Tail Risk Measures (Expected Shortfall, Spectral, Expectile)
Tail Risk Measures (Expected Shortfall, Spectral and Expectile Risk) · Also known as: expected shortfall, conditional value at risk, CVaR, spectral risk measure, expectile risk measure, coherent risk measure, Kuyruk Riski Ölçüleri (ES, Spectral, Expectile)
Tail risk measures quantify the loss distribution beyond Value-at-Risk (VaR). Expected Shortfall — the expected loss given that VaR is exceeded — is the leading coherent risk measure, formalised by Artzner, Delbaen, Eber and Heath (1999) and shown to be coherent by Acerbi and Tasche (2002). Spectral and expectile-based measures generalise it.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use tail risk measures when you must quantify downside risk for a continuous return or loss series, especially with fat-tailed data (kurtosis > 3) where normal VaR understates extreme losses. They need a reasonably long history (about 250 observations or more) and are the natural choice under Basel III/IV, which adopts Expected Shortfall at 97.5% as the standard regulatory measure. Prefer them to plain VaR whenever subadditivity (a coherent, diversification-respecting measure) matters.
Strengths & limitations
- Expected Shortfall is a coherent risk measure: unlike VaR it is subadditive, so diversification never increases measured risk.
- Captures the magnitude of losses beyond the VaR threshold, not just their probability of being exceeded.
- Adopted as the regulatory standard under Basel III/IV (Expected Shortfall at 97.5%), and generalises smoothly to spectral and expectile measures that encode investor risk aversion.
- Estimation is sensitive in the deep tail: with limited data the average beyond VaR rests on very few observations and can be unstable.
- Backtesting Expected Shortfall is harder than backtesting VaR and requires dedicated tests such as McNeil & Frey (2000) or Acerbi & Szekely (2014).
- Results depend on how the tail is modelled — historical estimates assume the past represents the future, while parametric estimates depend on the chosen distribution.
Frequently asked
How is Expected Shortfall different from VaR?
VaR is a threshold: the loss that will not be exceeded with a given confidence. Expected Shortfall is the average loss in the tail beyond that threshold, so it measures how severe the bad outcomes are, not just where they begin. Crucially, Expected Shortfall is coherent (subadditive) while VaR is not.
Why is coherence (subadditivity) important?
A subadditive measure never reports the risk of a combined portfolio as larger than the sum of its parts, so it always rewards diversification. VaR can violate this, occasionally penalising diversified portfolios. Artzner et al. (1999) showed coherence is a prerequisite for a well-behaved risk measure, and Acerbi & Tasche (2002) proved Expected Shortfall satisfies it.
What are spectral and expectile risk measures?
They generalise Expected Shortfall. A spectral measure integrates VaR across probability levels weighted by a risk-aversion spectrum that grows toward the tail; expectile-based measures define risk through asymmetrically weighted squared deviations. Both remain coherent and let you tune how heavily extreme losses are penalised.
Why does Basel III/IV prefer Expected Shortfall?
Because Expected Shortfall is coherent and captures the size of losses deep in the tail, regulators adopted it at the 97.5% level as the standard market-risk measure, replacing VaR which ignores how bad the worst losses can be.
Sources
- Artzner, P., Delbaen, F., Eber, J.-M. & Heath, D. (1999). Coherent Measures of Risk. Mathematical Finance, 9(3), 203–228. DOI: 10.1111/1467-9965.00068 ↗
- Acerbi, C. & Tasche, D. (2002). On the Coherence of Expected Shortfall. Journal of Banking & Finance, 26(7), 1487–1503. DOI: 10.1016/S0378-4266(02)00283-2 ↗
How to cite this page
ScholarGate. (2026, June 1). Tail Risk Measures (Expected Shortfall, Spectral and Expectile Risk). ScholarGate. https://scholargate.app/en/finance/tail-risk-measures
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Extreme Value TheoryFinance↔ compare
- GARCH ModelEconometrics↔ compare
- OLS RegressionEconometrics↔ compare
- Quantile RegressionEconometrics↔ compare
- Regime-Switching ModelFinance↔ compare