Comparison of VaR and CVaR
| :--- | :--- |
| VaR | – Easy to understand.
– Easy to interpret statistically, widely used in risk management.
– Can express risk with single value. | – Possibility to ignore extreme loss.
– When assuming normal distribution, may not match actual distribution.
– Inaccurate if distribution of risk is asymmetric. |
| CVaR | – Consider extreme loss. | – Depends on quality and quantity of data.
– Value becomes large because considering outliers. |
Risk evaluation is possible using either, but it is good to select according to personal goal whether to consider rare large loss or not.
