Extreme M-quantiles as risk measures: from L1 to Lp optimization

The class of quantiles lies at the heart of extreme-value theory and is one of the basic tools in risk management. The alternative family of expectiles is based on squared rather than absolute error loss minimization. It has recently been receiving a lot of attention in actuarial science, econometri...

Full description

Bibliographic Details
Main Authors: Daouia, Abdelaati, Girard, Stéphane, Stupfler, Gilles
Format: Article
Published: Bernoulli Society for Mathematical Statistics and Probability 2019
Subjects:
Online Access:https://eprints.nottingham.ac.uk/45682/
_version_ 1848797173385265152
author Daouia, Abdelaati
Girard, Stéphane
Stupfler, Gilles
author_facet Daouia, Abdelaati
Girard, Stéphane
Stupfler, Gilles
author_sort Daouia, Abdelaati
building Nottingham Research Data Repository
collection Online Access
description The class of quantiles lies at the heart of extreme-value theory and is one of the basic tools in risk management. The alternative family of expectiles is based on squared rather than absolute error loss minimization. It has recently been receiving a lot of attention in actuarial science, econometrics and statistical finance. Both quantiles and expectiles can be embedded in a more general class of M-quantiles by means of Lp optimization. These generalized Lp-quantiles steer an advantageous middle course between ordinary quantiles and expectiles without sacrificing their virtues too much for 1 < p < 2. In this paper, we investigate their estimation from the perspective of extreme values in the class of heavy-tailed distributions. We construct estimators of the intermediate Lp-quantiles and establish their asymptotic normality in a dependence framework motivated by financial and actuarial applications, before extrapolating these estimates to the very far tails. We also investigate the potential of extreme Lp-quantiles as a tool for estimating the usual quantiles and expectiles themselves. We show the usefulness of extreme Lp-quantiles and elaborate the choice of p through applications to some simulated and financial real data.
first_indexed 2025-11-14T19:59:40Z
format Article
id nottingham-45682
institution University of Nottingham Malaysia Campus
institution_category Local University
last_indexed 2025-11-14T19:59:40Z
publishDate 2019
publisher Bernoulli Society for Mathematical Statistics and Probability
recordtype eprints
repository_type Digital Repository
spelling nottingham-456822020-05-04T19:05:17Z https://eprints.nottingham.ac.uk/45682/ Extreme M-quantiles as risk measures: from L1 to Lp optimization Daouia, Abdelaati Girard, Stéphane Stupfler, Gilles The class of quantiles lies at the heart of extreme-value theory and is one of the basic tools in risk management. The alternative family of expectiles is based on squared rather than absolute error loss minimization. It has recently been receiving a lot of attention in actuarial science, econometrics and statistical finance. Both quantiles and expectiles can be embedded in a more general class of M-quantiles by means of Lp optimization. These generalized Lp-quantiles steer an advantageous middle course between ordinary quantiles and expectiles without sacrificing their virtues too much for 1 < p < 2. In this paper, we investigate their estimation from the perspective of extreme values in the class of heavy-tailed distributions. We construct estimators of the intermediate Lp-quantiles and establish their asymptotic normality in a dependence framework motivated by financial and actuarial applications, before extrapolating these estimates to the very far tails. We also investigate the potential of extreme Lp-quantiles as a tool for estimating the usual quantiles and expectiles themselves. We show the usefulness of extreme Lp-quantiles and elaborate the choice of p through applications to some simulated and financial real data. Bernoulli Society for Mathematical Statistics and Probability 2019-02-01 Article PeerReviewed Daouia, Abdelaati, Girard, Stéphane and Stupfler, Gilles (2019) Extreme M-quantiles as risk measures: from L1 to Lp optimization. Bernoulli, 25 (1). pp. 264-309. ISSN 1573-9759 Asymptotic normality; Dependent observations; Expectiles; Extrapolation; Extreme values; Heavy tails; Lp optimization; Mixing; Quantiles; Tail risk https://projecteuclid.org/euclid.bj/1544605247
spellingShingle Asymptotic normality; Dependent observations; Expectiles; Extrapolation; Extreme values; Heavy tails; Lp optimization; Mixing; Quantiles; Tail risk
Daouia, Abdelaati
Girard, Stéphane
Stupfler, Gilles
Extreme M-quantiles as risk measures: from L1 to Lp optimization
title Extreme M-quantiles as risk measures: from L1 to Lp optimization
title_full Extreme M-quantiles as risk measures: from L1 to Lp optimization
title_fullStr Extreme M-quantiles as risk measures: from L1 to Lp optimization
title_full_unstemmed Extreme M-quantiles as risk measures: from L1 to Lp optimization
title_short Extreme M-quantiles as risk measures: from L1 to Lp optimization
title_sort extreme m-quantiles as risk measures: from l1 to lp optimization
topic Asymptotic normality; Dependent observations; Expectiles; Extrapolation; Extreme values; Heavy tails; Lp optimization; Mixing; Quantiles; Tail risk
url https://eprints.nottingham.ac.uk/45682/
https://eprints.nottingham.ac.uk/45682/