When the P&L distribution has fat tails, very large losses are more likely than a normal distribution assumes. As a result, a variance-covariance VaR that assumes a normal distribution may understate losses. Expected shortfall, which averages the losses beyond VaR, reflects fat tails more strongly than VaR does.
Fat-tailed distribution
- Japanese
- 裾の厚い分布
- Last reviewed
- September 26, 2026
This is a draft, shown for review before publication.
A fat-tailed distribution is a probability distribution in which values far from the mean, such as very large losses, are more likely than in a normal distribution.