FRM Part I · FRM Exam Part I · Measures of Financial Risk
A risk manager reviews four properties of a coherent risk measure: monotonicity, translation invariance, positive homogeneity and subadditivity. Which statement about Value-at-Risk (VaR) is correct?
VaR satisfies monotonicity, translation invariance and positive homogeneity but can fail subadditivity for non-elliptical or discrete loss distributions, so it is not a coherent risk measure. Diversification can then appear to increase measured risk.
- AVaR satisfies all four properties for any return distribution
- BVaR satisfies monotonicity but can violate subadditivityCorrect
- CVaR always satisfies subadditivity but violates monotonicity
- DVaR violates translation invariance but satisfies subadditivity
Explanation
VaR is monotonic, translation invariant and positively homogeneous, but it is not subadditive in general, so combining portfolios can show a VaR larger than the sum of the stand-alone VaRs. Hence it is not coherent. The option claiming VaR is always subadditive is wrong because discrete or skewed loss distributions give counterexamples.
Did you get it right without looking?
One question tells you little. A timed set on Measures of Financial Risk shows your real accuracy, how long you take and where you lose marks.
More Measures of Financial Risk questions
- A loss L takes values 0, 10 and 20 with probabilities 0.36, 0.48 and 0.16. A distortion risk measure is defined as ρ(L) = ∫ g(S(x)) dx over …
- A portfolio manager estimates a one-day 99% VaR of USD 5 million. Over 250 trading days, how many days should the manager expect the loss to…
- A risk manager has a one-day 95% VaR of USD 3.0 million for a portfolio with normally distributed returns and zero mean. Using z-values of 1…
- Portfolio returns are normally distributed with mean zero and a daily standard deviation of USD 1.0 million in loss terms. The 99% VaR is 2.…
- A bank holds two independent loans, each of USD 10 million. Each defaults with probability 4% over the year, causing a total loss of the USD…
- A portfolio's one-period loss is normally distributed with mean 0 and standard deviation 10 (in USD million). Using the normal distribution,…