Monte Carlo Simulation & VaR Flashcards
7 cards from real CRA practice questions. Tap to flip, then mark Knew It or Still Learning — missed cards come back until you master them.
Read the first 7 Monte Carlo Simulation & VaR flashcards as text
Under the Fundamental Review of the Trading Book (FRTB), which risk measure replaces VaR as the primary regulatory capital metric for market risk?
Answer: Expected Shortfall at 97.5%
FRTB replaced VaR with Expected Shortfall calibrated at the 97.5th percentile, which is roughly equivalent in tail coverage to 99% VaR but is a coherent risk measure.
In importance sampling (a Monte Carlo variance reduction technique), how is the distribution altered to improve VaR estimation efficiency?
Answer: More probability mass is placed on tail scenarios and reweighted to correct for the shift
Importance sampling shifts the sampling distribution toward the tail of interest and applies a likelihood ratio correction (Radon-Nikodym derivative) to obtain unbiased estimates with lower variance.
A risk analyst is modeling credit portfolio losses using Monte Carlo simulation and a one-factor Gaussian copula. What does the single systematic factor represent?
Answer: The common macroeconomic or market factor driving asset value correlations
In the one-factor Gaussian copula, a single systematic factor represents a common macroeconomic driver that creates correlation between obligors' asset values, linking their default probabilities.
Which of the following VaR methods is most sensitive to the choice of the lookback window length?
Answer: Historical Simulation VaR
Historical Simulation directly uses observed returns over a chosen window; a short window may miss crises while a long window may include irrelevant old regimes, making window length a critical choice.
A delta-gamma approximation in Monte Carlo VaR is used to:
Answer: Improve the accuracy of option price changes by including second-order Greeks
The delta-gamma (or Taylor series) approximation improves on the linear delta approximation by including gamma (second-order sensitivity to price changes), better capturing option non-linearity.
In Monte Carlo simulation, quasi-random (low-discrepancy) sequences such as Sobol or Halton sequences are used instead of pseudo-random numbers primarily to:
Answer: Improve coverage of the sample space and reduce convergence error
Low-discrepancy sequences fill the sample space more uniformly than pseudo-random numbers, reducing the convergence error of Monte Carlo estimates for a given number of draws.
When a Monte Carlo VaR model is 'backtested' and produces too many exceptions (actual losses exceeding VaR), which of the following is a likely root cause?
Answer: Volatility is underestimated or fat tails are not modeled
Excess VaR exceptions typically indicate that the model underestimates volatility or uses distributions that are too thin-tailed, causing VaR to be systematically understated.