Explanation

Option A is correct: Only the parametric method assumes that return distributions for risk factors are normal.

Key differences between VaR methods:

Parametric Method (Variance-Covariance Method):

• Assumes normal distribution of returns
• Uses mean, variance, and covariance of returns
• Relies on the assumption that returns follow a normal distribution
• Most efficient computationally but may be inaccurate if returns are not normally distributed

Historical Simulation Method:

• Does NOT assume any specific distribution
• Uses actual historical returns to simulate future outcomes
• Captures non-normal characteristics like fat tails and skewness
• More realistic but requires sufficient historical data

Monte Carlo Simulation:

• Can assume various distributions (not necessarily normal)
• More flexible in modeling complex distributions

Therefore, only the parametric method explicitly assumes normal distribution of returns for risk factors.