Forecasting risk with Markov-switching GARCH models: A large-scale performance study

David Ardia, Keven Bluteau, Kris Boudt, Leopoldo Catania

Research output: Contribution to JournalArticleAcademicpeer-review

Abstract

We perform a large-scale empirical study in order to compare the forecasting performances of single-regime and Markov-switching GARCH (MSGARCH) models from a risk management perspective. We find that MSGARCH models yield more accurate Value-at-Risk, expected shortfall, and left-tail distribution forecasts than their single-regime counterparts for daily, weekly, and ten-day equity log-returns. Also, our results indicate that accounting for parameter uncertainty improves the left-tail predictions, independently of the inclusion of the Markov-switching mechanism.

Original languageEnglish
Pages (from-to)733-747
Number of pages15
JournalInternational Journal of Forecasting
Volume34
Issue number4
DOIs
Publication statusPublished - Oct 2018

Fingerprint

Markov switching
GARCH model
Value at risk
Regime switching
Parameter uncertainty
Risk management
Equity
Forecasting performance
Prediction
Expected shortfall
Empirical study
Inclusion

Keywords

  • Expected shortfall
  • Forecasting performance
  • GARCH
  • Large-scale study
  • MSGARCH
  • Risk management
  • Value-at-risk

Cite this

Ardia, David ; Bluteau, Keven ; Boudt, Kris ; Catania, Leopoldo. / Forecasting risk with Markov-switching GARCH models : A large-scale performance study. In: International Journal of Forecasting. 2018 ; Vol. 34, No. 4. pp. 733-747.
@article{1a72c66fa10447879b0cf52ec8b4d0aa,
title = "Forecasting risk with Markov-switching GARCH models: A large-scale performance study",
abstract = "We perform a large-scale empirical study in order to compare the forecasting performances of single-regime and Markov-switching GARCH (MSGARCH) models from a risk management perspective. We find that MSGARCH models yield more accurate Value-at-Risk, expected shortfall, and left-tail distribution forecasts than their single-regime counterparts for daily, weekly, and ten-day equity log-returns. Also, our results indicate that accounting for parameter uncertainty improves the left-tail predictions, independently of the inclusion of the Markov-switching mechanism.",
keywords = "Expected shortfall, Forecasting performance, GARCH, Large-scale study, MSGARCH, Risk management, Value-at-risk",
author = "David Ardia and Keven Bluteau and Kris Boudt and Leopoldo Catania",
year = "2018",
month = "10",
doi = "10.1016/j.ijforecast.2018.05.004",
language = "English",
volume = "34",
pages = "733--747",
journal = "International Journal of Forecasting",
issn = "0169-2070",
publisher = "Elsevier",
number = "4",

}

Forecasting risk with Markov-switching GARCH models : A large-scale performance study. / Ardia, David; Bluteau, Keven; Boudt, Kris; Catania, Leopoldo.

In: International Journal of Forecasting, Vol. 34, No. 4, 10.2018, p. 733-747.

Research output: Contribution to JournalArticleAcademicpeer-review

TY - JOUR

T1 - Forecasting risk with Markov-switching GARCH models

T2 - A large-scale performance study

AU - Ardia, David

AU - Bluteau, Keven

AU - Boudt, Kris

AU - Catania, Leopoldo

PY - 2018/10

Y1 - 2018/10

N2 - We perform a large-scale empirical study in order to compare the forecasting performances of single-regime and Markov-switching GARCH (MSGARCH) models from a risk management perspective. We find that MSGARCH models yield more accurate Value-at-Risk, expected shortfall, and left-tail distribution forecasts than their single-regime counterparts for daily, weekly, and ten-day equity log-returns. Also, our results indicate that accounting for parameter uncertainty improves the left-tail predictions, independently of the inclusion of the Markov-switching mechanism.

AB - We perform a large-scale empirical study in order to compare the forecasting performances of single-regime and Markov-switching GARCH (MSGARCH) models from a risk management perspective. We find that MSGARCH models yield more accurate Value-at-Risk, expected shortfall, and left-tail distribution forecasts than their single-regime counterparts for daily, weekly, and ten-day equity log-returns. Also, our results indicate that accounting for parameter uncertainty improves the left-tail predictions, independently of the inclusion of the Markov-switching mechanism.

KW - Expected shortfall

KW - Forecasting performance

KW - GARCH

KW - Large-scale study

KW - MSGARCH

KW - Risk management

KW - Value-at-risk

UR - http://www.scopus.com/inward/record.url?scp=85051373011&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85051373011&partnerID=8YFLogxK

U2 - 10.1016/j.ijforecast.2018.05.004

DO - 10.1016/j.ijforecast.2018.05.004

M3 - Article

VL - 34

SP - 733

EP - 747

JO - International Journal of Forecasting

JF - International Journal of Forecasting

SN - 0169-2070

IS - 4

ER -