Statistical measures such as value at risk describe losses in normal conditions. But the losses that threaten organisations usually come from rare, severe events — a deep recession, a currency crisis, a pandemic, a sudden loss of funding. Stress testing asks directly: what would happen to us if a severe but plausible event occurred? Since the 2008 financial crisis, stress tests have become a core tool for banks, regulators, insurers and companies to understand vulnerabilities and plan responses.
What you will learn
- Distinguish sensitivity tests, scenario tests and reverse stress tests
- Design a severe but plausible scenario and translate it into losses
- Calculate the effect of a stress scenario on a bank’s capital ratio
- Use stress test results in decisions and capital planning
Types of stress test
| Type | What it does | Example |
|---|---|---|
| Sensitivity test | Changes one risk factor, often by a large amount | Interest rates rise by 2 percentage points |
| Historical scenario | Replays a past crisis | The 2008 global financial crisis; a past currency devaluation |
| Hypothetical scenario | Combines several shocks into a coherent, forward-looking story | Recession with rising unemployment, falling property prices and higher rates |
| Reverse stress test | Starts from failure and asks what would cause it | What loss would push capital below the minimum? |
Stress tests in risk management build on the same ideas as scenario and sensitivity analysis in financial models, but focus on severe, adverse outcomes.
Designing a scenario
A good scenario is severe but plausible, internally consistent and relevant to the organisation’s specific vulnerabilities. Steps:
- Identify vulnerabilities — concentrations, funding dependence, key markets.
- Write a narrative — what happens, and why.
- Set macroeconomic and market variables — GDP, unemployment, interest rates, exchange rates, property and share prices — over the stress horizon (often one to three years).
- Translate into risk parameters — higher default probabilities, lower collateral values, market losses, deposit outflows.
- Project the impact on profit, capital and liquidity, including second-round effects.
- Consider management actions that could realistically be taken.
Worked example: a bank capital stress test
Worked example: one-year severe recession
A bank has capital of 100 and risk-weighted assets (RWA) of 800 — a capital ratio of 12.5%. The regulatory minimum including buffers is 10%. The scenario: a severe recession with rising defaults and a 2-percentage-point rise in interest rates (illustrative figures).
| Item | Effect on capital |
|---|---|
| Starting capital | 100 |
| Credit losses as non-performing loans rise | −30 |
| Losses on a 200 bond portfolio (duration 4 × 2% rate rise) | −16 |
| Pre-provision profit earned during the year | +15 |
| Capital after stress | 69 |
With lending reduced, RWA falls to 760. Stressed capital ratio = 69 ÷ 760 = 9.08% — below the 10% minimum.
The test reveals a vulnerability. Responses might include retaining earnings instead of paying dividends, raising capital, reducing the bond portfolio’s duration, or tightening lending to riskier sectors — see capital adequacy and non-performing loans.
Reverse stress testing
Instead of asking "what happens if…", a reverse stress test asks "what would break us?"
In the example, the ratio falls to 10% when capital reaches 10% × 760 = 76. That means a net loss of 24 — or gross losses of 39 after the 15 of pre-provision profit. Management can then ask how plausible a 39 loss is: for instance, how much would non-performing loans need to rise, or rates move, to cause it? Reverse stress tests often uncover combinations of events that standard scenarios miss.
Liquidity stress tests
Liquidity stress tests project cash inflows and outflows under scenarios such as large deposit withdrawals, loss of wholesale funding or collateral calls, and measure how long the organisation can survive — its survival horizon. Regulatory measures such as the LCR embed a standard 30-day stress — see liquidity risk.
Supervisory stress tests
Many central banks and supervisors run system-wide stress tests, applying common scenarios to major banks and publishing results. Bangladesh Bank requires banks to conduct regular stress tests covering credit, market and liquidity risks, and the Basel Committee’s Stress Testing Principles (2018) set international expectations for governance, scenarios, methods and use of results.
Stress testing for companies and individuals
- Companies test cash flows and covenants against falls in sales, rises in costs and interest rates, or the loss of a major customer — see budgeting and cash flow models.
- Individuals can ask: could I pay my loan if rates rose 3 points? How long could I cover essentials if my income stopped? The answers guide emergency fund size and borrowing decisions.
Using the results
Stress tests matter only if they influence decisions:
- Capital and liquidity planning — setting buffers above minimums.
- Risk appetite — limits defined in terms of stressed outcomes — see risk appetite.
- Contingency and recovery plans — actions prepared in advance.
- Strategy and pricing — avoiding concentrations that perform badly under stress.
- Board discussion of vulnerabilities and trade-offs.
Limitations
- Scenarios rely on judgement and may miss the next crisis.
- Models estimated on normal data can underestimate stressed relationships.
- Second-round effects — fire sales, contagion, feedback between the financial system and the economy — are hard to model.
- Management actions may be less effective in a real crisis than assumed.
Using several scenarios, reverse stress tests and expert challenge helps reduce these weaknesses.
Common mistakes
- Choosing scenarios too mild to be informative.
- Testing risks one at a time when they would occur together.
- Assuming optimistic management actions.
- Treating stress testing as a regulatory exercise disconnected from decisions.
- Ignoring liquidity effects alongside capital effects.
Key terms
- Stress test — an assessment of the impact of a severe but plausible event.
- Sensitivity test — a stress on a single risk factor.
- Scenario — a coherent set of assumptions about adverse developments.
- Reverse stress test — identifying events that would cause failure.
- Survival horizon — how long an organisation can meet obligations under liquidity stress.
Frequently asked questions
How severe should a stress scenario be?
Severe enough to be informative but plausible — often comparable to or worse than the worst historical episodes relevant to the business.
What is the difference between stress testing and VaR?
VaR estimates losses in normal conditions at a confidence level; stress tests examine specific extreme scenarios outside that range.
How often should stress tests be run?
Banks typically run comprehensive tests at least annually and more targeted tests more frequently, or when conditions change.
Key takeaways
- Stress tests examine severe but plausible events using sensitivity, historical, hypothetical and reverse tests.
- Scenarios are translated into losses, capital and liquidity effects over a set horizon.
- Reverse stress tests identify what would cause failure.
- Results should drive capital, liquidity, appetite, contingency planning and strategy.
Explore this in ARBIMETRA LAB
Apply these ideas to your own data. Files are analysed in your browser and are not uploaded.
- Conditional VaR (CVaR): Expected loss beyond VaR.
- Sensitivity & Scenario Analysis: One- and two-way tests, tornado chart.
