Forecasting tail risk measures for financial time series: an extreme value approach with covariates

Robert James; Henry Leung; Jessica Wai Yin Leung; Artem Prokhorov

doi:10.1016/j.jempfin.2023.01.002

Forecasting tail risk measures for financial time series: an extreme value approach with covariates

Robert James, Henry Leung, Jessica Wai Yin Leung, Artem Prokhorov

Econometrics & Business Statistics

Research output: Contribution to journal › Article › Research › peer-review

4 Citations (Scopus)

Abstract

The paper develops a tail risk forecasting model that incorporates the wealth of economic and financial information available to risk managers. The approach can be viewed as a regularized extension of the two-stage GARCH-EVT model of McNeil and Frey (2000) where we permit a time-varying data-driven selection of a sparse set of covariates affecting the scale of the extreme value distribution of risk. We use a rich data set from the U.S. equity market to explore when this additional information improves Value-at-Risk and Expected Shortfall forecasts compared to popular tail risk forecasting methods such as the traditional and non-regularized GARCH-EVT models, and the GJR-GARCH(1,1), Hawkes POT model, CaViaR and CARE models. Under an extensive set of performance criteria and tests we demonstrate that our approach produces competitive risk forecasts, particularly during periods of financial distress.

Original language	English
Pages (from-to)	29-50
Number of pages	22
Journal	Journal of Empirical Finance
Volume	71
DOIs	https://doi.org/10.1016/j.jempfin.2023.01.002
Publication status	Published - Mar 2023

Keywords

Expected shortfall
Extreme value theory
GARCH models
Regularization
Value-at-risk
Variable selection

Access to Document

10.1016/j.jempfin.2023.01.002

Cite this

@article{db40f6357b6a4a77aeb8a59a2eba25b5,

title = "Forecasting tail risk measures for financial time series: an extreme value approach with covariates",

abstract = "The paper develops a tail risk forecasting model that incorporates the wealth of economic and financial information available to risk managers. The approach can be viewed as a regularized extension of the two-stage GARCH-EVT model of McNeil and Frey (2000) where we permit a time-varying data-driven selection of a sparse set of covariates affecting the scale of the extreme value distribution of risk. We use a rich data set from the U.S. equity market to explore when this additional information improves Value-at-Risk and Expected Shortfall forecasts compared to popular tail risk forecasting methods such as the traditional and non-regularized GARCH-EVT models, and the GJR-GARCH(1,1), Hawkes POT model, CaViaR and CARE models. Under an extensive set of performance criteria and tests we demonstrate that our approach produces competitive risk forecasts, particularly during periods of financial distress.",

keywords = "Expected shortfall, Extreme value theory, GARCH models, Regularization, Value-at-risk, Variable selection",

author = "Robert James and Henry Leung and Leung, {Jessica Wai Yin} and Artem Prokhorov",

note = "Funding Information: Helpful comments from seminar participants at Monash University, and participants of the 41st International Symposium on Forecasting and the International Conference on Econometrics and Business Analytics are gratefully acknowledged. The use of the University of Sydney{\textquoteright}s high performance computing cluster, Artemis, is acknowledged. This research was supported in part by the Monash eResearch Centre and eSolutions-Research Support Services through the use of the MonARCH HPC Cluster. Research for this paper was supported by grants from Australian Research Council (James, Project DP200103549 ) and Russian Science Foundation (Prokhorov, Project No. 22-18-00588 ) for various and non-overlapping parts of this research. Funding Information: Helpful comments from seminar participants at Monash University, and participants of the 41st International Symposium on Forecasting and the International Conference on Econometrics and Business Analytics are gratefully acknowledged. The use of the University of Sydney's high performance computing cluster, Artemis, is acknowledged. This research was supported in part by the Monash eResearch Centre and eSolutions-Research Support Services through the use of the MonARCH HPC Cluster. Research for this paper was supported by grants from Australian Research Council (James, Project DP200103549) and Russian Science Foundation (Prokhorov, Project No. 22-18-00588) for various and non-overlapping parts of this research. Publisher Copyright: {\textcopyright} 2023 Elsevier B.V.",

