Skip to content
CyclepediaThe trading encyclopedia
ITENESPT-BRFRDE
Learning path Gold Professional operator

Stress testing

A stress test evaluates the effects of stated adverse conditions over a defined scope and horizon. It includes sensitivities, historical or hypothetical scenarios and reverse stress; it is neither a forecast nor proof of safety.

On this page

Who this entry is for — Anyone who wants to expose vulnerabilities in a portfolio, strategy or process when ordinary data and relationships are not enough.

A stress test applies defined adverse conditions to a scope and estimates their effects on relevant quantities: value and P&L, capital, margin, collateral, liquidity, funding, limits or operational continuity. The conditions may come from historical events, future hypotheses or a breaking outcome explored in reverse.

The result is conditional: it describes what that stress design produces under those assumptions. It does not automatically assign a probability to the scenario and is not a forecast.

In plain terms — A stress test does not ask “what will happen?” It asks “where are we vulnerable if these conditions occur?”.

From question to scenario—not from forecast to certainty A useful test exposes scope, shocks, transmission, results and limitations From question to scenario—not from forecast to certainty A useful test exposes scope, shocks, transmission, results and limitations A useful test exposes scope, shocks, transmission, results and limitations 1 · OBJECTIVE Scope, horizon, metric anddecision to be informed 2 · DESIGN Sensitivity, historical orhypothetical scenario, revers… 3 · TRANSMISSION Coherent factors,dependencies, non-linearity,… 4 · RESULTS Losses, resources,concentrations,… NOT A FORECAST Passing a test does not prove safety; failing it does not automaticallyassign a probability to the scenario. Cyclepedia · source-checked visual explainer
An interpretable test exposes its objective, design, transmission, results and limitations.

A family of methods

The Basel Committee places several methods inside stress testing:

Method Starting point Typical question
Sensitivity analysis One or a few changed factors How does the result respond to this input?
Historical scenario Movements observed in a past episode, applied to the current scope How would today's system respond to a similar configuration?
Hypothetical scenario A narrative and coherent paths built for the purpose Which vulnerabilities emerge from this combination?
Reverse stress test A predefined outcome such as illiquidity or a threshold breach Which combinations could lead to that outcome?

Scenario analysis is therefore one methodology in the family, not merely a list of playbook actions.

Building an interpretable test

  1. Objective and decision — State why the test is run and which decision it should inform.
  2. Scope, horizon and metrics — Define positions, entities, currency, date, resources and time interval.
  3. Material risks — Include relevant market factors, non-linearity, concentrations, counterparty, liquidity, funding and operations.
  4. Stress design — Disclose severity, shock sources, paths and consistency among variables.
  5. Transmission — Expose pricing, dependencies, counterparty actions, second-round effects and behavioural assumptions.
  6. Results and uncertainty — Separate losses, resources, vulnerabilities, assumptions, excluded risks and sensitivity of results.
  7. Governance and use — Document approval, independent review, challenge, escalation and feasibility of potential actions.

A static shock can produce an instantaneous loss, but it does not by itself determine maximum drawdown or recovery time. Those quantities require a path, a revaluation rule and assumptions about actions over time.

Severity and frequency depend on purpose

“Index −20%, spread ×2, slippage ×3” can be an illustration, not universal calibration. A useful scenario must be relevant to the scope, severe enough to expose vulnerabilities and internally consistent. History can inform design without constraining it: emerging risks may require combinations never jointly observed.

There is no single valid frequency either. Basel principles link frequency to the objective, scope, size and complexity of the institution and changes in the environment. Sector rules can set specific schedules — for example, the SEC requires at least weekly stress testing for certain funds subject to Rule 18f-4 — but that requirement does not become a general rule for every trader or portfolio.

Reading the result

  • Passing a test does not prove safety: other scenarios, models or omitted risks can produce different outcomes.
  • Failing a test does not by itself estimate scenario probability: it identifies a conditional vulnerability to investigate.
  • A mitigating action counts only if it is feasible under stress; selling, hedging or raising funding may become slower or more costly when needed.
  • The transmission model introduces model risk and requires review and challenge.

Typical mistake — Looking for a pass/fail traffic light without reading coverage, assumptions and vulnerabilities. The test's value is in diagnosis and traceable use of results.

Sources

Suggest an edit My proposals