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
Stressed VaR (SVaR), required under Basel 2.5, differs from regular VaR primarily because it:
Answer: Is calibrated to a continuous 12-month stress period
Stressed VaR uses a continuous 12-month window of significant financial stress (e.g., 2007–2008) to calibrate the model, capturing tail risk not present in recent calm periods.
In a Monte Carlo simulation of credit risk, which parameter most directly drives the probability of joint defaults in a correlated default model?
Answer: Asset correlation
Asset correlation (often proxied by equity correlation) drives the tendency for obligors to default together, making it the key parameter in correlated default simulations.
A risk manager increases the number of Monte Carlo simulations from 1,000 to 10,000. By approximately what factor does the standard error of the VaR estimate decrease?
Answer: 3.16x
Standard error decreases by √(10,000/1,000) = √10 ≈ 3.16, since Monte Carlo error is proportional to 1/√N.
Which backtesting framework, adopted by Basel, counts the number of days over a 250-day window that actual losses exceed the reported VaR?
Answer: Traffic light approach
The Basel traffic light approach categorizes model performance into green, yellow, and red zones based on the number of VaR exceptions in a 250-trading-day window.
Which of the following is NOT a limitation of parametric (variance-covariance) VaR that Monte Carlo simulation is designed to address?
Answer: Difficulty modeling correlation structure between many assets
Modeling correlation structure among many assets is a challenge for all VaR methods, including Monte Carlo; it is not a specific limitation of parametric VaR that Monte Carlo uniquely solves.
A Monte Carlo simulation generates 5,000 portfolio loss scenarios. To find the 99% VaR, you would:
Answer: Identify the loss at the 50th largest observation
At 99% confidence with 5,000 scenarios, the VaR is the 50th largest loss (5,000 × 1% = 50), i.e., the 50th worst outcome.
Which of the following describes 'model risk' specific to Monte Carlo VaR?
Answer: The risk that incorrect distributional or parameter assumptions produce misleading risk estimates
Model risk in Monte Carlo VaR arises from incorrect assumptions about return distributions, volatility processes, or correlations, which can systematically under- or overstate true risk.