FRM Part 2 Market Risk: VaR, ES & Stress Testing
Why Market Risk Is the Make-or-Break Topic in FRM Part 2
If you're working through FRM Part 2, you already know the exam doesn't reward memorization. GARP isn't asking you to recite a formula — it's asking you to apply it under conditions you haven't seen before. Nowhere is that more punishing than in the market risk section.
Market risk accounts for a significant portion of FRM Part 2, and the conceptual demands are steep. You need to understand Value at Risk (VaR) at a level deep enough to critique it, know when Expected Shortfall (ES) is the right tool instead, and explain how stress testing fits into a real risk management framework — not just as a checkbox exercise.
This post breaks down each of those three pillars the way a seasoned risk manager would explain them to a junior analyst who needs to be exam-ready in weeks, not years.
---
Value at Risk: What GARP Actually Wants You to Know
VaR is the entry point for FRM Part 2 market risk. At the most basic level, you know the definition: VaR is the maximum loss not exceeded with a given confidence level over a specified time horizon.
But GARP tests at a much deeper layer. Here's what actually shows up:
The Three VaR Methods — And Their Failure Modes
Historical Simulation uses actual past returns to build a loss distribution. It makes no distributional assumptions, which is its strength. Its weakness? It's entirely backward-looking. If the last 500 days were calm, your VaR looks deceptively low — right before a regime change.
Parametric (Variance-Covariance) VaR assumes returns are normally distributed. It's fast and elegant, but the normal distribution systematically underestimates tail risk. Fat tails are real. The 2008 financial crisis was a parade of "25-sigma" events that the normal distribution said should happen once every universe.
Monte Carlo Simulation is the most flexible method — it can model non-linear instruments and non-normal distributions — but it's computationally intensive and only as good as the model assumptions driving the simulations.
On exam day, GARP loves questions that ask you to identify which method is being described, which limitation applies, or — critically — what a portfolio manager should do when a method fails. Know not just what each method does, but when it breaks.
Backtesting VaR
Backtesting is how you validate a VaR model. You compare predicted VaR breaches (exceedances) against actual losses. Under the Basel framework, banks are evaluated using traffic light zones based on the number of exceedances over 250 trading days.
- Green zone: 0–4 exceedances — model is acceptable
- Yellow zone: 5–9 exceedances — increased scrutiny
- Red zone: 10+ exceedances — model rejection, higher capital multiplier
This isn't trivia. It connects market risk measurement directly to regulatory capital — a thread GARP pulls on throughout Part 2.
---
Expected Shortfall: The Upgrade That Basel III Demanded
If you've been treating Expected Shortfall (ES) as just "VaR's cousin," that mindset will cost you points.
ES — also called Conditional VaR (CVaR) or the Expected Tail Loss — answers a question VaR deliberately ignores: Given that we've already exceeded the VaR threshold, how bad does it get on average?
Formally: ES is the expected loss in the tail beyond the VaR confidence level.
Why ES Matters in the FRM Context
VaR has a well-documented flaw: it tells you nothing about the shape or severity of the tail. A portfolio could have a 99% VaR of $10M and occasionally produce losses of $500M — VaR wouldn't differentiate that from a portfolio with losses clustered just above $10M.
ES fixes this by averaging over the tail. It's a coherent risk measure (it satisfies subadditivity, which VaR under certain conditions does not), making it theoretically superior for portfolio aggregation.
Under Basel III/FRTB (Fundamental Review of the Trading Book), regulators formally moved from 99% VaR to 97.5% ES as the standard for internal models. This is a tested fact — and more importantly, a tested concept. GARP will ask you to explain why ES was adopted, not just that it was.
Study tip: Build a mental model of ES as the answer to: "VaR tells me the door. ES tells me what's behind it."
---
Stress Testing: Risk Management Beyond the Model
VaR and ES are statistical tools — they rely on historical data or assumptions about distributions. Stress testing exists precisely because the most dangerous scenarios are often the ones that don't appear in your historical window.
Scenario Analysis vs. Sensitivity Analysis
These two terms are often used loosely but test differently:
- Sensitivity analysis asks: what happens to my portfolio if one risk factor moves by a fixed amount? (e.g., rates +100bps)
- Scenario analysis asks: what happens if a full macro scenario unfolds simultaneously? (e.g., 2008-style credit crisis, sovereign default, equity crash + liquidity freeze)
Scenario analysis is more operationally complex but more realistic — correlated risk factors move together in real crises.
Reverse Stress Testing
This is a concept GARP has elevated in recent exam cycles. Reverse stress testing flips the question: instead of asking "what loss does this scenario produce?", it asks "what scenario would cause a loss large enough to threaten viability?"
It's a tool for identifying firm-specific vulnerabilities that statistical models miss entirely. Know this concept cold — it shows up in both the market risk and operational risk sections.
Regulatory Context: FRTB and Stressed VaR
The Fundamental Review of the Trading Book (FRTB) represents the most significant overhaul of market risk capital rules since Basel II. Key points GARP tests:
- Banks must identify a stressed period (typically a 12-month window with significant financial stress relevant to the portfolio) and use it for ES calculations
- The internal models approach (IMA) requires desk-level approval — a bank can't apply IMA selectively across the entire trading book
- The standardized approach (SA) becomes mandatory fallback for desks that fail backtesting requirements
FRTB questions often test whether you understand the purpose of the framework — to reduce variability in risk-weighted asset calculations across banks — not just the mechanics.
---
Connecting the Dots: A Framework for Exam Day
Here's how to think about these three tools as a unified system — which is how GARP frames them:
1. VaR gives you a threshold under normal conditions. Fast, widely used, regulatory baseline (pre-FRTB). 2. ES extends into the tail under stress. More informative, now the regulatory standard under FRTB. 3. Stress Testing operates outside the distribution entirely. It catches what models miss — correlated market meltdowns, liquidity crises, idiosyncratic shocks.
A firm using all three is building overlapping defenses. VaR fails at the tail; ES catches the tail; stress testing catches the unmodeled. That logical structure appears directly in GARP exam questions that ask you to evaluate a risk management framework and identify what's missing.
---
The FRM Part 2 Mindset Shift
As a serious FRM candidate, the biggest trap in market risk is treating these as formulas to recall rather than tools to evaluate. GARP will describe a scenario — a bank's VaR model is producing too few exceedances, or a risk manager is choosing between ES and VaR for a non-linear derivatives book — and ask you to reason through the implications.
That reasoning only comes from genuine conceptual mastery. Not from re-reading a formula sheet the night before.
This is exactly where Clavis is built to help. Rather than drilling static flashcards, Clavis generates adaptive questions that test your ability to apply, evaluate, and critique — the same cognitive level GARP demands. Built by finance professionals who've sat through these exams, Clavis tracks what you actually understand versus what you've only seen before, so you walk into exam day with a verified picture of your readiness — not just a feeling.
If market risk is your weak spot, start stress-testing your knowledge now. The real exam won't give you a warm-up round.