Skip to content

FRM Exam Part II · Market-Driven Scenarios: An Approach for Plausible Scenario Construction

Stress Testing and the Role of Scenario Analysis

Updated 11 October 2026 · Fact-checked

Scenario analysis applies a defined set of market shocks to a portfolio and measures the profit or loss. It complements VaR because VaR is a statistical estimate that says little about losses beyond its confidence level. Historical scenarios replay past events; hypothetical ones are designed. Both can miss tail risk if poorly built.

Understand Stress Testing and the Role of Scenario Analysis

VaR gives a loss level that should be exceeded only with a small probability over a set horizon, for example 99% over 10 days. It rests on a statistical model and on past data. It tells you nothing about how bad the loss is once the threshold is breached.

Scenario analysis works differently. You pick a set of shocks to risk factors, such as equity indices falling 30%, credit spreads widening 300 basis points, and the USD strengthening. You then revalue the portfolio under those shocks. The result is a loss figure for a specific, nameable event. It does not carry a probability unless you assign one separately.

That is why the two tools are complements. VaR is good for day-to-day limits and comparison across desks. Scenarios cover what VaR misses: extreme moves, changes in correlation, liquidity dry-ups, and structural breaks that the data window does not contain. Scenarios also help management discuss risks in plain terms, such as 'what if 2008 happened again?'

There are two main types. Historical scenarios reapply the factor moves of a past event, for example the 1987 equity crash, the 1998 LTCM episode or the 2008 crisis. They are concrete and easy to explain, but they assume the future resembles the past and only cover shocks that have actually happened. Hypothetical scenarios are constructed by judgement about events that could happen, such as a sudden geopolitical shock or a sovereign default. They can be more forward looking, but they are subjective, and the risk is that the designer shocks only a few factors and ignores how the other factors would move.

The Market-Driven Scenarios chapter addresses this last weakness. A plausible scenario shocks a few key factors, then sets the remaining factors using their statistical relationships to those shocked factors. This keeps the scenario internally consistent and makes it more plausible than a set of arbitrary shocks. Even so, no scenario set can cover every possible tail event.

Key formulas to remember

Scenario P&L (linear approximation)
ΔP ≈ Σ (exposure_i × shock_i)
Add exposure times shock across risk factors. For options and other nonlinear positions, use full revaluation, not this approximation.
VaR definition
P(Loss > VaR) = 1 − confidence level
VaR is a threshold. It gives no information on the size of losses beyond it.
Scenario versus VaR
VaR = statistical, probability-based; scenario = event-based, usually no probability
Use this contrast to pick the right tool in a question.

How to solve Stress Testing and the Role of Scenario Analysis questions

Use this method for any question on scenario analysis, stress testing or their link to VaR.

  1. 1Identify what the question asks: a comparison, a limitation, a calculation or a design choice.
  2. 2Decide the scenario type: historical (past event replayed) or hypothetical (designed shocks).
  3. 3If a calculation is needed, list the exposure to each shocked risk factor and its shock, then multiply and add.
  4. 4Check whether the portfolio is nonlinear. If so, linear exposures understate or misstate the loss under large shocks.
  5. 5Ask whether unshocked factors are moved consistently, for example through correlations. If not, plausibility is weak.
  6. 6State the interpretation: the loss is for this event only and is not a probability-weighted worst case.
  7. 7Match the answer to the limitation: past dependence, subjectivity, missed factors or missing probabilities.

Quickest way: Three-question screen

When to use it: Use it for conceptual MCQs where two options sound alike.

  1. Does the option say VaR shows losses beyond the confidence level? If so, it is wrong.
  2. Does it say historical scenarios cover events that have never happened? If so, it is wrong.
  3. Does it say scenarios give a probability of loss by default? If so, it is wrong. Prefer the option that says scenarios complement VaR.

Common mistakes in Stress Testing and the Role of Scenario Analysis

  • Treating a scenario loss as a worst-case loss.

    A large number looks like a bound.

    Fix: Remember that a scenario is one chosen event. A worse event can exist.

  • Saying VaR tells you the average loss when VaR is exceeded.

    Students mix up VaR and expected shortfall.

    Fix: VaR is a threshold. Expected shortfall is the average loss beyond it.

  • Believing historical scenarios capture all tail risk.

    Past crises feel extreme enough to cover everything.

    Fix: They only include events that occurred. New structures and new risk factors may behave differently.

  • Shocking only one or two factors and leaving others unchanged.

    It keeps the arithmetic simple.

    Fix: Set other factors consistently, for example using their correlations with the shocked factors.

  • Applying linear exposures to options under large shocks.

    Delta works well for small moves.

    Fix: Use full revaluation, or add gamma terms, for large shocks.

Worked examples

Example 1

A portfolio holds USD 80 million of equities with beta 1 to the index and USD 40 million of corporate bonds with spread duration 5. Scenario: equities fall 25% and credit spreads widen by 2%. Estimate the scenario loss using linear exposures, ignoring other effects.

Show the solution
  1. Equity loss = 80 million × 1 × 25% = USD 20 million.
  2. Bond loss = 40 million × 5 × 2% = USD 4 million.
  3. Total loss = 20 + 4 = USD 24 million.

Answer: The scenario loss is about USD 24 million.

Example 2

A risk committee says its 99% one-day VaR is USD 5 million, so it does not need stress tests. Give the best response.

Show the solution
  1. VaR is a threshold that is exceeded about 1% of the time under the model.
  2. It does not describe how large the loss is beyond USD 5 million.
  3. It relies on a historical or modelled distribution that may not include crisis events or correlation shifts.
  4. Scenarios show losses for specific extreme events and support management discussion and action.

Answer: Stress tests are still needed, because VaR does not measure losses beyond its confidence level or events outside its data. Scenario analysis complements it.

Exam tips

  • Expect conceptual contrast questions: VaR versus scenarios, and historical versus hypothetical.
  • Read the wording on probability. Scenarios usually carry none unless stated.
  • For calculations, check whether duration or spread duration must be multiplied by the shock.
  • Link plausibility to consistent factor moves, which is the main theme of this chapter.
  • Prefer answers that say tools complement each other over answers that call one superior.

Practice questions from Market-Driven Scenarios: An Approach for Plausible Scenario Construction

Stress Testing and the Role of Scenario Analysis in other exams

The same ground in other exams, if you are preparing for more than one or want another angle on it.

Stress Testing and the Role of Scenario Analysis: frequently asked questions

Why use scenario analysis if I already have VaR?

VaR does not show losses beyond its confidence level and depends on past data. Scenarios test specific extreme events and structural shifts. Together they give a fuller picture.

What is the difference between historical and hypothetical scenarios?

Historical scenarios replay factor moves from a past event such as 2008. Hypothetical scenarios are designed for events that could happen but may not have occurred. The first is concrete but backward looking; the second is forward looking but subjective.

Do scenarios have probabilities?

Not by default. A scenario is an event-based loss estimate. Some frameworks attach judgemental probabilities, but that is an extra step.

What makes a scenario plausible?

The unshocked factors move in a way consistent with the shocked ones, usually based on their statistical relationships. This avoids unrealistic combinations of moves.