How Level II tests VaR methods, scenario and sensitivity risk, backtesting, and simulation judgment in market-risk analysis.
Market-risk questions at Level II are about whether the risk tool fits the portfolio and the decision. A single VaR number is not the answer. It is only one way of summarizing potential loss, and the exam expects you to know when that summary is useful, when it is weak, and what other evidence should sit beside it.
Candidates often memorize three VaR methods and stop there. Level II goes further.
A simplified parametric form is often written as:
$$ \text{VaR}\alpha \approx z\alpha \sigma_P V $$
where the number depends on the confidence level, portfolio volatility, and exposure size.
| Method | Main idea | Typical strength | Typical weakness |
|---|---|---|---|
| Parametric or variance-covariance | Uses distribution assumptions and moments | Fast and tractable | Can understate nonlinear or fat-tail risk |
| Historical simulation | Replays realized history | Intuitive and distribution-light | History may not include the right stress |
| Monte Carlo simulation | Simulates many possible paths | Flexible for complex portfolios | Model assumptions can dominate the answer |
The exam often asks which method best matches the portfolio structure, not which one sounds most sophisticated.
| Tool | What it adds beyond VaR |
|---|---|
| Sensitivity measures | Show how value changes with small moves in a risk factor |
| Scenario measures | Show what happens under a specific shock or regime |
| Stress testing | Focuses on adverse but plausible outcomes outside ordinary assumptions |
That is why Level II often treats VaR as one tool in a broader risk dashboard.
| Limit type | What it tries to control |
|---|---|
| Risk budget | Total active or market-risk usage |
| Position limit | Concentration in one exposure |
| Scenario limit | Loss under a specific stress event |
| Stop-loss limit | Drawdown response discipline |
The exam may ask which limit type best addresses the risk described in the vignette.
flowchart TD
A["Define strategy or portfolio"] --> B["Choose assumptions, data window, and rules"]
B --> C["Run backtest or simulation"]
C --> D["Evaluate returns, drawdowns, and risk metrics"]
D --> E["Check for bias, unrealistic assumptions, and weak robustness"]
E --> F["Decide whether the result is decision-useful"]
This is where many weak answers fail. They read the performance summary but never question the design quality of the test.
| Red flag | Why it matters |
|---|---|
| Look-ahead bias | Uses information that would not have been known at the time |
| Survivorship bias | Overstates performance by excluding failures |
| Overfitting to history | Good past fit may collapse out of sample |
| Unrealistic turnover or cost assumptions | Net implementable performance may be much lower |
Level II often tests whether you can diagnose these issues from the setup, not just from the return chart.
A simulated strategy shows excellent historical returns with low VaR, but trading costs were ignored and the rule set was tuned repeatedly on the same sample.
A weak answer highlights the strong return and low apparent risk.
A stronger answer questions whether the simulation is overfit and whether the risk estimate is understating the true implementable risk.
What is the strongest reason to supplement VaR with scenario analysis?
Best answer: Scenario analysis can reveal losses under specific stress conditions that a single VaR summary may not capture well.
Why: Level II often tests whether you understand that risk management needs multiple lenses, not a single headline statistic.