than VaR: VaR asks the loss at a confidence level, stress testing asks the forward-looking impact if a defined scenario occurs (no confidence level attaches to a scenario).
- Reverse stress testing is expected by both the Fed and EBA: it asks what would break the firm and surfaces vulnerabilities forward testing misses.
- Scenario coherence is the hard part of design: random shock combinations are not credible economic states.
- Schuermann's evolution: SCAP (2009 one-time emergency) led to CCAR (annual quantitative plus qualitative) and DFAST (statutory annual); EBA tests share the design with different governance and disclosure.
Exam shortcut
When a question presents revenues, expected loss, expenses, taxes, and capital, the answer is almost always RAROC. Walk the components in order. Capital charge income is the most-forgotten line item; the test writer adds it as a distractor option and most candidates leave it out. When a question describes a stress scenario, ask first whether it is sensitivity (single factor), scenario (multiple factors), or reverse (outcome-driven).
The full lesson (about 3,208 words, 21 min read) adds 2 worked examples, all 6 common mistakes, a self-check, free in the app.
Learning objectives
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
- 13
- 14
- 15
- 16
- 17
- 18
- 19
- 20
- 21
- 22
- 23
- 24
Browse all free FRM Part II lessons or jump into free FRM Part II practice questions.