Market Risk Measurement and Management
Market risk management connects sensitivities, loss distributions, stress scenarios and model checks to limits and hedges. Use VaR for a threshold, ES for tail severity, and stress tests for specific vulnerabilities; none is a complete risk description alone.
The topic at a glance
| Concept | How it works | Watch for |
|---|---|---|
| Measure | Report factor exposure, horizon, confidence and portfolio scope. | Comparing risk numbers with different horizons. |
| Challenge | Use backtesting and stress analysis to test different weaknesses. | Treating a passed backtest as evidence that all tail risks are covered. |
| Act | Connect limits to hedges, exposure reductions and escalation. | Changing a model merely to make a limit breach disappear. |
01Measure
How it works: Report factor exposure, horizon, confidence and portfolio scope.
Watch for: Comparing risk numbers with different horizons.
02Challenge
How it works: Use backtesting and stress analysis to test different weaknesses.
Watch for: Treating a passed backtest as evidence that all tail risks are covered.
03Act
How it works: Connect limits to hedges, exposure reductions and escalation.
Watch for: Changing a model merely to make a limit breach disappear.
Original worked example
A limit decision needs more than VaR
A desk reports 99% one-day VaR of 2 million within a 2.5 million limit. A specified curve-and-spread stress loses 12 million, above a separate 10 million stress limit.
- STEP 1
The VaR limit has 0.5 million headroom.
- STEP 2
The stress limit is exceeded by 2 million.
- STEP 3
Check the concentrated factors and an executable hedge or reduction for the stressed exposure.
Being within the VaR limit does not authorize the stress-limit breach. Escalate using the relevant limit and exposure.
Illustrative assumptions and calculations for learning. Not observed market data or an official exam item.
How the ideas fit together
Before you start: Part I sensitivities, VaR, expected shortfall and stress scenarios.
Use complementary limits for different vulnerabilities
A sensitivity limit constrains a local exposure; a VaR limit constrains a modelled quantile; a stress limit constrains loss under a specified scenario. Passing one does not override another. Compare outputs using their stated horizon, portfolio scope and assumptions. A concentrated position may look acceptable in quiet historical data while breaching a forward-looking stress limit.
Turn model diagnostics into management actions
Backtesting investigates how forecasts compare with subsequent outcomes. Stress analysis tests specified adverse states. Position and data reconciliation check that the model describes the actual book. A response might reduce exposure, improve factor coverage or restrict use of an unreliable model. Identify the weakness before choosing the remedy.
Do not confuse regulatory and internal measures
An institution’s management limits can use different horizons, portfolios or models from a regulatory capital framework. Label the measure before applying a rule. The Basel market-risk framework provides regulatory context for expected shortfall, but a simplified classroom VaR calculation is not a capital calculation or proof of regulatory compliance.
Work through the topic
Choose market-risk measures for a specific decision, connect them to valuation and limits, and understand why a satisfactory headline number still requires investigation.
Define the portfolio, drivers and decision horizon
Market risk concerns changes in value associated with movements in market prices and risk factors. The practical starting point is a reconciled position inventory and a valuation framework. Rates, credit spreads, equity prices, currencies, commodity prices and implied volatilities can interact. A desk name does not provide a complete factor map, and a hedge in one factor can introduce exposure to another.
State the decision before choosing a measure. A trader assessing an incremental hedge needs local sensitivities. A committee setting a loss tolerance needs portfolio scenarios and aggregation. A funding team needs the cash consequences of margining. The measures may use the same positions but different horizons and assumptions. A daily risk report should make these differences explicit rather than place incomparable figures beside one another.
Regulatory market-risk frameworks add specific definitions, eligibility criteria and calculation requirements. The Basel reference supplies that context, including the role of expected shortfall in the internal models approach. A classroom ES calculation is not, by itself, a compliant regulatory capital calculation. Applicable rules, trading-book boundaries and local implementation require separate verification.
Use sensitivities, tail measures and scenarios together
A sensitivity asks how value changes for a small move in a selected factor while other assumptions are held fixed. VaR estimates a loss quantile. Expected shortfall measures an average in a defined upper tail. A stress result estimates the consequence of specified adverse conditions. They answer different questions, so selecting a single favorite measure leaves predictable blind spots.
