What FRM Part 1 Valuation & Risk Models actually tests
Valuation and Risk Models is 30% of the Part 1 exam:
| Area | What it covers |
|---|---|
| Value at Risk (VaR) | Parametric, historical simulation, Monte Carlo methods |
| Expected Shortfall | Definition, advantages over VaR, regulatory context (Basel III) |
| Backtesting | Kupiec test, Basel traffic-light framework |
| Bond Valuation & Duration | YTM, spot/forward rates, Macaulay/modified duration, convexity |
| Options Pricing | Black-Scholes-Merton, binomial model, put-call parity |
| Options Greeks | Delta, gamma, theta, vega, rho, delta hedging |
| Volatility Surfaces | Implied volatility, smile, skew |
| Credit Risk Fundamentals | Default probability, loss given default, structural models (Merton) |
Why this topic is the exam's center of gravity
Valuation and Risk Models is where the quant tools from earlier topics (probability distributions, regression) and the instruments from Financial Markets & Products (bonds, options) converge into the actual risk-management output — a VaR number, a duration figure, a Greek. It's tested heavily because it's genuinely the connective tissue of the whole Part 1 syllabus.
Sample question: Parametric VaR
Valuation & Risk Models · Medium difficulty
A portfolio has a daily standard deviation of $500,000. Using the parametric (delta-normal) method, what is the 1-day 99% VaR? (z = 2.33 for 99% confidence)
A. $825,000
B. $1,165,000
C. $1,325,000
D. $1,500,000
The correct answer is B — $1,165,000.
Parametric VaR = z × σ = 2.33 × $500,000 = $1,165,000. This assumes returns are normally distributed and the portfolio has no significant options/nonlinear exposure — for portfolios with meaningful optionality, historical simulation or Monte Carlo VaR is more appropriate than the parametric method.
Parametric VaR = z × σ = 2.33 × $500,000 = $1,165,000. This assumes returns are normally distributed and the portfolio has no significant options/nonlinear exposure — for portfolios with meaningful optionality, historical simulation or Monte Carlo VaR is more appropriate than the parametric method.