year = "2023",

month = mar,

doi = "10.1016/j.jempfin.2023.01.002",

language = "English",

volume = "71",

pages = "29--50",

journal = "Journal of Empirical Finance",

issn = "0927-5398",

publisher = "Elsevier",

}

TY - JOUR

T1 - Forecasting tail risk measures for financial time series

T2 - an extreme value approach with covariates

AU - James, Robert

AU - Leung, Henry

AU - Leung, Jessica Wai Yin

AU - Prokhorov, Artem

N1 - Funding Information: Helpful comments from seminar participants at Monash University, and participants of the 41st International Symposium on Forecasting and the International Conference on Econometrics and Business Analytics are gratefully acknowledged. The use of the University of Sydney’s high performance computing cluster, Artemis, is acknowledged. This research was supported in part by the Monash eResearch Centre and eSolutions-Research Support Services through the use of the MonARCH HPC Cluster. Research for this paper was supported by grants from Australian Research Council (James, Project DP200103549 ) and Russian Science Foundation (Prokhorov, Project No. 22-18-00588 ) for various and non-overlapping parts of this research. Funding Information: Helpful comments from seminar participants at Monash University, and participants of the 41st International Symposium on Forecasting and the International Conference on Econometrics and Business Analytics are gratefully acknowledged. The use of the University of Sydney's high performance computing cluster, Artemis, is acknowledged. This research was supported in part by the Monash eResearch Centre and eSolutions-Research Support Services through the use of the MonARCH HPC Cluster. Research for this paper was supported by grants from Australian Research Council (James, Project DP200103549) and Russian Science Foundation (Prokhorov, Project No. 22-18-00588) for various and non-overlapping parts of this research. Publisher Copyright: © 2023 Elsevier B.V.

PY - 2023/3

Y1 - 2023/3

N2 - The paper develops a tail risk forecasting model that incorporates the wealth of economic and financial information available to risk managers. The approach can be viewed as a regularized extension of the two-stage GARCH-EVT model of McNeil and Frey (2000) where we permit a time-varying data-driven selection of a sparse set of covariates affecting the scale of the extreme value distribution of risk. We use a rich data set from the U.S. equity market to explore when this additional information improves Value-at-Risk and Expected Shortfall forecasts compared to popular tail risk forecasting methods such as the traditional and non-regularized GARCH-EVT models, and the GJR-GARCH(1,1), Hawkes POT model, CaViaR and CARE models. Under an extensive set of performance criteria and tests we demonstrate that our approach produces competitive risk forecasts, particularly during periods of financial distress.

AB - The paper develops a tail risk forecasting model that incorporates the wealth of economic and financial information available to risk managers. The approach can be viewed as a regularized extension of the two-stage GARCH-EVT model of McNeil and Frey (2000) where we permit a time-varying data-driven selection of a sparse set of covariates affecting the scale of the extreme value distribution of risk. We use a rich data set from the U.S. equity market to explore when this additional information improves Value-at-Risk and Expected Shortfall forecasts compared to popular tail risk forecasting methods such as the traditional and non-regularized GARCH-EVT models, and the GJR-GARCH(1,1), Hawkes POT model, CaViaR and CARE models. Under an extensive set of performance criteria and tests we demonstrate that our approach produces competitive risk forecasts, particularly during periods of financial distress.

KW - Expected shortfall

KW - Extreme value theory

KW - GARCH models

KW - Regularization

KW - Value-at-risk

KW - Variable selection

UR - http://www.scopus.com/inward/record.url?scp=85147810306&partnerID=8YFLogxK

U2 - 10.1016/j.jempfin.2023.01.002

DO - 10.1016/j.jempfin.2023.01.002

M3 - Article

AN - SCOPUS:85147810306

SN - 0927-5398

VL - 71

SP - 29

EP - 50

JO - Journal of Empirical Finance

JF - Journal of Empirical Finance

ER -

Forecasting tail risk measures for financial time series: an extreme value approach with covariates

Abstract

Keywords

Access to Document

Other files and links

ARC Training Centre in Optimisation Technologies, Integrated Methodologies, and Applications (OPTIMA)

Cite this

Forecasting tail risk measures for financial time series: an extreme value approach with covariates

Abstract

Keywords

Access to Document

Other files and links

Projects

ARC Training Centre in Optimisation Technologies, Integrated Methodologies, and Applications (OPTIMA)

Cite this