Expected Shortfall (ES / CVaR)
Quant ResearchLevel: Intermediate
Expected Shortfall (ES / CVaR)
Hints
- Specify the horizon, loss distribution and probability level.
- Consider tail behavior separately from scenarios absent in historical data.
Answer
For integrable loss, ES_α=(1/(1−α))∫_α¹VaR_u du: the average of the worst 1−α probability mass. For a continuous loss distribution this equals E[L | L≥VaR_α], but atoms at the quantile require fractional weighting rather than that naive conditional average. ES is coherent, including subadditivity, and captures tail severity beyond VaR. Its estimation is data-hungry and model-sensitive because relatively few observations lie in the relevant tail.
Source and licence: Quantitative Finance: From Mathematical Foundations to Trading Desk Reality — zcoderz · MIT · answer: Quantoros editorial