Definition
A loss quantile at probability alpha is a value q such that the cumulative probability of loss at or below q reaches alpha, with a chosen convention for discontinuities. Quantiles turn a distribution into a threshold and are the building block for loss-based VaR. Sample quantiles require an interpolation or order-statistic rule.
How it is calculated
Loss quantile q-alpha of L is the infimum of loss values ell whose loss CDF F-L reaches at least alpha.
| Symbol | Meaning | Unit |
|---|---|---|
| q | quantile | loss units |
| alpha | quantile probability | probability |
| L | loss random variable | loss units |
| ell | candidate loss value | loss units |
| F | loss cumulative distribution | probability |
Interpretation
A quantile is a position marker in an ordered distribution. In a loss convention, moving alpha toward one moves the marker deeper into worse outcomes. With a finite sample there may be a jump rather than a smooth point, so the convention matters.
Assumptions and conventions
-
The CDF and quantile convention are specified.
-
The probability satisfies 0<alpha<1.
-
For empirical data, interpolation and ties are handled explicitly.
Loss units; alpha and F are probabilities.
Limitations and common misreadings
-
A quantile is not necessarily an observed value when interpolation is used.
-
For L=-R, upper loss and lower return quantiles describe the same adverse outcomes; discrete quantile endpoints require care.
References
- Alexander J. McNeil. (2005). Quantitative risk management: concepts, techniques, and tools. Chapter 2, Basic Concepts in Risk Management. isbn:9780691122557. Source