Suppose a desk’s reported 99% one-day VaR falls from 2.0 million to 1.5 million while its loss under a funding-disruption scenario rises from 6 million to 9 million. The figures are not necessarily contradictory. The estimation sample may have become quieter while the portfolio accumulated an exposure that performs poorly under the selected stress. Investigate positions, volatility, dependence and valuation changes before interpreting the lower VaR as improvement.
A useful decomposition holds one input set fixed at a time. Recalculate yesterday’s portfolio with today’s model and today’s portfolio with yesterday’s model, where feasible. Review factor and desk contributions, recognizing that allocated contributions depend on the method used. The aim is a causal explanation of the report’s movement, not simply an attractive total.
| Question | Measure to inspect | Main limitation |
|---|---|---|
| What does a small rate move do? | DV01 or key-rate sensitivities | Local approximation and factor definition |
| Where is a modeled loss threshold? | VaR at a stated horizon and confidence | No description of losses beyond the threshold |
| How severe is the modeled tail? | Expected shortfall | Tail-data and model uncertainty |
| What if funding and markets deteriorate together? | Integrated stress scenario | Severity is not an assigned probability |
Match the market-risk question to the measure
- What does a small rate move do?
- Measure to inspect: DV01 or key-rate sensitivities
- Main limitation: Local approximation and factor definition
- Where is a modeled loss threshold?
- Measure to inspect: VaR at a stated horizon and confidence
- Main limitation: No description of losses beyond the threshold
- How severe is the modeled tail?
- Measure to inspect: Expected shortfall
- Main limitation: Tail-data and model uncertainty
- What if funding and markets deteriorate together?
- Measure to inspect: Integrated stress scenario
- Main limitation: Severity is not an assigned probability
Turn measurement into limits and escalation
A limit framework translates appetite into decisions about positions, sensitivities, concentration and scenario losses. A desk can remain within its VaR limit and still breach a concentration or stress limit. Specify who monitors each measure, how often it is refreshed and who can approve an exception. An exception process should preserve the evidence and the time-bound remediation plan.
For an original case, a desk has a stress-loss limit of 8 million and a new result of 9 million. A proposed hedge reduces the modeled stress loss to 7.5 million but requires 1 million of initial cash collateral. The market-risk effect may resolve the limit breach, while treasury still needs to establish that the collateral can be funded. Approval should consider both consequences and the residual exposure if the hedge cannot be executed at the assumed price.
Validation then checks whether the model remains suitable for its use. Backtesting compares forecasts with consistent realized outcomes; sensitivity analysis challenges inputs; independent review assesses conceptual and implementation weaknesses. An acceptable test result supports a particular aspect of the model. It cannot establish that every material market driver is present or that liquidation will work under all conditions.
Why study market-risk management beyond the VaR formula?
The formula assumes positions, data, valuation, dependence and the loss definition have already been chosen correctly. Management adds the decisions: which exposures are acceptable, which result triggers action, whether a hedge can be funded and who owns the response.
Can a hedge increase another risk measure?
Yes. It may reduce a market sensitivity while increasing counterparty exposure, margin needs, basis risk or concentration. Evaluate the combined portfolio and the operational ability to maintain the hedge rather than its intended benefit in isolation.
Try a changed assumption
VaR is within its limit, but the stated stress loss exceeds its approved limit by 2 million. Can the desk report that all risk limits are satisfied?
Show answer and reasoning
No. The stress constraint is breached. A favorable VaR result does not cancel a separate approved constraint.
Original learning check. No email or account needed.
Does a larger ES always mean the model is worse?
No. It can reflect a riskier portfolio or a more severe estimated tail. Model quality concerns calibration, assumptions and fitness for use, not simply the size of the output.
Sources and scope
Links checked 2026-10-10. GARP establishes exam scope; the technical references support the methods identified below. These are selected study notes and original examples, not complete curriculum coverage or official exam questions.
- GARP FRM study materials and 2026 learning objectives
Exam-domain scope and the official route to the annual learning objectives. This page does not certify coverage of every reading.
- Basel Committee: minimum capital requirements for market risk (2019)
Regulatory context for expected shortfall and market-risk model assessment. Local implementation can differ.
- MIT: Value at Risk Models, lecture 7
Loss quantiles, estimation methods and the limitations of VaR.
- Basel Committee: corporate governance principles
Risk appetite, responsibilities of the board and management, independent risk oversight and internal audit.
Prepared by MiloPrep editorial team. MiloPrep is independent of GARP. References do not imply endorsement or independent review of these notes.