Market-driven scenarios (MDS) connect an economic narrative to consistent shocks across a portfolio's risk factors. Start with a clearly defined event and horizon. Specify alternative outcomes, identify the market variables most directly affected and estimate their shocks. Use a risk model to propagate those shocks to the remaining variables, assess plausibility and revalue the portfolio.
A historical replay applies an observed event to today's holdings. A hypothetical scenario can examine an event that has not occurred in the available sample. MDS adds statistical discipline to that hypothetical construction while retaining expert judgment about the event. Historical correlations are useful evidence, but a political or policy shock can change the relationships themselves.
The policy variables are the deliberately shocked variables, such as equity returns, a government yield or an exchange rate. Choose a small set that expresses the event without imposing contradictory information twice. A yield shock must specify its units and tenor. A currency shock must specify which exchange-rate quotation rises. Use a common time horizon for both shocks and covariance estimates.
Common mistakes
- Calling distance a probability. A distance of 3 across several variables is not a univariate three-standard-deviation tail probability.
- Misreading the correlation score. It measures how the shock combination fits correlations. It is not the number of standard deviations by which correlation itself has risen.
- Treating gross sales as cash. A $17 million sale with costs does not produce $17 million for redemptions.
Bottom line
- MDS links a defined event to direct shocks, propagated moves and portfolio losses.
- Scenario severity separates individual shock size from correlation consistency.
- Historical plausibility scores are conditional on the model, not event probabilities.
- Liquidity models need executable sizes, costs, settlement dates and uncertainty.
Exam shortcut
For scenario questions, identify the number of variables before converting distance to a scenario score. Keep shock magnitudes separate from their correlation pattern. For liquidity questions, build a dated cash calculation using net sale proceeds. For illiquid allocations, recompute weights after market moves, calls and spending instead of relying on the starting allocation.
The full lesson (about 5,512 words, 37 min read) adds 5 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
Browse all free FRM Part II lessons or jump into free FRM Part II practice questions.