Learning risk measures is like learning chess openings. Memorizing the moves for Value at Risk (VaR) is easy, but understanding why it fails against an unexpected threat, the kind Expected Shortfall (ES) is designed to see, is what separates a passing candidate from the 51% who fail.
Measures of financial risk for the FRM exam focus on Value at Risk (VaR), Expected Shortfall (ES), coherent risk measures, and three simulation methods: parametric, historical, and Monte Carlo. Mastering these concepts from Domain 4 is about judging which tool to use for a given scenario, not just calculating a number.
Key facts
- Official Body
- Global Association of Risk Professionals (GARP)
- Curriculum Location
- Domain 4: Valuation and Risk Models (Chapters 10 & 11)
- Domain 4 Weighting
- 30% of the FRM Part I exam
- Key Concepts
- VaR, Expected Shortfall (ES), Coherent Risk Measures, Parametric Simulation, Historical Simulation, Monte Carlo Simulation
- May 2023 Pass Rate (Part I)
- 49%
- Recommended Study Hours
- 200-250 hours for Part I
GARP reports the official FRM Part I pass rate for the May 2023 exam administration was 49% (GARP).
What is the Difference Between VaR and ES?
The core distinction is that VaR tells you a boundary, while ES describes the severity of what lies beyond that boundary. Value at Risk (VaR) is the maximum loss you can expect over a given time horizon at a specific confidence level (e.g., 99%). A one-day 99% VaR of $1 million means there is only a 1% chance of losing more than $1 million. It answers: "How bad can things get in a normal bad day?"
Studying for FRM Part 1? Benchmark your score in 5 minutes.
Get an instant weak-spot assessment and a custom 12-week study plan PDF generated for your exam window.
Expected Shortfall (ES), also called Conditional VaR (CVaR), answers a more critical question: "When a bad day happens, how bad is it on average?" It calculates the average of all losses in the tail of the distribution beyond the VaR threshold. This single change makes ES a "coherent" risk measure, which is a requirement for robust risk aggregation. Try VoraPrep's free FRM practice questions to see how this distinction is tested.
A risk measure is coherent if it meets four criteria. To remember them, use the mnemonic S.T.M.P. (some people prefer S.H.M.T.).
- Subadditivity: Risk(A+B) ≤ Risk(A) + Risk(B). The risk of a combined portfolio should not be greater than the sum of its parts. This reflects diversification. VaR fails this test.
- Translational Invariance: Risk(Portfolio + Cash) = Risk(Portfolio) − Cash. Adding cash to a portfolio reduces its risk by that amount.
- Monotonicity: If Portfolio A always performs worse than Portfolio B, then Risk(A) ≥ Risk(B).
- Positive Homogeneity: Risk(k × Portfolio) = k × Risk(Portfolio). Doubling the size of a portfolio doubles its risk.
Because VaR fails subadditivity, it can discourage diversification. ES passes all four tests, which is why regulators, especially under Basel's FRTB framework, have pushed for its adoption.
Choosing Your Method: Parametric vs. Historical vs. Monte Carlo
The FRM exam requires you to know not just what VaR and ES are, but how to calculate them using different models. Your choice of model depends on the data available and the nature of the portfolio assets.
| Method | How It Works | Best For | Major Weakness |
|---|---|---|---|
| Parametric | Assumes returns follow a known distribution (e.g., normal) and uses statistical parameters (mean, standard deviation) to calculate risk. | Simple, linear portfolios (stocks, bonds) where returns are reasonably normal. | Fails to capture "fat tails" or skewness; bad for portfolios with options. |
| Historical | Uses a window of past returns to simulate the future. Ranks historical returns and finds the loss at the desired percentile. | Portfolios with non-linear assets (options) or non-normal returns. Captures real-world fat tails. | Assumes the past is representative of the future; slow to react to new market regimes. |
| Monte Carlo | Generates thousands of random future return paths based on specified models and parameters. Creates a simulated P&L distribution. | Complex, non-linear portfolios. Allows for stress testing and modeling many risk factors. | Computationally intensive and highly dependent on the accuracy of the underlying pricing models. |
A classic exam question will describe a portfolio (e.g., one containing exotic options) and ask for the most appropriate simulation method. If you see non-linear payoffs, think Historical or Monte Carlo, not Parametric.
Calculate Your FRM Part 1 Score & Study Hours by Domain
Adjust your practice scores across official FRM Part 1 blueprint domains to predict your pass probability and generate a domain-weighted study schedule.
Worked Example: Calculating VaR and ES Under Exam Pressure
You will be asked to perform these calculations. The key is to be methodical and avoid the common traps related to percentile ranking.
Using the historical simulation method, calculate the 1-day 95% VaR and 95% ES.
Solution Path:Step 1: Identify the VaR Cutoff
We have 100 observations. A 95% confidence level corresponds to the worst 5% of outcomes (100% - 95% = 5%). The cutoff point is the 5th worst observation (100 days × 5% = 5th day).Step 2: Calculate the 95% VaR
The 5th worst P&L observation is -$2.5 million. The VaR is expressed as a positive number representing the loss. Therefore, the 1-day 95% VaR is $2.5 million.Step 3: Calculate the 95% ES
The 95% ES is the average of all losses in the worst 5% of cases. This means we average the 5 worst observations. Average = ($4.5 + $3.8 + $3.2 + $2.9 + $2.5) ÷ 5 Average = $16.9 ÷ 5 = $3.38 million. Therefore, the 1-day 95% ES is $3.38 million.Your 7-Day Sprint to Master Financial Risk Measures
Don't just read about these concepts. Use this focused plan to build deep, exam-ready knowledge in one week.
Day 1: Lock Down the Core Concepts- Action: Read GARP Chapter 10. Write down the definitions of VaR and ES in your own words. Focus on the intuition: VaR is a threshold, ES is the average beyond it.
- Checkpoint: Explain to a colleague (or yourself in the mirror) why ES is coherent and VaR is not. You must use the word "subadditivity."
- Action: Work through the formulas for calculating VaR from mean and standard deviation. Practice scaling a 1-day VaR to a 10-day VaR using the square-root-of-time rule.
- Checkpoint: Complete 5 numerical problems on parametric VaR. VoraPrep's adaptive question bank can isolate these for you.
- Action: Use the worked example in this article as a template. Create your own small dataset (10 numbers) and manually calculate historical VaR and ES. Then, read about the process for Monte Carlo.
- Checkpoint: Write one sentence explaining when you would choose Historical over Parametric, and one sentence for when you'd choose Monte Carlo over Historical.
- Action: Revisit the S.T.M.P. mnemonic. For each of the four properties, think of a simple two-asset portfolio example that illustrates the principle.
- Checkpoint: Do a flashcard drill on the properties of coherent risk measures until you can recall all four instantly.
- Action: Focus on qualitative questions. Find practice problems that describe a situation and ask for the best risk measure or the most significant model limitation.
- Checkpoint: Complete 10 scenario-based questions from the VoraPrep FRM course, focusing on the "why" explanations for each answer.
- Action: Take a 20-question timed quiz focused exclusively on Domain 4, Chapters 10 and 11. Simulate exam pressure.
- Checkpoint: For every question you get wrong, write down the specific concept you misunderstood. Use Vory, your 24/7 tutor, to get a step-by-step explanation until it clicks.
- Action: Review your notes and incorrect answers from the week. Re-read the key takeaways from this article.
- Checkpoint: You should now be able to create a one-page "cheat sheet" summarizing VaR, ES, coherence, and the three simulation methods from memory.